diff --git a/.gitattributes b/.gitattributes index 3da8dab2d02de28e331d6f3af589171fe46429e5..97af60d27ea9c214269b58bc438f47dffbe680a3 100644 --- a/.gitattributes +++ b/.gitattributes @@ -55,3 +55,4 @@ ultraep/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__ultraep__matched__seq8192__gbs ultraep/tierbtrace/tierbtrace_gbs8/QWEN2_57B__ultraep__matched__seq4096__gbs8__cap1p08984375/nsys/QWEN2_57B__ultraep__matched__seq4096__gbs8__cap1p08984375.nsys-rep filter=lfs diff=lfs merge=lfs -text moonep/tierbtrace/tierbtrace_gbs8/DSV2_LITE_L24__moonep__matched__seq8192__gbs8__capnone/nsys/DSV2_LITE_L24__moonep__matched__seq8192__gbs8__capnone.nsys-rep filter=lfs diff=lfs merge=lfs -text moonep/tierbtrace/tierbtrace_gbs8/QWEN2_57B__moonep__matched__seq4096__gbs8__capnone/nsys/QWEN2_57B__moonep__matched__seq4096__gbs8__capnone.nsys-rep filter=lfs diff=lfs merge=lfs -text +tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep filter=lfs diff=lfs merge=lfs -text diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/manifest.txt b/tierb-fixed-moonep-full/echo/tierbtrace/manifest.txt new file mode 100644 index 0000000000000000000000000000000000000000..9dca62a9b286083924585689984828f03bfad18e --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/manifest.txt @@ -0,0 +1,19 @@ +timestamp: 2026-08-30T00:10:30+00:00 +host: sp-aiinfra-interactive-aghosh-7jtjl-8xb200-moe-cg-0 +gpu: NVIDIA B200 x8 +driver: 580.95.05 +study_scripts_branch: granularity-echo +study_scripts_commit: 5054c89 +campaign: echo +backend: echo +phase: tierbtrace +gbs: 8 +seq: 4096 +configs: G1 G2 G3 G4 +capacity: 1.25 +repeats: 3 +nsys: 1 +megatron_repo: /workspace/Megatron-MoE-Echo +megatron_branch_required: bracing-regions-through-kernels +megatron_branch_actual: bracing-regions-through-kernels +megatron_commit: 9a236793 diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/pip_freeze.txt b/tierb-fixed-moonep-full/echo/tierbtrace/pip_freeze.txt new file mode 100644 index 0000000000000000000000000000000000000000..dd1d721bf3311308e8d733e6dfad151058f38459 --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/pip_freeze.txt @@ -0,0 +1,317 @@ +absl-py==2.3.0 +aiohappyeyeballs @ file:///rapids/aiohappyeyeballs-2.6.1-py3-none-any.whl#sha256=f349ba8f4b75cb25c99c5c2d84e997e485204d2902a9597802b0371f09331fb8 +aiohttp @ file:///rapids/aiohttp-3.12.7-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=94f98e0e5a49f89b252e115844f756c04fc8050f38252a32a3dd994ce8121f10 +aiosignal @ file:///rapids/aiosignal-1.3.2-py2.py3-none-any.whl#sha256=45cde58e409a301715980c2b01d0c28bdde3770d8290b5eb2173759d9acb31a5 +annotated-types==0.7.0 +anyio==4.9.0 +apex @ file:///opt/pytorch/apex/dist/apex-0.1-cp312-cp312-linux_x86_64.whl#sha256=a1508e4a386362e9c8f46c92ccada9110e421212b964c694235d0fd5a224b170 +argon2-cffi==25.1.0 +argon2-cffi-bindings==21.2.0 +arrow==1.3.0 +asciitree @ file:///rapids/asciitree-0.3.3-py3-none-any.whl#sha256=3e507ccfaeb417823158091fdb9d26b0205f6a59ba680b479a71db65c73d01d4 +asttokens==3.0.0 +astunparse==1.6.3 +async-lru==2.0.5 +attrs==25.3.0 +audioread==3.0.1 +babel==2.17.0 +beautifulsoup4==4.13.4 +black==25.1.0 +bleach==6.2.0 +blis==0.7.11 +cachetools @ file:///rapids/cachetools-6.0.0-py3-none-any.whl#sha256=82e73ba88f7b30228b5507dce1a1f878498fc669d972aef2dde4f3a3c24f103e +catalogue==2.0.10 +certifi==2025.4.26 +cffi==1.17.1 +charset-normalizer==3.4.2 +click==8.5.0 +cloudpathlib==0.21.1 +cloudpickle @ file:///rapids/cloudpickle-3.1.1-py3-none-any.whl#sha256=c8c5a44295039331ee9dad40ba100a9c7297b6f988e50e87ccdf3765a668350e +cmake==3.31.6 +comm==0.2.2 +confection==0.1.5 +contourpy==1.3.2 +cuda-bindings @ file:///rapids/cuda_bindings-12.9.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=ff0e28d1e34758654b9c961e1f55e4786e49aee6a4dbceaf3cc24c46c672df7e +cuda-python @ file:///rapids/cuda_python-12.9.0-py3-none-any.whl#sha256=926acba49b2c0a0374c61b7c98f337c085199cf51cdfe4d6423c4129c20547a7 +cudf @ file:///rapids/cudf-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b779a8e8c0e8577b9b560895958496da211bfac49cdc4f11a49e8d18cb82e348 +cudf-polars @ file:///rapids/cudf_polars-25.4.0-py3-none-any.whl#sha256=227b6c5147e93ea86282c8f7219118f916769fa2ca0cdd931e97a148aa66b869 +cugraph @ file:///rapids/cugraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b837c2dc2f7b235c702e4f8df64479b8fcbb342f848516d5faf0cdc4b5f31f29 +cugraph-service-client @ file:///rapids/cugraph_service_client-25.4.0-py3-none-any.whl#sha256=7e568c3e41154b9752f5173555ffcc282ac52d72343feeb79653b2e2b1232470 +cugraph-service-server @ file:///rapids/cugraph_service_server-25.4.0-py3-none-any.whl#sha256=a2830dcffafcccb16e12ebf85b98f8f247869d38c01ab4dd8664a1e7439d31be +cuml @ file:///rapids/cuml-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=1e90a464a18b094e24797f1676688c5d7696c4cf97c934504214a53d5aa09aa5 +cupy-cuda12x @ file:///rapids/cupy_cuda12x-13.3.0-cp312-cp312-linux_x86_64.whl#sha256=6a2dfa0aa1e8189a10d1534276c08f1d2f73b747fd28bfbe7b9f4ef66d9af73f +cuvs @ file:///rapids/cuvs-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b12f6b9f80e053a3b38542c2947d2516e468c02cad54a13d8b1a4a61d5aa7914 +cycler==0.12.1 +cymem==2.0.11 +Cython==3.1.1 +dask @ file:///rapids/dask-2025.2.0-py3-none-any.whl#sha256=f0fdeef6ceb0a06569d456c9e704f220f7f54e80f3a6ea42ab98cea6bc642b6e +dask-cuda @ file:///rapids/dask_cuda-25.4.0-py3-none-any.whl#sha256=442d87da0a95d00ca2a262b64c76e785a32ad905c82000fa2227932f81390e19 +dask-cudf @ file:///rapids/dask_cudf-25.4.0-py3-none-any.whl#sha256=5e7cea0b26daea17da6e0bb457435e24460f76ba740358a3a18e239d6015c95c +datasets==5.0.1 +debugpy==1.8.14 +decorator==5.2.1 +-e git+https://github.com/deepseek-ai/DeepEP.git@34152ae28f80bcc3ee38d7a12cb2ad87cfd4ea72#egg=deep_ep +defusedxml==0.7.1 +dill==0.4.1 +distributed @ file:///rapids/distributed-2025.2.0-py3-none-any.whl#sha256=368462084be47092d0a835f6c64c9b852da64dc88302b76d5f93bd5131228c67 +distributed-ucxx @ file:///rapids/distributed_ucxx-0.43.0-py3-none-any.whl#sha256=94f60d0ccb0c94ba8dbfa1eaf455874d18aaab17e2ae218c38590e9fdc2d721d +dm-tree==0.1.9 +einops==0.8.1 +execnet==2.1.1 +executing==2.2.0 +expecttest==0.3.0 +fasteners @ file:///rapids/fasteners-0.19-py3-none-any.whl#sha256=758819cb5d94cdedf4e836988b74de396ceacb8e2794d21f82d131fd9ee77237 +fastjsonschema==2.21.1 +fastrlock @ file:///rapids/fastrlock-0.8.3-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_28_x86_64.whl#sha256=5f13ec08f1adb1aa916c384b05ecb7dbebb8df9ea81abd045f60941c6283a670 +filelock==3.18.0 +flash_attn @ file:///opt/pytorch/flash_attn-2.7.4.post1-cp312-cp312-linux_x86_64.whl#sha256=a10455d6c56b387785b7413fd0bbcb392b42b3d4f4b88fda5d86cc8a9ce9182a +fonttools==4.58.1 +fqdn==1.5.1 +frozenlist @ file:///rapids/frozenlist-1.6.0-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=4e1be9111cb6756868ac242b3c2bd1f09d9aea09846e4f5c23715e7afb647103 +fsspec==2025.5.1 +gast==0.6.0 +grpcio==1.62.1 +h11==0.16.0 +hf-xet==1.6.0 +httpcore==1.0.9 +httpx==0.28.1 +huggingface_hub==1.29.0 +hypothesis==6.130.8 +idna==3.10 +importlib_metadata==8.7.0 +iniconfig==2.1.0 +intel-openmp==2021.4.0 +ipykernel==6.29.5 +ipython==9.3.0 +ipython_pygments_lexers==1.1.1 +isoduration==20.11.0 +isort==6.0.1 +jedi==0.19.2 +Jinja2==3.1.6 +joblib==1.5.1 +json5==0.12.0 +jsonpointer==3.0.0 +jsonschema==4.24.0 +jsonschema-specifications==2025.4.1 +jupyter-events==0.12.0 +jupyter-lsp==2.2.5 +jupyter_client==8.6.3 +jupyter_core==5.8.1 +jupyter_server==2.16.0 +jupyter_server_terminals==0.5.3 +jupyterlab==4.4.3 +jupyterlab_code_formatter==3.0.2 +jupyterlab_pygments==0.3.0 +jupyterlab_server==2.27.3 +jupyterlab_tensorboard_pro @ file:///builder/jupyterlab_tensorboard_pro-4.0.0-py2.py3-none-any.whl#sha256=6bfcb5ced3358eb72150f4b2ab91312faef2f0ca14d845e8379d8fa52df6902e +jupytext==1.17.2 +kiwisolver==1.4.8 +kvikio @ file:///rapids/kvikio-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=8b40fd8ea9e4a2974c04b50b7fc26b06094ac807f9a233281d0bafbaa71b1d85 +langcodes==3.5.0 +language_data==1.3.0 +lazy_loader==0.4 +libcudf @ file:///rapids/libcudf-25.4.0-py3-none-linux_x86_64.whl#sha256=238320b4958c1737bba2a609673e2539a78216b811b5498a68e00395d9b021a4 +libcugraph @ file:///rapids/libcugraph-25.4.0-py3-none-linux_x86_64.whl#sha256=29085e901b2f7569ff2653bb1355fe6fad20f2fcaa585f5649da9867d8df766d +libcuml @ file:///rapids/libcuml-25.4.0-py3-none-linux_x86_64.whl#sha256=8e2e444c72b0e995e15a48da37ff12fb410a47a2baa69ba86ccc9b1a3db91e8c +libcuvs @ file:///rapids/libcuvs-25.4.0-py3-none-linux_x86_64.whl#sha256=e2f8db61c92b769c590dd752f43439ddd62b2f6592a3c337362b6c1ef220c3b9 +libkvikio @ file:///rapids/libkvikio-25.4.0-py3-none-linux_x86_64.whl#sha256=da8a5713b05fc478bf40529a3e0fb61a2d605bb5a70ca4d461f96a6005dbe8b7 +libraft @ file:///rapids/libraft-25.4.0-py3-none-linux_x86_64.whl#sha256=b921dec005aa4a321dd156776a18376feb47c96a5b67eec125c8a8e19edd1e24 +librmm @ file:///rapids/librmm-25.4.0-py3-none-linux_x86_64.whl#sha256=ff951811bda729d4fa1109ab906225ae1f489a6ec2eeab5d1b674cc34bd4422f +librmm-cu12 @ file:///rapids/librmm_cu12-25.4.0-py3-none-any.whl#sha256=29296ef2b0e90679a0db280e86015be12d6ab9f4e62a181aa085f62dbadd06b3 +librosa==0.11.0 +libucxx @ file:///rapids/libucxx-0.43.0-py3-none-linux_x86_64.whl#sha256=f15849d7ac2fb68f0fd8998838a13321ddcfb0c594f7e6cbc401a3bcc4750566 +lightning-thunder @ file:///opt/pytorch/lightning-thunder +lightning-utilities==0.14.3 +lintrunner==0.12.7 +llvmlite==0.42.0 +locket @ file:///rapids/locket-1.0.0-py2.py3-none-any.whl#sha256=b6c819a722f7b6bd955b80781788e4a66a55628b858d347536b7e81325a3a5e3 +looseversion==1.3.0 +marisa-trie==1.2.1 +Markdown==3.8 +markdown-it-py==3.0.0 +MarkupSafe==3.0.2 +matplotlib==3.10.3 +matplotlib-inline==0.1.7 +mdit-py-plugins==0.4.2 +mdurl==0.1.2 +-e git+ssh://git@github.com/sriram-17-17/my-Megatron-MoE-Echo.git@9a23679390d7ab1fe462bc61fa31d90d393db49c#egg=megatron_core +mistune==3.1.3 +mkl==2021.1.1 +mkl-devel==2021.1.1 +mkl-include==2021.1.1 +ml_dtypes==0.6.0 +mock==5.2.0 +-e git+ssh://git@github.com/sriram-17-17/MoonEP.git@d8bc6667fad7177e6e09b637d773fa67f08bae93#egg=moonep +mpmath==1.3.0 +msgpack==1.1.0 +multidict @ file:///rapids/multidict-6.4.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=a920f9cf2abdf6e493c519492d892c362007f113c94da4c239ae88429835bad1 +multiprocess==0.70.19 +murmurhash==1.0.13 +mypy_extensions==1.1.0 +nbclient==0.10.2 +nbconvert==7.16.6 +nbformat==5.10.4 +nest-asyncio==1.6.0 +networkx==3.5 +ninja==1.11.1.4 +notebook==7.4.3 +notebook_shim==0.2.4 +numba==0.59.1 +numba-cuda @ file:///rapids/numba_cuda-0.4.0-py3-none-any.whl#sha256=5b95a5f42068771e37142944d44fe844fa87f1b2f89051461a4cb0af91bcc28a +numcodecs @ file:///rapids/numcodecs-0.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=da2230484e6102e5fa3cc1a5dd37ca1f92dfbd183d91662074d6f7574e3e8f53 +numpy==1.26.4 +nvdlfw_inspect @ git+https://github.com/NVIDIA/nvidia-dlfw-inspect.git@6b60eb0e675606fb2fbbfcb12667ecbec75eaf87 +nvfuser==0.2.27a0+9bf5aca +nvidia-cuda-cccl-cu12==12.9.27 +nvidia-cudnn-frontend @ file:///opt/pytorch/lightning-thunder/tmp/cudnn_frontend +nvidia-cutlass-dsl==4.4.2 +nvidia-cutlass-dsl-libs-base==4.4.2 +nvidia-dali-cuda120==1.50.0 +nvidia-ml-py @ file:///rapids/nvidia_ml_py-12.575.51-py3-none-any.whl#sha256=eb8641800d98ce40a22f479873f34b482e214a7e80349c63be51c3919845446e +nvidia-modelopt==0.29.0 +nvidia-modelopt-core==0.29.0 +nvidia-nvcomp-cu12==4.2.0.14 +nvidia-nvimgcodec-cu12==0.5.0.13 +nvidia-nvjpeg-cu12==12.4.0.16 +nvidia-nvjpeg2k-cu12==0.8.1.40 +nvidia-nvshmem-cu12==3.7.2 +nvidia-nvtiff-cu12==0.5.0.67 +nvidia-resiliency-ext==0.4.0 +nvtx @ file:///rapids/nvtx-0.2.11-cp312-cp312-linux_x86_64.whl#sha256=73521e37cba42b4db0b3cd68fc5eb814e55c500a61eeed6edcdf6746315ac53e +nx-cugraph @ file:///rapids/nx_cugraph-25.4.0-py3-none-any.whl#sha256=e625444486cc899a77bd48e7da76a2edd95c0006dd95fbf323745feae7f33111 +onnx @ file:///opt/pytorch/pytorch/third_party/onnx +onnx-ir==1.0.0 +onnxscript==0.7.1 +opt_einsum==3.4.0 +optree==0.16.0 +overrides==7.7.0 +packaging==23.2 +pandas @ file:///rapids/pandas-2.2.3-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=fffb8ae78d8af97f849404f21411c95062db1496aeb3e56f146f0355c9989319 +pandocfilters==1.5.1 +parso==0.8.4 +partd @ file:///rapids/partd-1.4.2-py3-none-any.whl#sha256=978e4ac767ec4ba5b86c6eaa52e5a2a3bc748a2ca839e8cc798f1cc6ce6efb0f +pathspec==0.12.1 +pexpect==4.9.0 +pillow==11.2.1 +platformdirs==4.3.8 +pluggy==1.6.0 +ply @ file:///rapids/ply-3.11-py2.py3-none-any.whl#sha256=096f9b8350b65ebd2fd1346b12452efe5b9607f7482813ffca50c22722a807ce +polars @ file:///rapids/polars-1.25.2-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=f7fcbb4f476784384ccda48757fca4e8c2e2c5a0a3aef3717aaf56aee4e30e09 +polygraphy==0.49.20 +pooch==1.8.2 +preshed==3.0.10 +prometheus_client==0.22.1 +prompt_toolkit==3.0.51 +propcache @ file:///rapids/propcache-0.3.