AbhishekGhosh1998 commited on
Commit
385e09f
·
verified ·
1 Parent(s): 124df86

granularity results granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace @ 2026-09-21T07:19:26Z from sp-aiinfra-interactive-aghosh-8-mb99k-8xb200-moe-cg-0

Browse files
.gitattributes CHANGED
@@ -109,3 +109,4 @@ granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K6_
109
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K10__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K10__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
110
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K16__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K16__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
111
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K8__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K8__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
 
 
109
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K10__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K10__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
110
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K16__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K16__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
111
  granularity_next_b200/tiera_topk_trace/moonep/tierbtrace/tierbtrace_gbs16/G4_K8__moonep__matched__seq4096__gbs16__capnone__async/nsys/G4_K8__moonep__matched__seq4096__gbs16__capnone__async.nsys-rep filter=lfs diff=lfs merge=lfs -text
112
+ granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875.nsys-rep filter=lfs diff=lfs merge=lfs -text
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/manifest.txt ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ timestamp: 2026-09-21T07:12:54+00:00
2
+ host: sp-aiinfra-interactive-aghosh-8-mb99k-8xb200-moe-cg-0
3
+ gpu: NVIDIA B200 x8
4
+ driver: 580.95.05
5
+ study_scripts_branch: granularity-echo
6
+ study_scripts_commit: b8bd100
7
+ campaign: echo
8
+ backend: echo
9
+ phase: tierbtrace
10
+ gbs: 8
11
+ seq: 4096
12
+ configs: G1 G2 G3 G4
13
+ capacity: 1.25
14
+ repeats: 3
15
+ nsys: 1
16
+ megatron_repo: /workspace/Megatron-MoE-Echo
17
+ megatron_branch_required: bracing-regions-through-kernels
18
+ megatron_branch_actual: bracing-regions-through-kernels
19
+ megatron_commit: 9a236793
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/pip_freeze.txt ADDED
@@ -0,0 +1,317 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ absl-py==2.3.0
2
+ aiohappyeyeballs @ file:///rapids/aiohappyeyeballs-2.6.1-py3-none-any.whl#sha256=f349ba8f4b75cb25c99c5c2d84e997e485204d2902a9597802b0371f09331fb8
3
+ aiohttp @ file:///rapids/aiohttp-3.12.7-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=94f98e0e5a49f89b252e115844f756c04fc8050f38252a32a3dd994ce8121f10
4
+ aiosignal @ file:///rapids/aiosignal-1.3.2-py2.py3-none-any.whl#sha256=45cde58e409a301715980c2b01d0c28bdde3770d8290b5eb2173759d9acb31a5