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=563f9d8c03ad645597b8d010ef4e9eab359faeb11a0a2ac9f7b4bc8c28ebef25 +protobuf==4.24.4 +psutil==7.0.0 +ptyprocess==0.7.0 +PuLP==3.2.1 +pure_eval==0.2.3 +pyarrow==25.0.1 +pybind11==2.13.6 +pybind11_global==2.13.6 +pycocotools @ git+https://github.com/nvidia/cocoapi.git@08350a157b41f0755a536441b5abec89ab934c5b#subdirectory=PythonAPI +pycparser==2.22 +pydantic==2.11.5 +pydantic_core==2.33.2 +Pygments==2.19.1 +pylibcudf @ file:///rapids/pylibcudf-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=401b8bfd022d06f84e120779ca7c3d01a28b68762aaa8f7873b10d6b1975cf6a +pylibcugraph @ file:///rapids/pylibcugraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=061c7389a88c815b54fd5d3a3584fb87bf693e5602f5be335260d0ba575ba557 +pylibcugraphops @ file:///rapids/pylibcugraphops-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=a25ef4b22166c80718d5c7e5bddb755e81a2fe340fa5becd0ac36ec3edfb208a +pylibraft @ file:///rapids/pylibraft-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=f2ecd23fe51f9c2caa87bbbe85cb32fcaff4ecfb52ceb9a9991bc50f82782c2e +pylibwholegraph @ file:///rapids/pylibwholegraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b82913d1e38bda76d8c272fbd8cc9072d987209d3d47cc93a464a37c2925aca2 +pynvjitlink @ file:///rapids/pynvjitlink-0.3.0-cp312-cp312-linux_x86_64.whl#sha256=066caf1314730643d3b48c9847bd65789defdd6a7c78993176eb3d8bbe3f4919 +pynvml @ file:///rapids/pynvml-12.0.0-py3-none-any.whl#sha256=fdff84b62a27dbe98e08e1a647eb77342bef1aebe0878bcd15e99a83fcbecb9e +pyparsing==3.2.3 +pytest==8.1.1 +pytest-flakefinder==1.1.0 +pytest-rerunfailures==15.1 +pytest-shard==0.1.2 +pytest-xdist==3.7.0 +python-dateutil==2.9.0.post0 +python-hostlist==2.2.1 +python-json-logger==3.3.0 +pytorch-triton @ file:///tmp/dist/pytorch_triton-3.3.0%2Bgit96316ce52.nvinternal-cp312-cp312-linux_x86_64.whl#sha256=6d6921772b9dfd0243c055e0d054de42a113e92cabcfdd1d3e80a5ad9c08cf7d +pytz @ file:///rapids/pytz-2023.4-py2.py3-none-any.whl#sha256=f90ef520d95e7c46951105338d918664ebfd6f1d995bd7d153127ce90efafa6a +PyYAML==6.0.2 +pyzmq==26.4.0 +raft-dask @ file:///rapids/raft_dask-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=36cf37ba6ce35454c4729c2cb597a360f22780eefb11d514a16a07137da5eb6c +rapids-dask-dependency @ file:///rapids/rapids_dask_dependency-25.4.0a0-py3-none-any.whl#sha256=484faca626e7760112a7d8ad3426f5bdd720b609006d8dca87c2607076a348bd +rapids-logger @ file:///rapids/rapids_logger-0.1.18-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl#sha256=3617b2d86291ac6cb957542895f20fc9737ace29773ca90337e6fb04c73bd120 +referencing==0.36.2 +regex==2024.11.6 +requests==2.32.3 +rfc3339-validator==0.1.4 +rfc3986-validator==0.1.1 +rich==14.0.0 +rmm @ file:///rapids/rmm-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=1824bcab3836296666980a997ddcb9c4bca24dfb873012cfa0a7ecea583ca712 +rpds-py==0.25.1 +safetensors==0.5.3 +scikit-learn==1.6.1 +scipy==1.15.3 +Send2Trash==1.8.3 +setuptools==78.1.1 +shellingham==1.5.4 +six==1.16.0 +smart-open==7.1.0 +sniffio==1.3.1 +sortedcontainers==2.4.0 +soundfile==0.13.1 +soupsieve==2.7 +soxr==0.5.0.post1 +spacy==3.7.5 +spacy-legacy==3.0.12 +spacy-loggers==1.0.5 +srsly==2.5.1 +stack-data==0.6.3 +sympy==1.14.0 +tabulate==0.9.0 +tbb==2021.13.1 +tblib @ file:///rapids/tblib-3.1.0-py3-none-any.whl#sha256=670bb4582578134b3d81a84afa1b016128b429f3d48e6cbbaecc9d15675e984e +tensorboard==2.16.2 +tensorboard-data-server==0.7.2 +tensorrt @ file:///workspace/TensorRT-10.11.0.33/python/tensorrt-10.11.0.33-cp312-none-linux_x86_64.whl#sha256=c2f18f72758a3d54a71901495c8fda9e7094a8bcdd2a5b6b58967fbf05733b3a +terminado==0.18.1 +thinc==8.2.5 +threadpoolctl==3.6.0 +thriftpy2 @ file:///rapids/thriftpy2-0.5.2-cp312-cp312-linux_x86_64.whl#sha256=c0c7d939bdc5ccf7739154959938445d5fdf72f1d6e12bd563947a36f67e2147 +tinycss2==1.4.0 +toolz @ file:///rapids/toolz-1.0.0-py3-none-any.whl#sha256=292c8f1c4e7516bf9086f8850935c799a874039c8bcf959d47b600e4c44a6236 +torch @ file:///opt/transfer/torch-2.8.0a0%2B5228986c39.nv25.6-29946474-cp312-cp312-linux_x86_64.whl#sha256=7a3c9ff5288d8102a31296b3d62491f3f47c8820a5c9aaefc7dc0fe20e1e8ad5 +torch_tensorrt @ file:///opt/pytorch/torch_tensorrt/dist/torch_tensorrt-2.8.0a0-cp312-cp312-linux_x86_64.whl#sha256=26f4d7a04e7e6e017a5b917b8b98b83720bb2bb0195f48d25f52d567d286bfe1 +torchao @ file:///opt/pytorch/ao +torchprofile==0.0.4 +torchvision @ file:///opt/transfer/torchvision-0.22.0a0%2B95f10a4e-29946474-cp312-cp312-manylinux_2_28_x86_64.whl#sha256=746a3c674cc61680a86f4a14b374b21af1b5c21f629c42765b40c9494f6b234d +tornado==6.5.1 +tqdm==4.67.1 +traitlets==5.14.3 +transformer_engine @ file:///pvcdata/aiinfra/b-aghosh/wheels/transformer_engine-2.10.0.dev0%2B10f6128d-cp312-cp312-linux_x86_64.whl#sha256=35cf5f4b83b2baad864e12e512bb422e0929409ac05610812fa521d4942d0189 +treelite @ file:///rapids/treelite-4.4.1-py3-none-linux_x86_64.whl#sha256=174741724205540a1a5b37da81483aeff55d89548f42b3461cf07acac7ae96eb +typer==0.16.0 +types-dataclasses==0.6.6 +types-python-dateutil==2.9.0.20250516 +typing-inspection==0.4.1 +typing_extensions==4.14.0 +tzdata @ file:///rapids/tzdata-2025.2-py2.py3-none-any.whl#sha256=1a403fada01ff9221ca8044d701868fa132215d84beb92242d9acd2147f667a8 +ucx-py @ file:///rapids/ucx_py-0.43.0-cp312-cp312-linux_x86_64.whl#sha256=4a452fd8d8c983283ada3d49086e3657dc2ee94fbad08efdd708d15ab6e0b73a +ucxx @ file:///rapids/ucxx-0.43.0-cp312-cp312-linux_x86_64.whl#sha256=f1b4e55cb45c66fe88a2d8f9c2b87751165f3cfec21fbdc9c47c4ddff75e8a5f +-e git+https://github.com/Dots-Infra/UltraEP.git@94cab099b44fffa99a82fea99e7c12d89cf65e4f#egg=ultra_ep +uri-template==1.3.0 +urllib3==2.0.7 +wasabi==1.1.3 +wcwidth==0.2.13 +weasel==0.4.1 +webcolors==24.11.1 +webencodings==0.5.1 +websocket-client==1.8.0 +Werkzeug==3.1.3 +wheel==0.45.1 +wrapt==1.17.2 +xdoctest==1.0.2 +xgboost @ file:///rapids/xgboost-2.1.4-py3-none-manylinux_2_39_x86_64.whl#sha256=25eeca27f4048fc491940b3a0b4b5782ee48c33cb3aec059cef71f8da542930b +xxhash==4.0.1 +yarl @ file:///rapids/yarl-1.20.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=bf099e2432131093cc611623e0b0bcc399b8cddd9a91eded8bfb50402ec35018 +zarr @ file:///rapids/zarr-2.18.7-py3-none-any.whl#sha256=ac3dc4033e9ae4e9d7b5e27c97ea3eaf1003cc0a07f010bd83d5134bf8c4b223 +zict @ file:///rapids/zict-3.0.0-py2.py3-none-any.whl#sha256=5796e36bd0e0cc8cf0fbc1ace6a68912611c1dbd74750a3f3026b9b9d6a327ae +zipp==3.22.0 diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/provenance.txt b/tierb-fixed-moonep-full/echo/tierbtrace/provenance.txt index b9a0351bb07858571557bb69d8b9892775c10933..3e11687821c3676c26b158f50f87c6949d1d26c6 100644 --- a/tierb-fixed-moonep-full/echo/tierbtrace/provenance.txt +++ b/tierb-fixed-moonep-full/echo/tierbtrace/provenance.txt @@ -27,3 +27,64 @@ command : run_ladder.sh --phase tierbtrace --gbs 8 --out /workspace/tierb- 7, 120 MHz, 1965 MHz, 1000.00 W, 28 NVLink topology: + GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 NIC8 CPU Affinity NUMA Affinity GPU NUMA ID + GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 PIX NODE SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A + GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 NODE PIX SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A + GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 SYS SYS PIX NODE SYS SYS SYS SYS SYS 28-55 1 N/A + GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 SYS SYS NODE PIX SYS SYS SYS SYS SYS 28-55 1 N/A + GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 SYS SYS SYS SYS PIX NODE SYS SYS SYS 56-83 2 N/A + GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 SYS SYS SYS SYS NODE PIX SYS SYS SYS 56-83 2 N/A + GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 SYS SYS SYS SYS SYS SYS PIX NODE SYS 84-111 3 N/A + GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X SYS SYS SYS SYS SYS SYS NODE PIX SYS 84-111 3 N/A + NIC0 PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS SYS NODE + NIC1 NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS SYS NODE + NIC2 SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS + NIC3 SYS SYS NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS + NIC4 SYS SYS SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS + unavailable + +---- toolchain ----------------------------------------------- + nvcc : release 12.9, V12.9.86 + gcc : gcc (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0 + python : 3.12.3 + torch : 2.8.0a0+5228986c39.nv25.06 cuda=12.9 nccl=2.27.3 + torch arch : ['sm_75', 'sm_80', 'sm_86', 'sm_90', 'sm_100', 'sm_120', 'compute_120'] + transformer_engine: 2.10.0.dev0+10f6128d [/usr/local/lib/python3.12/dist-packages/transformer_engine/__init__.py] + deep_ep : 1.2.1+34152ae [/workspace/DeepEP/deep_ep/__init__.py] + ultra_ep : 1.0.0 [/workspace/UltraEP/ultra_ep/__init__.py] + moonep : 0.0.1 [/workspace/MoonEP/moonep/__init__.py] + megatron : 0.14.0rc5 [None] + flash_attn : 2.7.4.post1 [/usr/local/lib/python3.12/dist-packages/flash_attn/__init__.py] + numpy : 1.26.4 [/usr/local/lib/python3.12/dist-packages/numpy/__init__.py] + huggingface_hub: 1.29.0 [/usr/local/lib/python3.12/dist-packages/huggingface_hub/__init__.py] + +---- repositories -------------------------------------------- + Megatron-MoE-Echo bracing-regions-through-kernels 9a23679390d7 + Megatron-UltraEP ultraep-token-drop-analysis eaf829e7ad21 + UltraEP main 94cab099b44f + DeepEP HEAD 34152ae28f80 (+3 untracked: build output) + MoonEP megatron-ddp-symmetric-params d8bc6667fad7 + blackwell_run_scripts granularity-echo 5054c894a84a + +---- environment --------------------------------------------- + CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0 + CUDA_DRIVER_VERSION=575.57.08 + CUDA_HOME=/usr/local/cuda + CUDA_MODULE_LOADING=LAZY + CUDA_VERSION=12.9.1.010 + HF_REMOTE_ROOT=tierb-fixed-moonep-full + NCCL_VERSION=2.27.3 + NVSHMEM_VERSION=3.2.5 + PYTORCH_BUILD_NUMBER=0 + PYTORCH_BUILD_VERSION=2.8.0a0+5228986 + PYTORCH_HOME=/opt/pytorch/pytorch + PYTORCH_VERSION=2.8.0a0+5228986 + TORCH_ALLOW_TF32_CUBLAS_OVERRIDE=1 + TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0+PTX + TORCH_NCCL_USE_COMM_NONBLOCKING=0 + +---- shared library resolution ------------------------------- + libnvshmem_host.so: not in ldconfig + libnccl.so: /lib/x86_64-linux-gnu/libnccl.so.2 + lrwxrwxrwx 1 root root 20 Aug 29 16:41 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so -> libnvshmem_host.so.3 + -rw-r--r-- 1 root root 179045744 Aug 29 16:37 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so.3 diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep new file mode 100644 index 0000000000000000000000000000000000000000..c478c0848289f16774c03125f5c2c957769bd58c --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab57a4219500aafc846fd89137bedeaf6df3fe39209ce2b8bc6b6d044cfce56d +size 10151124 diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/run.env b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/run.env new file mode 100644 index 0000000000000000000000000000000000000000..c99ac8daca8d307a95314b781e747506ac726e94 --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/run.env @@ -0,0 +1,70 @@ +backend=echo +config=QWEN3_235B +family=tierB +repo=/workspace/Megatron-MoE-Echo +branch=bracing-regions-through-kernels +commit=9a236793 +num_layers=12 +hidden=4096 +ffn_hidden=12288 +moe_ffn_hidden=1536 +num_experts=128 +topk=8 +experts_per_rank=16 +expert_params=18874368 +heads=64 +kv_groups=4 +kv_channels=128 +qk_layernorm=1 +shared_ffn=0 +shared_expert_overlap=0 +moe_token_dispatcher=flex +mla=0 +dense_first_layers=0 +moe_layers=12 +rotary_base=1000000 +rotary_scaling= +qkv_bias=0 +router_balancing=aux_loss +aux_loss_coeff=1e-3 +router_pre_softmax=0 +router_fusion=0 +router_dtype=fp32 +device_initiated_grouped_gemm=1 +norm_epsilon=1e-6 +seq_length=8192 +micro_batch_size=1 +global_batch_size=8 +num_microbatches=1 +assignments_per_rank=65536 +mean_tokens_per_expert=4096 +slot_policy=matched +slots_per_rank=16 +capacity_mode=static +capacity=1.083984375 +cuda_graph=1 +probe_mode=0 +async_grad_reduce=0 +branch_role=timing +memory_dump_mode=none +memory_snapshot_path= +cuda_graph_warmup_steps=3 +recompute=none +force_balance=0 +train_iters=10 +ep_size=8 +echo_dispatch_overlap=1 +echo_dispatch_recompute=0 +ultraep_quota_min_tokens= +ultraep_weight_sync_join= +profile_enable=1 +kernel_trace=1 +fp8_mode=mxfp8 +fp8_recipe=mxfp8 +training_seed=1234 +learning_rate=1e-4 +min_learning_rate=1.0e-5 +lr_warmup_iters=5 +state_dump_enable=0 +state_dump_dir= +moonep_assert_no_pending_reduce=0 diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/train.log b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/train.log new file mode 100644 index 0000000000000000000000000000000000000000..85d03a354828b22bff93bcbd505d758559f7f5ae --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/train.log @@ -0,0 +1,1452 @@ +[00:14:44] WARN: bracing-regions-through-kernels lacks --moe-router-fusion; QWEN3_235B runs unfused (timing only, routing identical) +============================================================== + QWEN3_235B / echo (slots/rank=16 policy=matched) + E=128 K=8 H'=1536 m=16 + L=12 seq=8192 gbs=8 (1 microbatches) + cap_mode=static cap=1.083984375 cg=1 recompute=none + memory_dump=none snapshots=/workspace/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/memory_snapshots/snapshot.pickle + repo=/workspace/Megatron-MoE-Echo @ 9a236793 +============================================================== +W0830 00:14:46.033000 291145 torch/distributed/run.py:766] +W0830 00:14:46.033000 291145 torch/distributed/run.py:766] ***************************************** +W0830 00:14:46.033000 291145 torch/distributed/run.py:766] Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. +W0830 00:14:46.033000 291145 torch/distributed/run.py:766] ***************************************** +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +using world size: 8, data-parallel size: 8, context-parallel size: 1, hierarchical context-parallel sizes: None, tensor-model-parallel size: 1, pipeline-model-parallel size: 1 +WARNING: overriding default arguments for tokenizer_type:GPT2BPETokenizer with tokenizer_type:GPT2BPETokenizer +Number of virtual stages per pipeline stage: None +accumulate and all-reduce gradients in fp32 for bfloat16 data type. +using torch.bfloat16 for parameters ... +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +------------------------ arguments ------------------------ + account_for_embedding_in_pipeline_split ......... False + account_for_loss_in_pipeline_split .............. False + accumulate_allreduce_grads_in_fp32 .............. True + adam_beta1 ...................................... 0.9 + adam_beta2 ...................................... 0.999 + adam_eps ........................................ 1e-08 + add_bias_linear ................................. False + add_position_embedding .......................... False + add_qkv_bias .................................... False + adlr_autoresume ................................. False + adlr_autoresume_interval ........................ 1000 + align_grad_reduce ............................... True + align_param_gather .............................. False + app_tag_run_name ................................ None + app_tag_run_version ............................. 0.0.0 + apply_layernorm_1p .............................. False + apply_query_key_layer_scaling ................... False + apply_residual_connection_post_layernorm ........ False + apply_rope_fusion ............................... True + async_save ...................................... None + async_tensor_model_parallel_allreduce ........... True + attention_backend ............................... AttnBackend.auto + attention_dropout ............................... 0.0 + attention_softmax_in_fp32 ....................... False + auto_detect_ckpt_format ......................... False + barrier_with_L1_time ............................ True + bert_binary_head ................................ True + bert_embedder_type .............................. megatron + bert_load ....................................... None + bf16 ............................................ True + bias_dropout_fusion ............................. True + bias_gelu_fusion ................................ False + bias_swiglu_fusion .............................. True + biencoder_projection_dim ........................ 0 + biencoder_shared_query_context_model ............ False + block_data_path ................................. None + calc_ft_timeouts ................................ False + calculate_per_token_loss ........................ False + check_for_large_grads ........................... False + check_for_nan_in_loss_and_grad .................. False + check_for_spiky_loss ............................ False + check_weight_hash_across_dp_replicas_interval ... None + ckpt_assume_constant_structure .................. False + ckpt_convert_format ............................. None + ckpt_convert_save ............................... None + ckpt_convert_update_legacy_dist_opt_format ...... False + ckpt_format ..................................... torch_dist + ckpt_fully_parallel_load ........................ False + ckpt_fully_parallel_save ........................ True + ckpt_fully_parallel_save_deprecated ............. False + ckpt_step ....................................... None + classes_fraction ................................ 1.0 + clip_grad ....................................... 1.0 + clone_scatter_output_in_embedding ............... True + config_logger_dir ............................... + consumed_train_samples .......................... 0 + consumed_valid_samples .......................... 0 + context_parallel_size ........................... 1 + cp_comm_type .................................... ['p2p'] + create_attention_mask_in_dataloader ............. True + cross_entropy_fusion_impl ....................... native + cross_entropy_loss_fusion ....................... False + cuda_graph_scope ................................ full_iteration + cuda_graph_warmup_steps ......................... 3 + data_args_path .................................. None + data_cache_path ................................. None + data_parallel_random_init ....................... False + data_parallel_sharding_strategy ................. no_shard + data_parallel_size .............................. 8 + data_path ....................................... ['/workspace/data/wikitext/wikitext103_gpt2_text_document'] + data_per_class_fraction ......................... 1.0 + data_sharding ................................... True + dataloader_type ................................. single + ddp_average_in_collective ....................... False + ddp_bucket_size ................................. None + ddp_num_buckets ................................. None + ddp_pad_buckets_for_high_nccl_busbw ............. False + decoder_first_pipeline_num_layers ............... None + decoder_last_pipeline_num_layers ................ None + decoder_num_layers .............................. None + decoder_seq_length .............................. None + decoupled_lr .................................... None + decoupled_min_lr ................................ None + decrease_batch_size_if_needed ................... False + defer_embedding_wgrad_compute ................... False + delay_wgrad_compute ............................. False + deprecated_use_mcore_models ..................... True + deterministic_mode .............................. False + dino_bottleneck_size ............................ 256 + dino_freeze_last_layer .......................... 1 + dino_head_hidden_size ........................... 2048 + dino_local_crops_number ......................... 10 + dino_local_img_size ............................. 96 + dino_norm_last_layer ............................ False + dino_teacher_temp ............................... 0.07 + dino_warmup_teacher_temp ........................ 0.04 + dino_warmup_teacher_temp_epochs ................. 30 + disable_bf16_reduced_precision_matmul ........... False + disable_mamba_mem_eff_path ...................... False + disable_straggler_on_startup .................... False + dist_ckpt_format_deprecated ..................... None + dist_ckpt_strictness ............................ assume_ok_unexpected + distribute_saved_activations .................... False + distributed_backend ............................. nccl + distributed_timeout_minutes ..................... 10 + embedding_init_method_std ....................... None + embedding_path .................................. None + empty_unused_memory_level ....................... 0 + enable_cuda_graph ............................... True + enable_experimental ............................. False + enable_ft_package ............................... False + enable_gloo_process_groups ...................... True + enable_msc ...................................... True + enable_one_logger ............................... True + encoder_num_layers .............................. 12 + encoder_seq_length .............................. 8192 + end_weight_decay ................................ 0.1 + eod_mask_loss ................................... False + error_injection_rate ............................ 0 + error_injection_type ............................ transient_error + eval_interval ................................... 1000000 + eval_iters ...................................... 0 + evidence_data_path .............................. None + exit_duration_in_mins ........................... None + exit_interval ................................... None + exit_on_missing_checkpoint ...................... False + exit_signal_handler ............................. False + exp_avg_dtype ................................... torch.float32 + exp_avg_sq_dtype ................................ torch.float32 + expert_model_parallel_size ...................... 8 + expert_tensor_parallel_size ..................... 1 + export_force_local_attention .................... False + export_kd_cfg ................................... None + export_kd_teacher_ckpt_format ................... None + export_kd_teacher_load .......................... None + export_kv_cache_quant ........................... False + export_legacy_megatron .......................... False + export_model_type ............................... GPTModel + export_moe_apply_probs_on_input ................. False + export_qk_l2_norm ............................... False + export_quant_cfg ................................ None + export_real_quant_cfg ........................... None + export_te_mcore_model ........................... False + external_cuda_graph ............................. False + ffn_hidden_size ................................. 12288 + finetune ........................................ False + finetune_data_split ............................. train + finetune_hf_dataset ............................. None + first_last_layers_bf16 .......................... False + flash_decode .................................... False + fp16 ............................................ False + fp16_lm_cross_entropy ........................... False + fp32_residual_connection ........................ False + fp8 ............................................. e4m3 + fp8_amax_compute_algo ........................... most_recent + fp8_amax_history_len ............................ 1 + fp8_interval .................................... 1 + fp8_margin ...................................... 0 + fp8_param_gather ................................ False + fp8_recipe ...................................... mxfp8 + fp8_wgrad ....................................... True + fsdp_double_buffer .............................. False + global_batch_size ............................... 8 + grad_reduce_in_bf16 ............................. False + gradient_accumulation_fusion .................... True + gradient_reduce_div_fusion ...................... True + group_query_attention ........................... True + head_lr_mult .................................... 1.0 + heterogeneous_layers_config_encoded_json ........ None + heterogeneous_layers_config_path ................ None + hidden_dropout .................................. 0.0 + hidden_size ..................................... 4096 + hierarchical_context_parallel_sizes ............. None + high_priority_stream_groups ..................... [] + hybrid_attention_ratio .......................... 0.0 + hybrid_mlp_ratio ................................ 0.0 + hybrid_override_pattern ......................... None + hysteresis ...................................... 2 + ict_head_size ................................... None + ict_load ........................................ None + img_h ........................................... 224 + img_w ........................................... 224 + indexer_batch_size .............................. 128 + indexer_log_interval ............................ 1000 + inference_batch_times_seqlen_threshold .......... -1 + inference_dynamic_batching ...................... False + inference_dynamic_batching_buffer_guaranteed_fraction 0.2 + inference_dynamic_batching_buffer_overflow_factor None + inference_dynamic_batching_buffer_size_gb ....... 40.0 + inference_dynamic_batching_chunk_size ........... 256 + inference_dynamic_batching_max_requests_override None + inference_dynamic_batching_max_tokens_override .. None + inference_dynamic_batching_num_cuda_graphs ...... 16 + inference_max_batch_size ........................ 8 + inference_max_seq_length ........................ 2560 + inference_rng_tracker ........................... False + init_method_std ................................. 0.01 + init_method_xavier_uniform ...................... False + init_model_with_meta_device ..................... False + initial_loss_scale .............................. 4294967296 + inprocess_active_world_size ..................... 8 + inprocess_barrier_timeout ....................... 120 + inprocess_completion_timeout .................... 120 + inprocess_empty_cuda_cache ...................... False + inprocess_granularity ........................... node + inprocess_hard_timeout .......................... 90 + inprocess_heartbeat_interval .................... 30 + inprocess_heartbeat_timeout ..................... 60 + inprocess_last_call_wait ........................ 1 + inprocess_max_iterations ........................ None + inprocess_monitor_process_interval .............. 1.0 + inprocess_monitor_thread_interval ............... 1.0 + inprocess_progress_watchdog_interval ............ 1.0 + inprocess_restart ............................... False + inprocess_soft_timeout .......................... 60 + inprocess_termination_grace_time ................ 1 + is_hybrid_model ................................. False + iter_per_epoch .................................. 1250 + iterations_to_skip .............................. [] + keep_fp8_transpose_cache_when_using_custom_fsdp . False + kitchen_config_file ............................. None + kitchen_recipe_number ........................... None + kv_channels ..................................... 128 + kv_lora_rank .................................... 32 + lazy_mpu_init ................................... None + load ............................................ None + load_main_params_from_ckpt ...................... None + load_model_opt_format ........................... False + local_rank ...................................... 0 + log_energy ...................................... False + log_interval .................................... 1 + log_loss_scale_to_tensorboard ................... True + log_memory_to_tensorboard ....................... False + log_num_zeros_in_grad ........................... False + log_params_norm ................................. False + log_progress .................................... False + log_straggler ................................... False + log_throughput .................................. True + log_timers_to_tensorboard ....................... False + log_validation_ppl_to_tensorboard ............... False + log_world_size_to_tensorboard ................... False + logging_level ................................... None + loss_scale ...................................... None + loss_scale_window ............................... 1000 + lr .............................................. 0.0001 + lr_decay_iters .................................. 8 + lr_decay_samples ................................ None + lr_decay_style .................................. cosine + lr_warmup_fraction .............................. None + lr_warmup_init .................................. 0.0 + lr_warmup_iters ................................. 5 + lr_warmup_samples ............................... 0 + lr_wsd_decay_iters .............................. None + lr_wsd_decay_samples ............................ None + lr_wsd_decay_style .............................. exponential + main_grads_dtype ................................ torch.float32 + main_params_dtype ............................... torch.float32 + make_vocab_size_divisible_by .................... 128 + mamba_head_dim .................................. 64 + mamba_num_groups ................................ 8 + mamba_num_heads ................................. None + mamba_state_dim ................................. 128 + manual_gc ....................................... False + manual_gc_eval .................................. True + manual_gc_interval .............................. 0 + mask_factor ..................................... 1.0 + mask_prob ....................................... 0.15 + mask_type ....................................... random + masked_softmax_fusion ........................... False + max_position_embeddings ......................... 32768 + max_tokens_to_oom ............................... 12000 + memory_snapshot_path ............................ snapshot.pickle + merge_file ...................................... /workspace/data/wikitext/gpt2-merges.txt + micro_batch_size ................................ 1 + microbatch_group_size_per_vp_stage .............. None + mid_level_dataset_surplus ....................... 0.005 + min_loss_scale .................................. 1.0 + min_lr .......................................... 1e-05 + mlp_chunks_for_prefill .......................... 1 + mmap_bin_files .................................. True + mock_data ....................................... False + moe_apply_probs_on_input ........................ False + moe_aux_loss_coeff .............................. 0.001 + moe_deepep_num_sms .............................. 20 + moe_echo_enable_random_offloading ............... False + moe_echo_expert_dispatch_overlap ................ True + moe_echo_expert_dispatcher_type ................. hybridep + moe_echo_recompute_expert_dispatch .............. False + moe_enable_deepep ............................... False + moe_enable_echo ................................. True + moe_enable_hybridep ............................. True + moe_expert_capacity_factor ...................... None + moe_extended_tp ................................. False + moe_ffn_hidden_size ............................. 1536 + moe_grouped_gemm ................................ True + moe_hybridep_num_sms ............................ 16 + moe_input_jitter_eps ............................ None + moe_layer_freq .................................. 