5
+ annotated-types==0.7.0
6
+ anyio==4.9.0
7
+ apex @ file:///opt/pytorch/apex/dist/apex-0.1-cp312-cp312-linux_x86_64.whl#sha256=a1508e4a386362e9c8f46c92ccada9110e421212b964c694235d0fd5a224b170
8
+ argon2-cffi==25.1.0
9
+ argon2-cffi-bindings==21.2.0
10
+ arrow==1.3.0
11
+ asciitree @ file:///rapids/asciitree-0.3.3-py3-none-any.whl#sha256=3e507ccfaeb417823158091fdb9d26b0205f6a59ba680b479a71db65c73d01d4
12
+ asttokens==3.0.0
13
+ astunparse==1.6.3
14
+ async-lru==2.0.5
15
+ attrs==25.3.0
16
+ audioread==3.0.1
17
+ babel==2.17.0
18
+ beautifulsoup4==4.13.4
19
+ black==25.1.0
20
+ bleach==6.2.0
21
+ blis==0.7.11
22
+ cachetools @ file:///rapids/cachetools-6.0.0-py3-none-any.whl#sha256=82e73ba88f7b30228b5507dce1a1f878498fc669d972aef2dde4f3a3c24f103e
23
+ catalogue==2.0.10
24
+ certifi==2025.4.26
25
+ cffi==1.17.1
26
+ charset-normalizer==3.4.2
27
+ click==8.5.0
28
+ cloudpathlib==0.21.1
29
+ cloudpickle @ file:///rapids/cloudpickle-3.1.1-py3-none-any.whl#sha256=c8c5a44295039331ee9dad40ba100a9c7297b6f988e50e87ccdf3765a668350e
30
+ cmake==3.31.6
31
+ comm==0.2.2
32
+ confection==0.1.5
33
+ contourpy==1.3.2
34
+ cuda-bindings @ file:///rapids/cuda_bindings-12.9.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=ff0e28d1e34758654b9c961e1f55e4786e49aee6a4dbceaf3cc24c46c672df7e
35
+ cuda-python @ file:///rapids/cuda_python-12.9.0-py3-none-any.whl#sha256=926acba49b2c0a0374c61b7c98f337c085199cf51cdfe4d6423c4129c20547a7
36
+ cudf @ file:///rapids/cudf-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b779a8e8c0e8577b9b560895958496da211bfac49cdc4f11a49e8d18cb82e348
37
+ cudf-polars @ file:///rapids/cudf_polars-25.4.0-py3-none-any.whl#sha256=227b6c5147e93ea86282c8f7219118f916769fa2ca0cdd931e97a148aa66b869
38
+ cugraph @ file:///rapids/cugraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b837c2dc2f7b235c702e4f8df64479b8fcbb342f848516d5faf0cdc4b5f31f29
39
+ cugraph-service-client @ file:///rapids/cugraph_service_client-25.4.0-py3-none-any.whl#sha256=7e568c3e41154b9752f5173555ffcc282ac52d72343feeb79653b2e2b1232470
40
+ cugraph-service-server @ file:///rapids/cugraph_service_server-25.4.0-py3-none-any.whl#sha256=a2830dcffafcccb16e12ebf85b98f8f247869d38c01ab4dd8664a1e7439d31be
41
+ cuml @ file:///rapids/cuml-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=1e90a464a18b094e24797f1676688c5d7696c4cf97c934504214a53d5aa09aa5
42
+ cupy-cuda12x @ file:///rapids/cupy_cuda12x-13.3.0-cp312-cp312-linux_x86_64.whl#sha256=6a2dfa0aa1e8189a10d1534276c08f1d2f73b747fd28bfbe7b9f4ef66d9af73f
43
+ cuvs @ file:///rapids/cuvs-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b12f6b9f80e053a3b38542c2947d2516e468c02cad54a13d8b1a4a61d5aa7914
44
+ cycler==0.12.1
45
+ cymem==2.0.11
46
+ Cython==3.1.1
47
+ dask @ file:///rapids/dask-2025.2.0-py3-none-any.whl#sha256=f0fdeef6ceb0a06569d456c9e704f220f7f54e80f3a6ea42ab98cea6bc642b6e