1 + moe_layer_recompute ............................. False + moe_num_echo_experts ............................ 128 + moe_pad_expert_input_to_capacity ................ False + moe_per_layer_logging ........................... False + moe_permute_fusion .............................. False + moe_received_token_capacity ..................... 1.083984375 + moe_router_bias_update_rate ..................... 0.001 + moe_router_dtype ................................ fp32 + moe_router_enable_expert_bias ................... False + moe_router_force_load_balancing ................. False + moe_router_group_topk ........................... None + moe_router_load_balancing_type .................. aux_loss + moe_router_num_groups ........................... None + moe_router_padding_for_fp8 ...................... False + moe_router_pre_softmax .......................... False + moe_router_score_function ....................... softmax + moe_router_topk ................................. 8 + moe_router_topk_scaling_factor .................. None + moe_shared_expert_intermediate_size ............. None + moe_shared_expert_overlap ....................... False + moe_token_dispatcher_type ....................... flex + moe_token_drop_policy ........................... probs + moe_upcycling_granularity ....................... 1 + moe_use_device_initiated_grouped_gemm ........... True + moe_use_legacy_grouped_gemm ..................... False + moe_use_upcycling ............................... False + moe_z_loss_coeff ................................ None + mrope_section ................................... None + mscale .......................................... 1.0 + mscale_all_dim .................................. 0.0 + mtp_loss_scaling_factor ......................... 0.1 + mtp_num_layers .................................. None + multi_latent_attention .......................... False + nccl_all_reduce_for_prefill ..................... False + nccl_communicator_config_path ................... None + nccl_ub ......................................... False + no_load_optim ................................... None + no_load_rng ..................................... None + no_persist_layer_norm ........................... False + no_rope_freq .................................... None + no_save_optim ................................... None + no_save_rng ..................................... None + non_persistent_ckpt_type ........................ None + non_persistent_global_ckpt_dir .................. None + non_persistent_local_ckpt_algo .................. fully_parallel + non_persistent_local_ckpt_dir ................... None + non_persistent_save_interval .................... None + norm_epsilon .................................... 1e-06 + normalization ................................... RMSNorm + num_attention_heads ............................. 64 + num_channels .................................... 3 + num_classes ..................................... 1000 + num_dataset_builder_threads ..................... 1 + num_distributed_optimizer_instances ............. 1 + num_experts ..................................... 128 + num_layers ...................................... 12 + num_layers_at_end_in_bf16 ....................... 1 + num_layers_at_start_in_bf16 ..................... 1 + num_layers_per_virtual_pipeline_stage ........... None + num_query_groups ................................ 4 + num_virtual_stages_per_pipeline_rank ............ None + num_workers ..................................... 2 + object_storage_cache_path ....................... None + one_logger_async ................................ False + one_logger_project .............................. megatron-lm + one_logger_run_name ............................. None + onnx_safe ....................................... None + openai_gelu ..................................... False + optimizer ....................................... adam + optimizer_cpu_offload ........................... False + optimizer_offload_fraction ...................... 1.0 + output_bert_embeddings .......................... False + overlap_cpu_optimizer_d2h_h2d ................... False + overlap_grad_reduce ............................. True + overlap_moe_expert_parallel_comm ................ False + overlap_p2p_comm ................................ False + overlap_p2p_comm_warmup_flush ................... False + overlap_param_gather ............................ True + overlap_param_gather_with_optimizer_step ........ False + override_opt_param_scheduler .................... False + params_dtype .................................... torch.bfloat16 + patch_dim ....................................... 16 + per_split_data_args_path ........................ None + perform_initialization .......................... True + pin_cpu_grads ................................... True + pin_cpu_params .................................. True + pipeline_model_parallel_comm_backend ............ None + pipeline_model_parallel_layout .................. None + pipeline_model_parallel_size .................... 1 + position_embedding_type ......................... rope + pretrained_checkpoint ........................... None + profile ......................................... True + profile_ranks ................................... [0, 1, 2, 3, 4, 5, 6, 7] + profile_step_end ................................ 8 + profile_step_start .............................. 6 + q_lora_rank ..................................... None + qk_head_dim ..................................... 128 + qk_l2_norm ...................................... False + qk_layernorm .................................... True + qk_pos_emb_head_dim ............................. 64 + query_in_block_prob ............................. 0.1 + rampup_batch_size ............................... None + rank ............................................ 0 + recompute_granularity ........................... None + recompute_method ................................ None + recompute_modules ............................... None + recompute_num_layers ............................ None + record_memory_history ........................... False + relative_attention_max_distance ................. 128 + relative_attention_num_buckets .................. 32 + replication ..................................... False + replication_factor .............................. 2 + replication_jump ................................ None + rerun_mode ...................................... validate_results + reset_attention_mask ............................ False + reset_position_ids .............................. False + result_rejected_tracker_filename ................ None + retriever_report_topk_accuracies ................ [] + retriever_score_scaling ......................... False + retriever_seq_length ............................ 256 + retro_add_retriever ............................. False + retro_attention_gate ............................ 1 + retro_cyclic_train_iters ........................ None + retro_encoder_attention_dropout ................. 0.1 + retro_encoder_hidden_dropout .................... 0.1 + retro_encoder_layers ............................ 2 + retro_num_neighbors ............................. 2 + retro_num_retrieved_chunks ...................... 2 + retro_project_dir ............................... None + retro_verify_neighbor_count ..................... True + reuse_grad_buf_for_mxfp8_param_ag ............... False + rope_scaling_factor ............................. 8.0 + rope_type ....................................... None + rotary_base ..................................... 1000000 + rotary_interleaved .............................. False + rotary_percent .................................. 1.0 + rotary_scaling_factor ........................... 1.0 + rotary_seq_len_interpolation_factor ............. None + run_workload_inspector_server ................... False + sample_rate ..................................... 1.0 + save ............................................ None + save_interval ................................... None + save_retain_interval ............................ None + scatter_gather_tensors_in_pipeline .............. True + seed ............................................ 1234 + seq_length ...................................... 8192 + sequence_parallel ............................... False + sft ............................................. False + sft_tokenizer_prompt_format ..................... nemotron-h-aligned + sgd_momentum .................................... 0.9 + short_seq_prob .................................. 0.1 + skip_train ...................................... False + skipped_train_samples ........................... 0 + spec ............................................ None + split ........................................... 100,0,0 + squared_relu .................................... False + start_weight_decay .............................. 0.1 + straggler_ctrlr_port ............................ 65535 + straggler_minmax_count .......................... 1 + suggested_communication_unit_size ............... None + swiglu .......................................... True + swin_backbone_type .............................. tiny + symmetric_ar_type ............................... None + te_rng_tracker .................................. True + tensor_model_parallel_size ...................... 1 + tensorboard_dir ................................. None + tensorboard_log_interval ........................ 1 + tensorboard_queue_size .......................... 1000 + test_data_path .................................. None + test_mode ....................................... False + tiktoken_num_special_tokens ..................... 1000 + tiktoken_pattern ................................ None + tiktoken_special_tokens ......................... None + timing_log_level ................................ 0 + timing_log_option ............................... minmax + titles_data_path ................................ None + tokenizer_model ................................. None + tokenizer_type .................................. GPT2BPETokenizer + torch_fsdp2_reshard_after_forward ............... True + tp_comm_bootstrap_backend ....................... nccl + tp_comm_bulk_dgrad .............................. True + tp_comm_bulk_wgrad .............................. True + tp_comm_overlap ................................. False + tp_comm_overlap_ag .............................. True + tp_comm_overlap_cfg ............................. None + tp_comm_overlap_rs .............................. True + tp_comm_overlap_rs_dgrad ........................ False + tp_comm_split_ag ................................ True + tp_comm_split_rs ................................ True + train_data_path ................................. None + train_iters ..................................... 10 + train_samples ................................... None + train_sync_interval ............................. None + transformer_impl ................................ transformer_engine + transformer_pipeline_model_parallel_size ........ 1 + untie_embeddings_and_output_weights ............. True + use_checkpoint_args ............................. False + use_checkpoint_opt_param_scheduler .............. False + use_cpu_initialization .......................... None + use_custom_fsdp ................................. False + use_dist_ckpt ................................... True + use_dist_ckpt_deprecated ........................ False + use_distributed_optimizer ....................... True + use_flash_attn .................................. False + use_legacy_models ............................... False + use_mp_args_from_checkpoint_args ................ False + use_one_sent_docs ............................... False + use_persistent_ckpt_worker ...................... False + use_precision_aware_optimizer ................... False + use_pytorch_profiler ............................ False + use_ring_exchange_p2p ........................... False + use_rope_scaling ................................ False + use_rotary_position_embeddings .................. False + use_sharp ....................................... False + use_tokenizer_model_from_checkpoint_args ........ True + use_torch_fsdp2 ................................. False + use_torch_optimizer_for_cpu_offload ............. False + use_tp_pp_dp_mapping ............................ False + v_head_dim ...................................... 128 + valid_data_path ................................. None + variable_seq_lengths ............................ False + virtual_pipeline_model_parallel_size ............ None + vision_backbone_type ............................ vit + vision_pretraining .............................. False + vision_pretraining_type ......................... classify + vocab_extra_ids ................................. 0 + vocab_file ...................................... /workspace/data/wikitext/gpt2-vocab.json + vocab_size ...................................... None + wandb_exp_name .................................. + wandb_project ................................... + wandb_save_dir .................................. + weight_decay .................................... 0.1 + weight_decay_incr_style ......................... constant + wgrad_deferral_limit ............................ 0 + world_size ...................................... 8 + yaml_cfg ........................................ None +-------------------- end of arguments --------------------- +INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 1 +> building GPT2BPETokenizer tokenizer ... +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") + > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304) +WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.VALIDATE_RESULTS +> initializing torch distributed ... +WARNING: one_logger package is required to enable e2e metrics tracking. please go to https://confluence.nvidia.com/display/MLWFO/Package+Repositories for details to install it +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +/workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled + warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled") +[Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank [Gloo] Rank 7 is connected to 7 peer ranks. 