48
+ dask-cuda @ file:///rapids/dask_cuda-25.4.0-py3-none-any.whl#sha256=442d87da0a95d00ca2a262b64c76e785a32ad905c82000fa2227932f81390e19
49
+ dask-cudf @ file:///rapids/dask_cudf-25.4.0-py3-none-any.whl#sha256=5e7cea0b26daea17da6e0bb457435e24460f76ba740358a3a18e239d6015c95c
50
+ datasets==5.0.1
51
+ debugpy==1.8.14
52
+ decorator==5.2.1
53
+ -e git+https://github.com/deepseek-ai/DeepEP.git@34152ae28f80bcc3ee38d7a12cb2ad87cfd4ea72#egg=deep_ep
54
+ defusedxml==0.7.1
55
+ dill==0.4.1
56
+ distributed @ file:///rapids/distributed-2025.2.0-py3-none-any.whl#sha256=368462084be47092d0a835f6c64c9b852da64dc88302b76d5f93bd5131228c67
57
+ distributed-ucxx @ file:///rapids/distributed_ucxx-0.43.0-py3-none-any.whl#sha256=94f60d0ccb0c94ba8dbfa1eaf455874d18aaab17e2ae218c38590e9fdc2d721d
58
+ dm-tree==0.1.9
59
+ einops==0.8.1
60
+ execnet==2.1.1
61
+ executing==2.2.0
62
+ expecttest==0.3.0
63
+ fasteners @ file:///rapids/fasteners-0.19-py3-none-any.whl#sha256=758819cb5d94cdedf4e836988b74de396ceacb8e2794d21f82d131fd9ee77237
64
+ fastjsonschema==2.21.1
65
+ 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
66
+ filelock==3.18.0
67
+ flash_attn @ file:///opt/pytorch/flash_attn-2.7.4.post1-cp312-cp312-linux_x86_64.whl#sha256=a10455d6c56b387785b7413fd0bbcb392b42b3d4f4b88fda5d86cc8a9ce9182a
68
+ fonttools==4.58.1
69
+ fqdn==1.5.1
70
+ 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
71
+ fsspec==2025.5.1
72
+ gast==0.6.0
73
+ grpcio==1.62.1
74
+ h11==0.16.0
75
+ hf-xet==1.6.0
76
+ httpcore==1.0.9
77
+ httpx==0.28.1
78
+ huggingface_hub==1.32.0
79
+ hypothesis==6.130.8
80
+ idna==3.10
81
+ importlib_metadata==8.7.0
82
+ iniconfig==2.1.0
83
+ intel-openmp==2021.4.0
84
+ ipykernel==6.29.5
85
+ ipython==9.3.0
86
+ ipython_pygments_lexers==1.1.1
87
+ isoduration==20.11.0
88
+ isort==6.0.1
89
+ jedi==0.19.2
90
+ Jinja2==3.1.6
91
+ joblib==1.5.1
92
+ json5==0.12.0
93
+ jsonpointer==3.0.0
94
+ jsonschema==4.24.0
95
+ jsonschema-specifications==2025.4.1
96
+ jupyter-events==0.12.0
97
+ jupyter-lsp==2.2.5
98
+ jupyter_client==8.6.3
99
+ jupyter_core==5.8.1
100
+ jupyter_server==2.16.0
101
+ jupyter_server_terminals==0.5.3
102
+ jupyterlab==4.4.3
103
+ jupyterlab_code_formatter==3.0.2
104
+ jupyterlab_pygments==0.3.0
105
+ jupyterlab_server==2.27.3
106
+ jupyterlab_tensorboard_pro @ file:///builder/jupyterlab_tensorboard_pro-4.0.0-py2.py3-none-any.whl#sha256=6bfcb5ced3358eb72150f4b2ab91312faef2f0ca14d845e8379d8fa52df6902e
107
+ jupytext==1.17.2
108
+ kiwisolver==1.4.8
109
+ kvikio @ file:///rapids/kvikio-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=8b40fd8ea9e4a2974c04b50b7fc26b06094ac807f9a233281d0bafbaa71b1d85
110
+ langcodes==3.5.0
111
+ language_data==1.3.0
112
+ lazy_loader==0.4
113
+ libcudf @ file:///rapids/libcudf-25.4.0-py3-none-linux_x86_64.whl#sha256=238320b4958c1737bba2a609673e2539a78216b811b5498a68e00395d9b021a4
114
+ libcugraph @ file:///rapids/libcugraph-25.4.0-py3-none-linux_x86_64.whl#sha256=29085e901b2f7569ff2653bb1355fe6fad20f2fcaa585f5649da9867d8df766d
115
+ libcuml @ file:///rapids/libcuml-25.4.0-py3-none-linux_x86_64.whl#sha256=8e2e444c72b0e995e15a48da37ff12fb410a47a2baa69ba86ccc9b1a3db91e8c
116
+ libcuvs @ file:///rapids/libcuvs-25.4.0-py3-none-linux_x86_64.whl#sha256=e2f8db61c92b769c590dd752f43439ddd62b2f6592a3c337362b6c1ef220c3b9
117
+ libkvikio @ file:///rapids/libkvikio-25.4.0-py3-none-linux_x86_64.whl#sha256=da8a5713b05fc478bf40529a3e0fb61a2d605bb5a70ca4d461f96a6005dbe8b7
118
+ libraft @ file:///rapids/libraft-25.4.0-py3-none-linux_x86_64.whl#sha256=b921dec005aa4a321dd156776a18376feb47c96a5b67eec125c8a8e19edd1e24
119
+ librmm @ file:///rapids/librmm-25.4.0-py3-none-linux_x86_64.whl#sha256=ff951811bda729d4fa1109ab906225ae1f489a6ec2eeab5d1b674cc34bd4422f
120
+ librmm-cu12 @ file:///rapids/librmm_cu12-25.4.0-py3-none-any.whl#sha256=29296ef2b0e90679a0db280e86015be12d6ab9f4e62a181aa085f62dbadd06b3
121
+ librosa==0.11.0
122
+ libucxx @ file:///rapids/libucxx-0.43.0-py3-none-linux_x86_64.whl#sha256=f15849d7ac2fb68f0fd8998838a13321ddcfb0c594f7e6cbc401a3bcc4750566
123
+ lightning-thunder @ file:///opt/pytorch/lightning-thunder
124
+ lightning-utilities==0.14.3
125
+ lintrunner==0.12.7
126
+ llvmlite==0.42.0
127
+ locket @ file:///rapids/locket-1.0.0-py2.py3-none-any.whl#sha256=b6c819a722f7b6bd955b80781788e4a66a55628b858d347536b7e81325a3a5e3
128
+ looseversion==1.3.0
129
+ marisa-trie==1.2.1
130
+ Markdown==3.8
131
+ markdown-it-py==3.0.0
132
+ MarkupSafe==3.0.2
133
+ matplotlib==3.10.3
134
+ matplotlib-inline==0.1.7
135
+ mdit-py-plugins==0.4.2
136
+ mdurl==0.1.2
137
+ -e git+ssh://git@github.com/sriram-17-17/my-Megatron-MoE-Echo.git@9a23679390d7ab1fe462bc61fa31d90d393db49c#egg=megatron_core
138
+ mistune==3.1.3
139
+ mkl==2021.1.1
140
+ mkl-devel==2021.1.1
141
+ mkl-include==2021.1.1
142
+ ml_dtypes==0.6.0
143
+ mock==5.2.0
144
+ -e git+ssh://git@github.com/sriram-17-17/MoonEP.git@d8bc6667fad7177e6e09b637d773fa67f08bae93#egg=moonep
145
+ mpmath==1.3.0
146
+ msgpack==1.1.0
147
+ multidict @ file:///rapids/multidict-6.4.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=a920f9cf2abdf6e493c519492d892c362007f113c94da4c239ae88429835bad1
148
+ multiprocess==0.70.19
149
+ murmurhash==1.0.13
150
+ mypy_extensions==1.1.0
151
+ nbclient==0.10.2
152
+ nbconvert==7.16.6
153
+ nbformat==5.10.4
154
+ nest-asyncio==1.6.0
155
+ networkx==3.5
156
+ ninja==1.11.1.4
157
+ notebook==7.4.3
158
+ notebook_shim==0.2.4
159
+ numba==0.59.1
160
+ numba-cuda @ file:///rapids/numba_cuda-0.4.0-py3-none-any.whl#sha256=5b95a5f42068771e37142944d44fe844fa87f1b2f89051461a4cb0af91bcc28a
161