6Expected number of connected peer ranks is : is connected to 77 peer ranks. Expected number of connected peer ranks is : +7 +[Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 6 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 7 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7 +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +[Gloo] Rank [Gloo] Rank 0 is connected to 00 peer ranks. is connected to Expected number of connected peer ranks is : 00 peer ranks. Expected number of connected peer ranks is : 0 + +[Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0 +> initialized tensor model parallel with size 1 +> initialized pipeline model parallel with size 1 +> setting random seeds to 1234 ... +[rank6]:[W830 00:14:59.010143577 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 6] using GPU 6 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank4]:[W830 00:14:59.010175864 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 4] using GPU 4 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank5]:[W830 00:14:59.010214001 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 5] using GPU 5 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank2]:[W830 00:14:59.010225851 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 2] using GPU 2 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank1]:[W830 00:14:59.010238715 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 1] using GPU 1 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank7]:[W830 00:14:59.012171699 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 7] using GPU 7 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +[rank3]:[W830 00:14:59.017243822 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 3] using GPU 3 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +> compiling dataset index builder ... +make: Entering directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets' +make: Nothing to be done for 'default'. +make: Leaving directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets' +>>> done with dataset index builder. Compilation time: 0.430 seconds +WARNING: constraints for invoking optimized fused softmax kernel are not met. We default back to unfused kernel invocations. +> compiling and loading fused kernels ... +[rank0]:[W830 00:15:00.620129346 ProcessGroupNCCL.cpp:4789] [PG ID 0 PG GUID 0 Rank 0] using GPU 0 as device used by this process is currently unknown. This can potentially cause a hang if this rank to GPU mapping is incorrect. You can specify device_id in init_process_group() to force use of a particular device. +>>> done with compiling and loading fused kernels. Compilation time: 3.696 seconds +time to initialize megatron (seconds): 9.068 +[after megatron is initialized] datetime: 2026-08-30 00:15:06 +building GPT model ... +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( +/usr/local/lib/python3.12/dist-packages/transformer_engine/pytorch/cpu_offload.py:708: DeprecationWarning: Offloading weights is deprecated. Using offload_weights=True does not have any effect. + warnings.warn( + > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 4898003968 +INFO:megatron.core.distributed.distributed_data_parallel:Setting up DistributedDataParallel with config DistributedDataParallelConfig(grad_reduce_in_fp32=True, overlap_grad_reduce=True, overlap_param_gather=True, align_param_gather=False, use_distributed_optimizer=True, num_distributed_optimizer_instances=1, check_for_nan_in_grad=False, check_for_large_grads=False, bucket_size=40000000, pad_buckets_for_high_nccl_busbw=False, average_in_collective=False, fp8_param_gather=False, reuse_grad_buf_for_mxfp8_param_ag=False, use_custom_fsdp=False, data_parallel_sharding_strategy='no_shard', gradient_reduce_div_fusion=True, suggested_communication_unit_size=None, preserve_fp32_weights=True, keep_fp8_transpose_cache_when_using_custom_fsdp=False, nccl_ub=False, fsdp_double_buffer=False) +INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 14 +Params for bucket 1 (206045184 elements, 206045184 padded size): + module.output_layer.weight +Params for bucket 2 (71840000 elements, 71840000 padded size): + module.decoder.layers.11.self_attention.linear_proj.weight + module.decoder.final_layernorm.weight + module.decoder.layers.11.self_attention.q_layernorm.weight + module.decoder.layers.11.mlp.router.weight + module.decoder.layers.11.pre_mlp_layernorm.weight + module.decoder.layers.11.self_attention.k_layernorm.weight + module.decoder.layers.11.self_attention.linear_qkv.weight + module.decoder.layers.11.self_attention.linear_qkv.layer_norm_weight +Params for bucket 3 (71835904 elements, 71835904 padded size): + module.decoder.layers.10.self_attention.k_layernorm.weight + module.decoder.layers.10.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.10.self_attention.linear_proj.weight + module.decoder.layers.10.mlp.router.weight + module.decoder.layers.10.self_attention.q_layernorm.weight + module.decoder.layers.10.pre_mlp_layernorm.weight + module.decoder.layers.10.self_attention.linear_qkv.weight +Params for bucket 4 (71835904 elements, 71835904 padded size): + module.decoder.layers.9.pre_mlp_layernorm.weight + module.decoder.layers.9.self_attention.linear_qkv.weight + module.decoder.layers.9.self_attention.q_layernorm.weight + module.decoder.layers.9.mlp.router.weight + module.decoder.layers.9.self_attention.k_layernorm.weight + module.decoder.layers.9.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.9.self_attention.linear_proj.weight +Params for bucket 5 (71835904 elements, 71835904 padded size): + module.decoder.layers.8.mlp.router.weight + module.decoder.layers.8.pre_mlp_layernorm.weight + module.decoder.layers.8.self_attention.q_layernorm.weight + module.decoder.layers.8.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.8.self_attention.linear_proj.weight + module.decoder.layers.8.self_attention.linear_qkv.weight + module.decoder.layers.8.self_attention.k_layernorm.weight +Params for bucket 6 (71835904 elements, 71835904 padded size): + module.decoder.layers.7.self_attention.k_layernorm.weight + module.decoder.layers.7.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.7.self_attention.linear_proj.weight + module.decoder.layers.7.mlp.router.weight + module.decoder.layers.7.self_attention.q_layernorm.weight + module.decoder.layers.7.pre_mlp_layernorm.weight + module.decoder.layers.7.self_attention.linear_qkv.weight +Params for bucket 7 (71835904 elements, 71835904 padded size): + module.decoder.layers.6.self_attention.linear_qkv.weight + module.decoder.layers.6.self_attention.q_layernorm.weight + module.decoder.layers.6.pre_mlp_layernorm.weight + module.decoder.layers.6.self_attention.linear_proj.weight + module.decoder.layers.6.mlp.router.weight + module.decoder.layers.6.self_attention.k_layernorm.weight + module.decoder.layers.6.self_attention.linear_qkv.layer_norm_weight +Params for bucket 8 (71835904 elements, 71835904 padded size): + module.decoder.layers.5.pre_mlp_layernorm.weight + module.decoder.layers.5.self_attention.linear_qkv.weight + module.decoder.layers.5.self_attention.q_layernorm.weight + module.decoder.layers.5.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.5.mlp.router.weight + module.decoder.layers.5.self_attention.k_layernorm.weight + module.decoder.layers.5.self_attention.linear_proj.weight +Params for bucket 9 (71835904 elements, 71835904 padded size): + module.decoder.layers.4.mlp.router.weight + module.decoder.layers.4.pre_mlp_layernorm.weight + module.decoder.layers.4.self_attention.linear_qkv.weight + module.decoder.layers.4.self_attention.q_layernorm.weight + module.decoder.layers.4.self_attention.k_layernorm.weight + module.decoder.layers.4.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.4.self_attention.linear_proj.weight +Params for bucket 10 (71835904 elements, 71835904 padded size): + module.decoder.layers.3.mlp.router.weight + module.decoder.layers.3.self_attention.k_layernorm.weight + module.decoder.layers.3.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.3.self_attention.linear_qkv.weight + module.decoder.layers.3.self_attention.q_layernorm.weight + module.decoder.layers.3.pre_mlp_layernorm.weight + module.decoder.layers.3.self_attention.linear_proj.weight +Params for bucket 11 (71835904 elements, 71835904 padded size): + module.decoder.layers.2.self_attention.linear_qkv.weight + module.decoder.layers.2.pre_mlp_layernorm.weight + module.decoder.layers.2.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.2.mlp.router.weight + module.decoder.layers.2.self_attention.q_layernorm.weight + module.decoder.layers.2.self_attention.k_layernorm.weight + module.decoder.layers.2.self_attention.linear_proj.weight +Params for bucket 12 (71835904 elements, 71835904 padded size): + module.decoder.layers.1.mlp.router.weight + module.decoder.layers.1.self_attention.k_layernorm.weight + module.decoder.layers.1.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.1.self_attention.linear_proj.weight + module.decoder.layers.1.self_attention.q_layernorm.weight + module.decoder.layers.1.pre_mlp_layernorm.weight + module.decoder.layers.1.self_attention.linear_qkv.weight +Params for bucket 13 (71835904 elements, 71835904 padded size): + module.decoder.layers.0.mlp.router.weight + module.decoder.layers.0.self_attention.linear_qkv.weight + module.decoder.layers.0.self_attention.linear_qkv.layer_norm_weight + module.decoder.layers.0.self_attention.linear_proj.weight + module.decoder.layers.0.self_attention.q_layernorm.weight + module.decoder.layers.0.pre_mlp_layernorm.weight + module.decoder.layers.0.self_attention.k_layernorm.weight +Params for bucket 14 (206045184 elements, 206045184 padded size): + module.embedding.word_embeddings.weight +INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 76 +Params for bucket 1 (44040192 elements, 44040192 padded size): + module.decoder.layers.11.mlp.experts.linear_fc2.weight12 + module.decoder.layers.11.mlp.experts.linear_fc2.weight14 + module.decoder.layers.11.mlp.experts.linear_fc2.weight10 + module.decoder.layers.11.mlp.experts.linear_fc2.weight15 + module.decoder.layers.11.mlp.experts.linear_fc2.weight13 + module.decoder.layers.11.mlp.experts.linear_fc2.weight11 + module.decoder.layers.11.mlp.experts.linear_fc2.weight9 +Params for bucket 2 (44040192 elements, 44040192 padded size): + module.decoder.layers.11.mlp.experts.linear_fc2.weight8 + module.decoder.layers.11.mlp.experts.linear_fc2.weight6 + module.decoder.layers.11.mlp.experts.linear_fc2.weight4 + module.decoder.layers.11.mlp.experts.linear_fc2.weight2 + module.decoder.layers.11.mlp.experts.linear_fc2.weight7 + module.decoder.layers.11.mlp.experts.linear_fc2.weight5 + module.decoder.layers.11.mlp.experts.linear_fc2.weight3 +Params for bucket 3 (50331648 elements, 50331648 padded size): + module.decoder.layers.11.mlp.experts.linear_fc1.weight15 + module.decoder.layers.11.mlp.experts.linear_fc2.weight0 + module.decoder.layers.11.mlp.experts.linear_fc1.weight13 + module.decoder.layers.11.mlp.experts.linear_fc2.weight1 + module.decoder.layers.11.mlp.experts.linear_fc1.weight14 +Params for bucket 4 (50331648 elements, 50331648 padded size): + module.decoder.layers.11.mlp.experts.linear_fc1.weight12 + module.decoder.layers.11.mlp.experts.linear_fc1.weight9 + module.decoder.layers.11.mlp.experts.linear_fc1.weight10 + module.decoder.layers.11.mlp.experts.linear_fc1.weight11 +Params for bucket 5 (50331648 elements, 50331648 padded size): + module.decoder.layers.11.mlp.experts.linear_fc1.weight8 + module.decoder.layers.11.mlp.experts.linear_fc1.weight6 + module.decoder.layers.11.mlp.experts.linear_fc1.weight7 + module.decoder.layers.11.mlp.experts.linear_fc1.weight5 +Params for bucket 6 (50331648 elements, 50331648 padded size): + module.decoder.layers.11.mlp.experts.linear_fc1.weight4 + module.decoder.layers.11.mlp.experts.linear_fc1.weight3 + module.decoder.layers.11.mlp.experts.linear_fc1.weight2 + module.decoder.layers.11.mlp.experts.linear_fc1.weight1 +Params for bucket 7 (44040192 elements, 44040192 padded size): + module.decoder.layers.10.mlp.experts.linear_fc2.weight15 + module.decoder.layers.10.mlp.experts.linear_fc2.weight13 + module.decoder.layers.10.mlp.experts.linear_fc2.weight11 + module.decoder.layers.11.mlp.experts.linear_fc1.weight0 + module.decoder.layers.10.mlp.experts.linear_fc2.weight12 + module.decoder.layers.10.mlp.experts.linear_fc2.weight14 +Params for bucket 8 (44040192 elements, 44040192 padded size): + module.decoder.layers.10.mlp.experts.linear_fc2.weight9 + module.decoder.layers.10.mlp.experts.linear_fc2.weight7 + module.decoder.layers.10.mlp.experts.linear_fc2.weight5 + module.decoder.layers.10.mlp.experts.linear_fc2.weight10 + module.decoder.layers.10.mlp.experts.linear_fc2.weight8 + module.decoder.layers.10.mlp.experts.linear_fc2.weight6 + module.decoder.layers.10.mlp.experts.linear_fc2.weight4 +Params for bucket 9 (50331648 elements, 50331648 padded size): + module.decoder.layers.10.mlp.experts.linear_fc2.weight3 + module.decoder.layers.10.mlp.experts.linear_fc1.weight15 + module.decoder.layers.10.mlp.experts.linear_fc2.weight1 + module.decoder.layers.10.mlp.experts.linear_fc2.weight2 + module.decoder.layers.10.mlp.experts.linear_fc2.weight0 + module.decoder.layers.10.mlp.experts.linear_fc1.weight14 +Params for bucket 10 (50331648 elements, 50331648 padded size): + module.decoder.layers.10.mlp.experts.linear_fc1.weight13 + module.decoder.layers.10.mlp.experts.linear_fc1.weight10 + module.decoder.layers.10.mlp.experts.linear_fc1.weight11 + module.decoder.layers.10.mlp.experts.linear_fc1.weight12 +Params for bucket 11 (50331648 elements, 50331648 padded size): + module.decoder.layers.10.mlp.experts.linear_fc1.weight9 + module.decoder.layers.10.mlp.experts.linear_fc1.weight7 + module.decoder.layers.10.mlp.experts.linear_fc1.weight8 + module.decoder.layers.10.mlp.experts.linear_fc1.weight6 +Params for bucket 12 (50331648 elements, 50331648 padded size): + module.decoder.layers.10.mlp.experts.linear_fc1.weight5 + module.decoder.layers.10.mlp.experts.linear_fc1.weight4 + module.decoder.layers.10.mlp.experts.linear_fc1.weight2 + module.decoder.layers.10.mlp.experts.linear_fc1.weight3 +Params for bucket 13 (44040192 elements, 44040192 padded size): + module.decoder.layers.10.mlp.experts.linear_fc1.weight1 + module.decoder.layers.9.mlp.experts.linear_fc2.weight15 + module.decoder.layers.9.mlp.experts.linear_fc2.weight13 + module.decoder.layers.9.mlp.experts.linear_fc2.weight14 + module.decoder.layers.10.mlp.experts.linear_fc1.weight0 +Params for bucket 14 (44040192 elements, 44040192 padded size): + module.decoder.layers.9.mlp.experts.linear_fc2.weight11 + module.decoder.layers.9.mlp.experts.linear_fc2.weight9 + module.decoder.layers.9.mlp.experts.linear_fc2.weight7 + module.decoder.layers.9.mlp.experts.linear_fc2.weight12 + module.decoder.layers.9.mlp.experts.linear_fc2.weight8 + module.decoder.layers.9.mlp.experts.linear_fc2.weight10 + module.decoder.layers.9.mlp.experts.linear_fc2.weight6 +Params for bucket 15 (50331648 elements, 50331648 padded size): + module.decoder.layers.9.mlp.experts.linear_fc2.weight5 + module.decoder.layers.9.mlp.experts.linear_fc2.weight1 + module.decoder.layers.9.mlp.experts.linear_fc2.weight3 + module.decoder.layers.9.mlp.experts.linear_fc2.weight2 + module.decoder.layers.9.mlp.experts.linear_fc2.weight4 + module.decoder.layers.9.mlp.experts.linear_fc2.weight0 + module.decoder.layers.9.mlp.experts.linear_fc1.weight15 +Params for bucket 16 (50331648 elements, 50331648 padded size): + module.decoder.layers.9.mlp.experts.linear_fc1.weight14 + module.decoder.layers.9.mlp.experts.linear_fc1.weight12 + module.decoder.layers.9.mlp.experts.linear_fc1.weight13 + module.decoder.layers.9.mlp.experts.linear_fc1.weight11 +Params for bucket 17 (50331648 elements, 50331648 padded size): + module.decoder.layers.9.mlp.experts.linear_fc1.weight10 + module.decoder.layers.9.mlp.experts.linear_fc1.weight7 + module.decoder.layers.9.mlp.experts.linear_fc1.weight9 + module.decoder.layers.9.mlp.experts.linear_fc1.weight8 +Params for bucket 18 (50331648 elements, 50331648 padded size): + module.decoder.layers.9.mlp.experts.linear_fc1.weight6 + module.decoder.layers.9.mlp.experts.linear_fc1.weight3 + module.decoder.layers.9.mlp.experts.linear_fc1.weight5 + module.decoder.layers.9.mlp.experts.linear_fc1.weight4 +Params for bucket 19 (44040192 elements, 44040192 padded size): + module.decoder.layers.9.mlp.experts.linear_fc1.weight2 + module.decoder.layers.9.mlp.experts.linear_fc1.weight0 + module.decoder.layers.9.mlp.experts.linear_fc1.weight1 + module.decoder.layers.8.mlp.experts.linear_fc2.weight15 +Params for bucket 20 (44040192 elements, 44040192 padded size): + module.decoder.layers.8.mlp.experts.linear_fc2.weight14 + module.decoder.layers.8.mlp.experts.linear_fc2.weight12 + module.decoder.layers.8.mlp.experts.linear_fc2.weight10 + module.decoder.layers.8.mlp.experts.linear_fc2.weight8 + module.decoder.layers.8.mlp.experts.linear_fc2.weight13 + module.decoder.layers.8.mlp.experts.linear_fc2.weight11 + module.decoder.layers.8.mlp.experts.linear_fc2.weight9 +Params for bucket 21 (44040192 elements, 44040192 padded size): + module.decoder.layers.8.mlp.experts.linear_fc2.weight6 + module.decoder.layers.8.mlp.experts.linear_fc2.weight4 + module.decoder.layers.8.mlp.experts.linear_fc2.weight2 + module.decoder.layers.8.mlp.experts.linear_fc2.weight7 + module.decoder.layers.8.mlp.experts.linear_fc2.weight5 + module.decoder.layers.8.mlp.experts.linear_fc2.weight3 + module.decoder.layers.8.mlp.experts.linear_fc2.weight1 +Params for bucket 22 (44040192 elements, 44040192 padded size): + module.decoder.layers.8.mlp.experts.linear_fc2.weight0 + module.decoder.layers.8.mlp.experts.linear_fc1.weight15 + module.decoder.layers.8.mlp.experts.linear_fc1.weight13 + module.decoder.layers.8.mlp.experts.linear_fc1.weight14 +Params for bucket 23 (50331648 elements, 50331648 padded size): + module.decoder.layers.8.mlp.experts.linear_fc1.weight12 + module.decoder.layers.8.mlp.experts.linear_fc1.weight9 + module.decoder.layers.8.mlp.experts.linear_fc1.weight10 + module.decoder.layers.8.mlp.experts.linear_fc1.weight11 +Params for bucket 24 (50331648 elements, 50331648 padded size): + module.decoder.layers.8.mlp.experts.linear_fc1.weight8 + module.decoder.layers.8.mlp.experts.linear_fc1.weight5 + module.decoder.layers.8.mlp.experts.linear_fc1.weight6 + module.decoder.layers.8.mlp.experts.linear_fc1.weight7 +Params for bucket 25 (50331648 elements, 50331648 padded size): + module.decoder.layers.8.mlp.experts.linear_fc1.weight4 + module.decoder.layers.8.mlp.experts.linear_fc1.weight2 + module.decoder.layers.8.mlp.experts.linear_fc1.weight3 + module.decoder.layers.8.mlp.experts.linear_fc1.weight1 +Params for bucket 26 (44040192 elements, 44040192 padded size): + module.decoder.layers.8.mlp.experts.linear_fc1.weight0 + module.decoder.layers.7.mlp.experts.linear_fc2.weight14 + module.decoder.layers.7.mlp.experts.linear_fc2.weight12 + module.decoder.layers.7.mlp.experts.linear_fc2.weight15 + module.decoder.layers.7.mlp.experts.linear_fc2.weight13 + module.decoder.layers.7.mlp.experts.linear_fc2.weight11 +Params for bucket 27 (44040192 elements, 44040192 padded size): + module.decoder.layers.7.mlp.experts.linear_fc2.weight10 + module.decoder.layers.7.mlp.experts.linear_fc2.weight6 + module.decoder.layers.7.mlp.experts.linear_fc2.weight8 + module.decoder.layers.7.mlp.experts.linear_fc2.weight4 + module.decoder.layers.7.mlp.experts.linear_fc2.weight7 + module.decoder.layers.7.mlp.experts.linear_fc2.weight9 + module.decoder.layers.7.mlp.experts.linear_fc2.weight5 +Params for bucket 28 (50331648 elements, 50331648 padded size): + module.decoder.layers.7.mlp.experts.linear_fc2.weight0 + module.decoder.layers.7.mlp.experts.linear_fc2.weight2 + module.decoder.layers.7.mlp.experts.linear_fc1.weight15 + module.decoder.layers.7.mlp.experts.linear_fc2.weight3 + module.decoder.layers.7.mlp.experts.linear_fc2.weight1 + module.decoder.layers.7.mlp.experts.linear_fc1.weight14 +Params for bucket 29 (50331648 elements, 50331648 padded size): + module.decoder.layers.7.mlp.experts.linear_fc1.weight13 + module.decoder.layers.7.mlp.experts.linear_fc1.weight10 + module.decoder.layers.7.mlp.experts.linear_fc1.weight11 + module.decoder.layers.7.mlp.experts.linear_fc1.weight12 +Params for bucket 30 (50331648 elements, 50331648 padded size): + module.decoder.layers.7.mlp.experts.linear_fc1.weight9 + module.decoder.layers.7.mlp.experts.linear_fc1.weight7 + module.decoder.layers.7.mlp.experts.linear_fc1.weight8 + module.decoder.layers.7.mlp.experts.linear_fc1.weight6 +Params for bucket 31 (50331648 elements, 50331648 padded size): + module.decoder.layers.7.mlp.experts.linear_fc1.weight5 + module.decoder.layers.7.mlp.experts.linear_fc1.weight3 + module.decoder.layers.7.mlp.experts.linear_fc1.weight4 + module.decoder.layers.7.mlp.experts.linear_fc1.weight2 +Params for bucket 32 (44040192 elements, 44040192 padded size): + module.decoder.layers.7.mlp.experts.linear_fc1.weight1 + module.decoder.layers.6.mlp.experts.linear_fc2.weight15 + module.decoder.layers.6.mlp.experts.linear_fc2.weight13 + module.decoder.layers.7.mlp.experts.linear_fc1.weight0 + module.decoder.layers.6.mlp.experts.linear_fc2.weight14 +Params for bucket 33 (44040192 elements, 44040192 padded size): + module.decoder.layers.6.mlp.experts.linear_fc2.weight11 + module.decoder.layers.6.mlp.experts.linear_fc2.weight9 + module.decoder.layers.6.mlp.experts.linear_fc2.weight7 + module.decoder.layers.6.mlp.experts.linear_fc2.weight12 + module.decoder.layers.6.mlp.experts.linear_fc2.weight8 + module.decoder.layers.6.mlp.experts.linear_fc2.weight10 + module.decoder.layers.6.mlp.experts.linear_fc2.weight6 +Params for bucket 34 (50331648 elements, 50331648 padded size): + module.decoder.layers.6.mlp.experts.linear_fc2.weight5 + module.decoder.layers.6.mlp.experts.linear_fc2.weight1 + module.decoder.layers.6.mlp.experts.linear_fc2.weight3 + module.decoder.layers.6.mlp.experts.linear_fc2.weight4 + module.decoder.layers.6.mlp.experts.linear_fc2.weight2 + module.decoder.layers.6.mlp.experts.linear_fc2.weight0 + module.decoder.layers.6.mlp.experts.linear_fc1.weight15 +Params for bucket 35 (50331648 elements, 50331648 padded size): + module.decoder.layers.6.mlp.experts.linear_fc1.weight14 + module.decoder.layers.6.mlp.experts.linear_fc1.weight13 + module.decoder.layers.6.mlp.experts.linear_fc1.weight11 + module.decoder.layers.6.mlp.experts.linear_fc1.weight12 +Params for bucket 36 (50331648 elements, 50331648 padded size): + module.decoder.layers.6.mlp.experts.linear_fc1.weight10 + module.decoder.layers.6.mlp.experts.linear_fc1.weight8 + module.decoder.layers.6.mlp.experts.linear_fc1.weight9 + module.decoder.layers.6.mlp.experts.linear_fc1.weight7 +Params for bucket 37 (50331648 elements, 50331648 padded size): + module.decoder.layers.6.mlp.experts.linear_fc1.weight6 + module.decoder.layers.6.mlp.experts.linear_fc1.weight5 + module.decoder.layers.6.mlp.experts.linear_fc1.weight4 + module.decoder.layers.6.mlp.experts.linear_fc1.weight3 +Params for bucket 38 (44040192 elements, 44040192 padded size): + module.decoder.layers.6.mlp.experts.linear_fc1.weight2 + module.decoder.layers.5.mlp.experts.linear_fc2.weight15 + module.decoder.layers.6.mlp.experts.linear_fc1.weight1 + module.decoder.layers.6.mlp.experts.linear_fc1.weight0 +Params for bucket 39 (44040192 elements, 44040192 padded size): + module.decoder.layers.5.mlp.experts.linear_fc2.weight14 + module.decoder.layers.5.mlp.experts.linear_fc2.weight10 + module.decoder.layers.5.mlp.experts.linear_fc2.weight12 + module.decoder.layers.5.mlp.experts.linear_fc2.weight8 + module.decoder.layers.5.mlp.experts.linear_fc2.weight13 + module.decoder.layers.5.mlp.experts.linear_fc2.weight11 + module.decoder.layers.5.mlp.experts.linear_fc2.weight9 +Params for bucket 40 (44040192 elements, 44040192 padded size): + module.decoder.layers.5.mlp.experts.linear_fc2.weight4 + module.decoder.layers.5.mlp.experts.linear_fc2.weight6 + module.decoder.layers.5.mlp.experts.linear_fc2.weight2 + module.decoder.layers.5.mlp.experts.linear_fc2.weight7 + module.decoder.layers.5.mlp.experts.linear_fc2.weight5 + module.decoder.layers.5.mlp.experts.linear_fc2.weight3 + module.decoder.layers.5.mlp.experts.linear_fc2.weight1 +Params for bucket 41 (44040192 elements, 44040192 padded size): + module.decoder.layers.5.mlp.experts.linear_fc2.weight0 + module.decoder.layers.5.mlp.experts.linear_fc1.weight14 + module.decoder.layers.5.mlp.experts.linear_fc1.weight15 + module.decoder.layers.5.mlp.experts.linear_fc1.weight13 +Params for bucket 42 (50331648 elements, 50331648 padded size): + module.decoder.layers.5.mlp.experts.linear_fc1.weight12 + module.decoder.layers.5.mlp.experts.linear_fc1.weight11 + module.decoder.layers.5.mlp.experts.linear_fc1.weight9 + module.decoder.layers.5.mlp.experts.linear_fc1.weight10 +Params for bucket 43 (50331648 elements, 50331648 padded size): + module.decoder.layers.5.mlp.experts.linear_fc1.weight8 + module.decoder.layers.5.mlp.experts.linear_fc1.weight5 + module.decoder.layers.5.mlp.experts.linear_fc1.weight6 + module.decoder.layers.5.mlp.experts.linear_fc1.weight7 +Params for bucket 44 (50331648 elements, 50331648 padded size): + module.decoder.layers.5.mlp.experts.linear_fc1.weight4 + module.decoder.layers.5.mlp.experts.linear_fc1.weight1 + module.decoder.layers.5.mlp.experts.linear_fc1.weight2 + module.decoder.layers.5.mlp.experts.linear_fc1.weight3 +Params for bucket 45 (44040192 elements, 44040192 padded size): + module.decoder.layers.4.mlp.experts.linear_fc2.weight13 + module.decoder.layers.4.mlp.experts.linear_fc2.weight15 + module.decoder.layers.4.mlp.experts.linear_fc2.weight11 + module.decoder.layers.5.mlp.experts.linear_fc1.weight0 + module.decoder.layers.4.mlp.experts.linear_fc2.weight14 + module.decoder.layers.4.mlp.experts.linear_fc2.weight12 +Params for bucket 46 (44040192 elements, 44040192 padded size): + module.decoder.layers.4.mlp.experts.linear_fc2.weight9 + module.decoder.layers.4.mlp.experts.linear_fc2.weight7 + module.decoder.layers.4.mlp.experts.linear_fc2.weight5 + module.decoder.layers.4.mlp.experts.linear_fc2.weight10 + module.decoder.layers.4.mlp.experts.linear_fc2.weight8 + module.decoder.layers.4.mlp.experts.linear_fc2.weight6 + module.decoder.layers.4.mlp.experts.linear_fc2.weight4 +Params for bucket 47 (50331648 elements, 50331648 padded size): + module.decoder.layers.4.mlp.experts.linear_fc2.weight3 + module.decoder.layers.4.mlp.experts.linear_fc2.weight1 + module.decoder.layers.4.mlp.experts.linear_fc1.weight14 + module.decoder.layers.4.mlp.experts.linear_fc2.weight2 + module.decoder.layers.4.mlp.experts.linear_fc1.weight15 + module.decoder.layers.4.mlp.experts.linear_fc2.weight0 +Params for bucket 48 (50331648 elements, 50331648 padded size): + module.decoder.layers.4.mlp.experts.linear_fc1.weight13 + module.decoder.layers.4.mlp.experts.linear_fc1.weight11 + module.decoder.layers.4.mlp.experts.linear_fc1.weight12 + module.decoder.layers.4.mlp.experts.linear_fc1.weight10 +Params for bucket 49 (50331648 elements, 50331648 padded size): + module.decoder.layers.4.mlp.experts.linear_fc1.weight9 + module.decoder.layers.4.mlp.experts.linear_fc1.weight8 + module.decoder.layers.4.mlp.experts.linear_fc1.weight6 + module.decoder.layers.4.mlp.experts.linear_fc1.weight7 +Params for bucket 50 (50331648 elements, 50331648 padded size): + module.decoder.layers.4.mlp.experts.linear_fc1.weight5 + module.decoder.layers.4.mlp.experts.linear_fc1.weight2 + module.decoder.layers.4.mlp.experts.linear_fc1.weight4 + module.decoder.layers.4.mlp.experts.linear_fc1.weight3 +Params for bucket 51 (44040192 elements, 44040192 padded size): + module.decoder.layers.4.mlp.experts.linear_fc1.weight0 + module.decoder.layers.3.mlp.experts.linear_fc2.weight14 + module.decoder.layers.4.mlp.experts.linear_fc1.weight1 + module.decoder.layers.3.mlp.experts.linear_fc2.weight15 + module.decoder.layers.3.mlp.experts.linear_fc2.weight13 +Params for bucket 52 (44040192 elements, 44040192 padded size): + module.decoder.layers.3.mlp.experts.linear_fc2.weight12 + module.decoder.layers.3.mlp.experts.linear_fc2.weight10 + module.decoder.layers.3.mlp.experts.linear_fc2.weight8 + module.decoder.layers.3.mlp.experts.linear_fc2.weight6 + module.decoder.layers.3.mlp.experts.linear_fc2.weight9 + module.decoder.layers.3.mlp.experts.linear_fc2.weight11 + module.decoder.layers.3.mlp.experts.linear_fc2.weight7 +Params for bucket 53 (50331648 elements, 50331648 padded size): + module.decoder.layers.3.mlp.experts.linear_fc2.weight2 + module.decoder.layers.3.mlp.experts.linear_fc2.weight4 + module.decoder.layers.3.mlp.experts.linear_fc2.weight0 + module.decoder.layers.3.mlp.experts.linear_fc1.weight15 + module.decoder.layers.3.mlp.experts.linear_fc2.weight5 + module.decoder.layers.3.mlp.experts.linear_fc2.weight3 + module.decoder.layers.3.mlp.experts.linear_fc2.weight1 +Params for bucket 54 (50331648 elements, 50331648 padded size): + module.decoder.layers.3.mlp.experts.linear_fc1.weight14 + module.decoder.layers.3.mlp.experts.linear_fc1.weight11 + module.decoder.layers.3.mlp.experts.linear_fc1.weight12 + module.decoder.layers.3.mlp.experts.linear_fc1.weight13 +Params for bucket 55 (50331648 elements, 50331648 padded size): + module.decoder.layers.3.mlp.experts.linear_fc1.weight10 + module.decoder.layers.3.mlp.experts.linear_fc1.weight8 + module.decoder.layers.3.mlp.experts.linear_fc1.weight9 + module.decoder.layers.3.mlp.experts.linear_fc1.weight7 +Params for bucket 56 (50331648 elements, 50331648 padded size): + module.decoder.layers.3.mlp.experts.linear_fc1.weight6 + module.decoder.layers.3.mlp.experts.linear_fc1.weight3 + module.decoder.layers.3.mlp.experts.linear_fc1.weight4 + module.decoder.layers.3.mlp.experts.linear_fc1.weight5 +Params for bucket 57 (44040192 elements, 44040192 padded size): + module.decoder.layers.3.mlp.experts.linear_fc1.weight2 + module.decoder.layers.3.mlp.experts.linear_fc1.weight0 + module.decoder.layers.2.mlp.experts.linear_fc2.weight15 + module.decoder.layers.3.mlp.experts.linear_fc1.weight1 +Params for bucket 58 (44040192 elements, 44040192 padded size): + module.decoder.layers.2.mlp.experts.linear_fc2.weight14 + module.decoder.layers.2.mlp.experts.linear_fc2.weight12 + module.decoder.layers.2.mlp.experts.linear_fc2.weight10 + module.decoder.layers.2.mlp.experts.linear_fc2.weight8 + module.decoder.layers.2.mlp.experts.linear_fc2.weight13 + module.decoder.layers.2.mlp.experts.linear_fc2.weight11 + module.decoder.layers.2.mlp.experts.linear_fc2.weight9 +Params for bucket 59 (44040192 elements, 44040192 padded size): + module.decoder.layers.2.mlp.experts.linear_fc2.weight6 + module.decoder.layers.2.mlp.experts.linear_fc2.weight4 + module.decoder.layers.2.mlp.experts.linear_fc2.weight2 + module.decoder.layers.2.mlp.experts.linear_fc2.weight7 + module.decoder.layers.2.mlp.experts.linear_fc2.weight3 + module.decoder.layers.2.mlp.experts.linear_fc2.weight5 + module.decoder.layers.2.mlp.experts.linear_fc2.weight1 +Params for bucket 60 (44040192 elements, 44040192 padded size): + module.decoder.layers.2.mlp.experts.linear_fc2.weight0 + module.decoder.layers.2.mlp.experts.linear_fc1.weight15 + module.decoder.layers.2.mlp.experts.linear_fc1.weight13 + module.decoder.layers.2.mlp.experts.linear_fc1.weight14 +Params for bucket 61 (50331648 elements, 50331648 padded size): + module.decoder.layers.2.mlp.experts.linear_fc1.weight12 + module.decoder.layers.2.mlp.experts.linear_fc1.weight9 + module.decoder.layers.2.mlp.experts.linear_fc1.weight10 + module.decoder.layers.2.mlp.experts.linear_fc1.weight11 +Params for bucket 62 (50331648 elements, 50331648 padded size): + module.decoder.layers.2.mlp.experts.linear_fc1.weight8 + module.decoder.layers.2.mlp.experts.linear_fc1.weight6 + module.decoder.layers.2.mlp.experts.linear_fc1.weight7 + module.decoder.layers.2.mlp.experts.linear_fc1.weight5 +Params for bucket 63 (50331648 elements, 50331648 padded size): + module.decoder.layers.2.mlp.experts.linear_fc1.weight4 + module.decoder.layers.2.mlp.experts.linear_fc1.weight3 + module.decoder.layers.2.mlp.experts.linear_fc1.weight1 + module.decoder.layers.2.mlp.experts.linear_fc1.weight2 +Params for bucket 64 (44040192 elements, 44040192 padded size): + module.decoder.layers.1.mlp.experts.linear_fc2.weight15 + module.decoder.layers.1.mlp.experts.linear_fc2.weight13 + module.decoder.layers.1.mlp.experts.linear_fc2.weight11 + module.decoder.layers.2.mlp.experts.linear_fc1.weight0 + module.decoder.layers.1.mlp.experts.linear_fc2.weight14 + module.decoder.layers.1.mlp.experts.linear_fc2.weight12 +Params for bucket 65 (44040192 elements, 44040192 padded size): + module.decoder.layers.1.mlp.experts.linear_fc2.weight7 + module.decoder.layers.1.mlp.experts.linear_fc2.weight9 + module.decoder.layers.1.mlp.experts.linear_fc2.weight5 + module.decoder.layers.1.mlp.experts.linear_fc2.weight10 + module.decoder.layers.1.mlp.experts.linear_fc2.weight8 + module.decoder.layers.1.mlp.experts.linear_fc2.weight6 + module.decoder.layers.1.mlp.experts.linear_fc2.weight4 +Params for bucket 66 (50331648 elements, 50331648 padded size): + module.decoder.layers.1.mlp.experts.linear_fc2.weight3 + module.decoder.layers.1.mlp.experts.linear_fc2.weight1 + module.decoder.layers.1.mlp.experts.linear_fc1.weight14 + module.decoder.layers.1.mlp.experts.linear_fc1.weight15 + module.decoder.layers.1.mlp.experts.linear_fc2.weight2 + module.decoder.layers.1.mlp.experts.linear_fc2.weight0 +Params for bucket 67 (50331648 elements, 50331648 padded size): + module.decoder.layers.1.mlp.experts.linear_fc1.weight13 + module.decoder.layers.1.mlp.experts.linear_fc1.weight12 + module.decoder.layers.1.mlp.experts.linear_fc1.weight11 + module.decoder.layers.1.mlp.experts.linear_fc1.weight10 +Params for bucket 68 (50331648 elements, 50331648 padded size): + module.decoder.layers.1.mlp.experts.linear_fc1.weight9 + module.decoder.layers.1.mlp.experts.linear_fc1.weight8 + module.decoder.layers.1.mlp.experts.linear_fc1.weight6 + module.decoder.layers.1.mlp.experts.linear_fc1.weight7 +Params for bucket 69 (50331648 elements, 50331648 padded size): + module.decoder.layers.1.mlp.experts.linear_fc1.weight5 + module.decoder.layers.1.mlp.experts.linear_fc1.weight2 + module.decoder.layers.1.mlp.experts.linear_fc1.weight3 + module.decoder.layers.1.mlp.experts.linear_fc1.weight4 +Params for bucket 70 (44040192 elements, 44040192 padded size): + module.decoder.layers.0.mlp.experts.linear_fc2.weight13 + module.decoder.layers.1.mlp.experts.linear_fc1.weight0 + module.decoder.layers.0.mlp.experts.linear_fc2.weight15 + module.decoder.layers.1.mlp.experts.linear_fc1.weight1 + module.decoder.layers.0.mlp.experts.linear_fc2.weight14 +Params for bucket 71 (44040192 elements, 44040192 padded size): + module.decoder.layers.0.mlp.experts.linear_fc2.weight9 + module.decoder.layers.0.mlp.experts.linear_fc2.weight11 + module.decoder.layers.0.mlp.experts.linear_fc2.weight7 + module.decoder.layers.0.mlp.experts.linear_fc2.weight12 + module.decoder.layers.0.mlp.experts.linear_fc2.weight10 + module.decoder.layers.0.mlp.experts.linear_fc2.weight8 + module.decoder.layers.0.mlp.experts.linear_fc2.weight6 +Params for bucket 72 (50331648 elements, 50331648 padded size): + module.decoder.layers.0.mlp.experts.linear_fc2.weight5 + module.decoder.layers.0.mlp.experts.linear_fc2.weight3 + module.decoder.layers.0.mlp.experts.linear_fc2.weight1 + module.decoder.layers.0.mlp.experts.linear_fc1.weight15 + module.decoder.layers.0.mlp.experts.linear_fc2.weight4 + module.decoder.layers.0.mlp.experts.linear_fc2.weight2 + module.decoder.layers.0.mlp.experts.linear_fc2.weight0 +Params for bucket 73 (50331648 elements, 50331648 padded size): + module.decoder.layers.0.mlp.experts.linear_fc1.weight14 + module.decoder.layers.0.mlp.experts.linear_fc1.weight12 + module.decoder.layers.0.mlp.experts.linear_fc1.weight13 + module.decoder.layers.0.mlp.experts.linear_fc1.weight11 +Params for bucket 74 (50331648 elements, 50331648 padded size): + module.decoder.layers.0.mlp.experts.linear_fc1.weight10 + module.decoder.layers.0.mlp.experts.linear_fc1.weight9 + module.decoder.layers.0.mlp.experts.linear_fc1.weight7 + module.decoder.layers.0.mlp.experts.linear_fc1.weight8 +Params for bucket 75 (50331648 elements, 50331648 padded size): + module.decoder.layers.0.mlp.experts.linear_fc1.weight6 + module.decoder.layers.0.mlp.experts.linear_fc1.weight4 + module.decoder.layers.0.mlp.experts.linear_fc1.weight3 + module.decoder.layers.0.mlp.experts.linear_fc1.weight5 +Params for bucket 76 (37748736 elements, 37748736 padded size): + module.decoder.layers.0.mlp.experts.linear_fc1.weight2 + module.decoder.layers.0.mlp.experts.linear_fc1.weight0 + module.decoder.layers.0.mlp.experts.linear_fc1.weight1 +INFO:megatron.core.optimizer:Setting up optimizer with config OptimizerConfig(optimizer='adam', lr=0.0001, min_lr=1e-05, decoupled_lr=None, decoupled_min_lr=None, weight_decay=0.1, fp8_recipe='mxfp8', fp16=False, bf16=True, reuse_grad_buf_for_mxfp8_param_ag=False, params_dtype=torch.bfloat16, use_precision_aware_optimizer=False, store_param_remainders=True, main_grads_dtype=torch.float32, main_params_dtype=torch.float32, exp_avg_dtype=torch.float32, exp_avg_sq_dtype=torch.float32, loss_scale=None, initial_loss_scale=4294967296, min_loss_scale=1.0, loss_scale_window=1000, hysteresis=2, adam_beta1=0.9, adam_beta2=0.999, adam_eps=1e-08, sgd_momentum=0.9, use_distributed_optimizer=True, overlap_param_gather_with_optimizer_step=False, optimizer_cpu_offload=False, optimizer_offload_fraction=1.0, use_torch_optimizer_for_cpu_offload=False, overlap_cpu_optimizer_d2h_h2d=False, pin_cpu_grads=True, pin_cpu_params=True, clip_grad=1.0, log_num_zeros_in_grad=False, barrier_with_L1_time=True, timers=, config_logger_dir='') +INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine +[after model, optimizer, and learning rate scheduler are built] datetime: 2026-08-30 00:15:06 +> building train, validation, and test datasets ... + > datasets target sizes (minimum size): + train: 80 + validation: 0 + test: 0 +INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 1.0), None, None] +> building train, validation, and test datasets for GPT ... +INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building GPTDataset splits with sizes=(80, 0, 0) and config=GPTDatasetConfig(random_seed=1234, sequence_length=8192, blend=(['/workspace/data/wikitext/wikitext103_gpt2_text_document'], None), blend_per_split=None, split='100,0,0', split_matrix=[(0, 1.0), None, None], num_dataset_builder_threads=1, path_to_cache=None, mmap_bin_files=True, mock=False, tokenizer=, mid_level_dataset_surplus=0.005, reset_position_ids=False, reset_attention_mask=False, eod_mask_loss=False, create_attention_mask=True, drop_last_partial_validation_sequence=True, add_extra_token_to_sequence=True, object_storage_cache_path=None) +INFO:megatron.core.datasets.indexed_dataset:Load the _IndexReader from /workspace/data/wikitext/wikitext103_gpt2_text_document.idx +INFO:megatron.core.datasets.indexed_dataset: Extract the sequence lengths +INFO:megatron.core.datasets.indexed_dataset: Extract the sequence pointers +INFO:megatron.core.datasets.indexed_dataset: Extract the document indices +INFO:megatron.core.datasets.indexed_dataset:> total number of sequences: 1165029 +INFO:megatron.core.datasets.indexed_dataset:> total number of documents: 1801350 +INFO:megatron.core.datasets.gpt_dataset:Build and save the GPTDataset train indices +INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 14536 +INFO:megatron.core.datasets.gpt_dataset:> total number of epochs: 1 +> finished creating GPT datasets ... +[after dataloaders are built] datetime: 2026-08-30 00:15:07 +done with setup ... +(min, max) time across ranks (ms): + model-and-optimizer-setup ......................: (643.92, 655.14) + train/valid/test-data-iterators-setup ..........: (146.06, 156.34) +training ... +Overwriting rerun_state_machine.current_iteration from -1 to 0... +[before the start of training step] datetime: 2026-08-30 00:15:07 +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future. + warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.") + [2026-08-30 00:15:52] iteration 1/ 10 | consumed samples: 8 | elapsed time per iteration (ms): 45750.5 | throughput per GPU (TFLOP/s/GPU): 4.0 | learning rate: 2.000000E-05 | global batch size: 8 | lm loss: 1.115992E+01 | loss scale: 1.0 | grad norm: 44.779 | number of skipped iterations: 0 | number of nan iterations: 0 | +Number of parameters in transformer block in billions: 29.85 +Number of parameters in embedding layers in billions: 0.41 +Total number of parameters in billions: 30.26 +Number of parameters in most loaded shard in billions: 30.2590 +Theoretical memory footprints: weight and optimizer=216428.96 MB +[Rank 0] (after 1 iterations) memory (MB) | allocated: 81339.25341796875 | max allocated: 87165.23095703125 | reserved: 97354.0 | max reserved: 97354.0 +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") +/workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future. + warnings.warn("The use_host_meta is deprecated, it will be removed in the future.") + [2026-08-30 00:15:53] iteration 2/ 10 | consumed samples: 16 | elapsed time per iteration (ms): 553.1 | throughput per GPU (TFLOP/s/GPU): 326.9 | learning rate: 4.000000E-05 | global batch size: 8 | lm loss: 1.117604E+01 | loss scale: 1.0 | grad norm: 44.306 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:15:53] iteration 3/ 10 | consumed samples: 24 | elapsed time per iteration (ms): 357.9 | throughput per GPU (TFLOP/s/GPU): 505.3 | learning rate: 6.000000E-05 | global batch size: 8 | lm loss: 9.693825E+00 | loss scale: 1.0 | grad norm: 66.342 | number of skipped iterations: 0 | number of nan iterations: 0 | +Capture CUDA graph for training!!! +CUDA graph capture done!!! + [2026-08-30 00:15:56] iteration 4/ 10 | consumed samples: 32 | elapsed time per iteration (ms): 2747.8 | throughput per GPU (TFLOP/s/GPU): 65.8 | learning rate: 8.000000E-05 | global batch size: 8 | lm loss: 1.338515E+01 | loss scale: 1.0 | grad norm: 19.956 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:15:56] iteration 5/ 10 | consumed samples: 40 | elapsed time per iteration (ms): 316.6 | throughput per GPU (TFLOP/s/GPU): 571.1 | learning rate: 1.000000E-04 | global batch size: 8 | lm loss: 1.010135E+01 | loss scale: 1.0 | grad norm: 83.843 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:15:57] iteration 6/ 10 | consumed samples: 48 | elapsed time per iteration (ms): 315.0 | throughput per GPU (TFLOP/s/GPU): 574.0 | learning rate: 7.750000E-05 | global batch size: 8 | lm loss: 1.357873E+01 | loss scale: 1.0 | grad norm: 121.634 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:16:00] iteration 7/ 10 | consumed samples: 56 | elapsed time per iteration (ms): 3026.7 | throughput per GPU (TFLOP/s/GPU): 59.7 | learning rate: 3.250000E-05 | global batch size: 8 | lm loss: 1.031766E+01 | loss scale: 1.0 | grad norm: 6.315 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:16:00] iteration 8/ 10 | consumed samples: 64 | elapsed time per iteration (ms): 327.3 | throughput per GPU (TFLOP/s/GPU): 552.5 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 9.886835E+00 | loss scale: 1.0 | grad norm: 5.614 | number of skipped iterations: 0 | number of nan iterations: 0 | +Capture range started in the application. +Capture range ended in the application. +Generating '/tmp/nsys-report-cfb6.qdstrm' + [1/1] [0% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-repProcessing events... + [1/1] [0% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [5% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [0% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [5% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [8% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [10% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [13% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=16% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [==19% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [==21% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [===24% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [====27% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [====28% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=====29% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=====30% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=====32% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [======33% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [======34% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [======35% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=======36% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=======37% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=======38% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [========40% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [========41% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [========42% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=========44% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=========45% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [==========47% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [==========48% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [===========50% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [===========51% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [===========53% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [============54% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [=============60% ] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [========================100%] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep [1/1] [========================100%] QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep + [2026-08-30 00:16:42] iteration 9/ 10 | consumed samples: 72 | elapsed time per iteration (ms): 41887.6 | throughput per GPU (TFLOP/s/GPU): 4.3 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 9.643233E+00 | loss scale: 1.0 | grad norm: 5.519 | number of skipped iterations: 0 | number of nan iterations: 0 | + [2026-08-30 00:16:42] iteration 10/ 10 | consumed samples: 80 | elapsed time per iteration (ms): 325.5 | throughput per GPU (TFLOP/s/GPU): 555.6 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 9.548412E+00 | loss scale: 1.0 | grad norm: 5.381 | number of skipped iterations: 0 | number of nan iterations: 0 | +[after training is done] datetime: 2026-08-30 00:16:42 +[rank0]:[W830 00:16:44.422994918 ProcessGroupNCCL.cpp:1505] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator()) +[rank7]:[W830 00:16:44.833124189 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +[rank5]:[W830 00:16:44.847165253 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +pure virtual method called +terminate called without an active exception +[rank0]:[W830 00:16:45.271746439 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +[rank3]:[W830 00:16:45.305610521 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +[rank6]:[W830 00:16:45.318662398 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +[rank1]:[W830 00:16:45.326520150 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +[rank4]:[W830 00:16:45.338510981 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +pure virtual method called +terminate called without an active exception +[rank2]:[W830 00:16:45.397463197 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator()) +pure virtual method called +terminate called without an active exception +W0830 00:16:45.639000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291263 closing signal SIGTERM +W0830 00:16:45.640000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291264 closing signal SIGTERM +W0830 00:16:45.640000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291265 closing signal SIGTERM +W0830 00:16:45.641000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291266 closing signal SIGTERM +W0830 00:16:45.641000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291267 closing signal SIGTERM +W0830 00:16:45.641000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291268 closing signal SIGTERM +W0830 00:16:45.641000 291145 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 291269 closing signal SIGTERM +E0830 00:16:48.511000 291145 torch/distributed/elastic/multiprocessing/api.py:874] failed (exitcode: -6) local_rank: 7 (pid: 291270) of binary: /usr/bin/python +Traceback (most recent call last): + File "/usr/local/bin/torchrun", line 8, in + sys.exit(main()) + ^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper + return f(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main + run(args) + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run + elastic_launch( + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 270, in launch_agent + raise ChildFailedError( +torch.distributed.elastic.multiprocessing.errors.ChildFailedError: +============================================================ +pretrain_gpt.py FAILED +------------------------------------------------------------ +Failures: + +------------------------------------------------------------ +Root Cause (first observed failure): +[0]: + time : 2026-08-30_00:16:45 + host : sp-aiinfra-interactive-aghosh-7jtjl-8xb200-moe-cg-0 + rank : 7 (local_rank: 7) + exitcode : -6 (pid: 291270) + error_file: + traceback : Signal 6 (SIGABRT) received by PID 291270 +============================================================ +Collection stop is not allowed in this state. +Generated: + /workspace/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/verify_trace_ranks.log b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/verify_trace_ranks.log new file mode 100644 index 0000000000000000000000000000000000000000..8c7063a359ddd42b6fe04fa3a815460aded79bf1 --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/verify_trace_ranks.log @@ -0,0 +1,6 @@ +report : /workspace/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep +kernels : 120211 +devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7] +processes: 1 +markers : 8832 echo_trace__ kernels on devices [0, 1, 2, 3, 4, 5, 6, 7] +PASS: all ranks present and backend markers emitted diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/results.tsv b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/results.tsv new file mode 100644 index 0000000000000000000000000000000000000000..d9690c679d3d751b83465f63e6318c6df1d2db06 --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/results.tsv @@ -0,0 +1 @@ +config backend policy slots seq gbs cap recompute rep status iter_ms_mean iter_ms_min peak_mib run_dir diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_extension_build.log b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_extension_build.log new file mode 100644 index 0000000000000000000000000000000000000000..df5d19e50b870c32e4431a2f9ad3de340ca2a76b --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_extension_build.log @@ -0,0 +1,74 @@ +Could not find an up-to-date installation of `packaging`. License expressions might not be validated. To enforce validation, please install `packaging>=24.2`. +/usr/local/lib/python3.12/dist-packages/setuptools/config/_apply_pyprojecttoml.py:82: SetuptoolsDeprecationWarning: `project.license` as a TOML table is deprecated +!! + + ******************************************************************************** + Please use a simple string containing a SPDX expression for `project.license`. You can also use `project.license-files`. (Both options available on setuptools>=77.0.0). + + This deprecation is overdue, please update your project and remove deprecated + calls to avoid build errors in the future. + + See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details. + ******************************************************************************** + +!! + corresp(dist, value, root_dir) +/usr/local/lib/python3.12/dist-packages/setuptools/config/_apply_pyprojecttoml.py:61: SetuptoolsDeprecationWarning: License classifiers are deprecated. +!! + + ******************************************************************************** + Please consider removing the following classifiers in favor of a SPDX license expression: + + License :: OSI Approved :: BSD License + + See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details. + ******************************************************************************** + +!! + dist._finalize_license_expression() +/usr/local/lib/python3.12/dist-packages/setuptools/dist.py:759: SetuptoolsDeprecationWarning: License classifiers are deprecated. +!! + + ******************************************************************************** + Please consider removing the following classifiers in favor of a SPDX license expression: + + License :: OSI Approved :: BSD License + + See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details. + ******************************************************************************** + +!! + self._finalize_license_expression() +running build_ext +building 'megatron.core.datasets.helpers_cpp' extension +creating /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/datasets +Emitting ninja build file /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/build.ninja... +Compiling objects... +Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) +[1/1] c++ -MMD -MF /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/datasets/helpers.o.d -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O2 -Wall -fPIC -I/usr/include/python3.12 -c -c /workspace/Megatron-MoE-Echo/megatron/core/datasets/helpers.cpp -o /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/datasets/helpers.o -I/usr/include/python3.12 -I/usr/local/lib/python3.12/dist-packages/pybind11/include -O3 -Wall -std=c++17 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=helpers_cpp +In file included from /workspace/Megatron-MoE-Echo/megatron/core/datasets/helpers.cpp:12: +In function ‘pybind11::ssize_t pybind11::detail::byte_offset_unsafe(const Strides&, pybind11::ssize_t, Ix ...) [with long int Dim = 0; Strides = std::array; Ix = {}]’, + inlined from ‘const T& pybind11::detail::unchecked_reference::operator()(Ix ...) const [with Ix = {long int}; T = long int; long int Dims = 1]’ at /usr/local/lib/python3.12/dist-packages/pybind11/include/pybind11/numpy.h:545:65, + inlined from ‘const T& pybind11::detail::unchecked_reference::operator[](pybind11::ssize_t) const [with long int D = 1; = void; T = long int; long int Dims = 1]’ at /usr/local/lib/python3.12/dist-packages/pybind11/include/pybind11/numpy.h:553:26, + inlined from ‘void build_exhaustive_blending_indices(pybind11::array_t&, pybind11::array_t&, const pybind11::array_t&, int32_t)’ at /workspace/Megatron-MoE-Echo/megatron/core/datasets/helpers.cpp:67:31: +/usr/local/lib/python3.12/dist-packages/pybind11/include/pybind11/numpy.h:494:14: warning: ‘error_argmax’ may be used uninitialized [-Wmaybe-uninitialized] + 494 | return i * strides[Dim] + byte_offset_unsafe(strides, index...); + | ~~^~~~~~~~~~ +/workspace/Megatron-MoE-Echo/megatron/core/datasets/helpers.cpp: In function ‘void build_exhaustive_blending_indices(pybind11::array_t&, pybind11::array_t&, const pybind11::array_t&, int32_t)’: +/workspace/Megatron-MoE-Echo/megatron/core/datasets/helpers.cpp:49:13: note: ‘error_argmax’ was declared here + 49 | int64_t error_argmax; + | ^~~~~~~~~~~~ +creating build/lib.linux-x86_64-cpython-312/megatron/core/datasets +x86_64-linux-gnu-g++ -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O2 -Wall -shared -Wl,-O1 -Wl,-Bsymbolic-functions /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/datasets/helpers.o -L/usr/lib/x86_64-linux-gnu -o build/lib.linux-x86_64-cpython-312/megatron/core/datasets/helpers_cpp.cpython-312-x86_64-linux-gnu.so +building 'megatron.core.transformer.moe._echo_trace_C' extension +creating /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe +Emitting ninja build file /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/build.ninja... +Compiling objects... +Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) +[1/2] c++ -MMD -MF /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_ops.o.d -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O2 -Wall -fPIC -I/usr/local/lib/python3.12/dist-packages/torch/include -I/usr/local/lib/python3.12/dist-packages/torch/include/torch/csrc/api/include -I/usr/local/cuda/include -I/usr/include/python3.12 -c -c /workspace/Megatron-MoE-Echo/megatron/core/transformer/moe/echo_trace_ops.cpp -o /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_ops.o -O3 -std=c++17 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=_echo_trace_C +[2/2] /usr/local/cuda/bin/nvcc --generate-dependencies-with-compile --dependency-output /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_kernels.o.d -I/usr/local/lib/python3.12/dist-packages/torch/include -I/usr/local/lib/python3.12/dist-packages/torch/include/torch/csrc/api/include -I/usr/local/cuda/include -I/usr/include/python3.12 -c -c /workspace/Megatron-MoE-Echo/megatron/core/transformer/moe/echo_trace_kernels.cu -o /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_kernels.o -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr --compiler-options ''"'"'-fPIC'"'"'' -O3 -std=c++17 -lineinfo -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1016"' -DTORCH_EXTENSION_NAME=_echo_trace_C -gencode=arch=compute_100,code=sm_100 -gencode=arch=compute_120,code=compute_120 -gencode=arch=compute_120,code=sm_120 -gencode=arch=compute_75,code=sm_75 -gencode=arch=compute_80,code=sm_80 -gencode=arch=compute_86,code=sm_86 -gencode=arch=compute_90,code=sm_90 +creating build/lib.linux-x86_64-cpython-312/megatron/core/transformer/moe +x86_64-linux-gnu-g++ -fno-strict-overflow -Wsign-compare -DNDEBUG -g -O2 -Wall -shared -Wl,-O1 -Wl,-Bsymbolic-functions /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_kernels.o /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/megatron/core/transformer/moe/echo_trace_ops.o -L/usr/local/lib/python3.12/dist-packages/torch/lib -L/usr/local/cuda/lib64 -L/usr/lib/x86_64-linux-gnu -lc10 -ltorch -ltorch_cpu -ltorch_python -lcudart -lc10_cuda -ltorch_cuda -o build/lib.linux-x86_64-cpython-312/megatron/core/transformer/moe/_echo_trace_C.cpython-312-x86_64-linux-gnu.so +copying build/lib.linux-x86_64-cpython-312/megatron/core/datasets/helpers_cpp.cpython-312-x86_64-linux-gnu.so -> megatron/core/datasets +copying build/lib.linux-x86_64-cpython-312/megatron/core/transformer/moe/_echo_trace_C.cpython-312-x86_64-linux-gnu.so -> megatron/core/transformer/moe +load check passed: /workspace/Megatron-MoE-Echo/megatron/core/transformer/moe/_echo_trace_C.cpython-312-x86_64-linux-gnu.so -> torch.ops.megatron_echo_trace.fwd_echo_forward_begin diff --git a/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_results.tsv b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_results.tsv new file mode 100644 index 0000000000000000000000000000000000000000..9556815149db9e1382d5f6847971594ac4e95e6e --- /dev/null +++ b/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/trace_results.tsv @@ -0,0 +1,2 @@ +model backend capacity status report async_overlap_json +QWEN3_235B echo 1.083984375 OK /workspace/tierb-fixed-moonep-full/echo/tierbtrace/tierbtrace_gbs8/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375/nsys/QWEN3_235B__echo__matched__seq8192__gbs8__cap1p083984375.nsys-rep