+ numcodecs @ file:///rapids/numcodecs-0.13.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=da2230484e6102e5fa3cc1a5dd37ca1f92dfbd183d91662074d6f7574e3e8f53
162
+ numpy==1.26.4
163
+ nvdlfw_inspect @ git+https://github.com/NVIDIA/nvidia-dlfw-inspect.git@6b60eb0e675606fb2fbbfcb12667ecbec75eaf87
164
+ nvfuser==0.2.27a0+9bf5aca
165
+ nvidia-cuda-cccl-cu12==12.9.27
166
+ nvidia-cudnn-frontend @ file:///opt/pytorch/lightning-thunder/tmp/cudnn_frontend
167
+ nvidia-cutlass-dsl==4.4.2
168
+ nvidia-cutlass-dsl-libs-base==4.4.2
169
+ nvidia-dali-cuda120==1.50.0
170
+ nvidia-ml-py @ file:///rapids/nvidia_ml_py-12.575.51-py3-none-any.whl#sha256=eb8641800d98ce40a22f479873f34b482e214a7e80349c63be51c3919845446e
171
+ nvidia-modelopt==0.29.0
172
+ nvidia-modelopt-core==0.29.0
173
+ nvidia-nvcomp-cu12==4.2.0.14
174
+ nvidia-nvimgcodec-cu12==0.5.0.13
175
+ nvidia-nvjpeg-cu12==12.4.0.16
176
+ nvidia-nvjpeg2k-cu12==0.8.1.40
177
+ nvidia-nvshmem-cu12==3.7.2
178
+ nvidia-nvtiff-cu12==0.5.0.67
179
+ nvidia-resiliency-ext==0.4.0
180
+ nvtx @ file:///rapids/nvtx-0.2.11-cp312-cp312-linux_x86_64.whl#sha256=73521e37cba42b4db0b3cd68fc5eb814e55c500a61eeed6edcdf6746315ac53e
181
+ nx-cugraph @ file:///rapids/nx_cugraph-25.4.0-py3-none-any.whl#sha256=e625444486cc899a77bd48e7da76a2edd95c0006dd95fbf323745feae7f33111
182
+ onnx @ file:///opt/pytorch/pytorch/third_party/onnx
183
+ onnx-ir==1.0.0
184
+ onnxscript==0.7.2
185
+ opt_einsum==3.4.0
186
+ optree==0.16.0
187
+ overrides==7.7.0
188
+ packaging==23.2
189
+ pandas @ file:///rapids/pandas-2.2.3-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=fffb8ae78d8af97f849404f21411c95062db1496aeb3e56f146f0355c9989319
190
+ pandocfilters==1.5.1
191
+ parso==0.8.4
192
+ partd @ file:///rapids/partd-1.4.2-py3-none-any.whl#sha256=978e4ac767ec4ba5b86c6eaa52e5a2a3bc748a2ca839e8cc798f1cc6ce6efb0f
193
+ pathspec==0.12.1
194
+ pexpect==4.9.0
195
+ pillow==11.2.1
196
+ platformdirs==4.3.8
197
+ pluggy==1.6.0
198
+ ply @ file:///rapids/ply-3.11-py2.py3-none-any.whl#sha256=096f9b8350b65ebd2fd1346b12452efe5b9607f7482813ffca50c22722a807ce
199
+ polars @ file:///rapids/polars-1.25.2-cp39-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=f7fcbb4f476784384ccda48757fca4e8c2e2c5a0a3aef3717aaf56aee4e30e09
200
+ polygraphy==0.49.20
201
+ pooch==1.8.2
202
+ preshed==3.0.10
203
+ prometheus_client==0.22.1
204
+ prompt_toolkit==3.0.51
205
+ propcache @ file:///rapids/propcache-0.3.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=563f9d8c03ad645597b8d010ef4e9eab359faeb11a0a2ac9f7b4bc8c28ebef25
206
+ protobuf==4.24.4
207
+ psutil==7.0.0
208
+ ptyprocess==0.7.0
209
+ PuLP==3.2.1
210
+ pure_eval==0.2.3
211
+ pyarrow==25.0.1
212
+ pybind11==2.13.6
213
+ pybind11_global==2.13.6
214
+ pycocotools @ git+https://github.com/nvidia/cocoapi.git@08350a157b41f0755a536441b5abec89ab934c5b#subdirectory=PythonAPI
215
+ pycparser==2.22
216
+ pydantic==2.11.5
217
+ pydantic_core==2.33.2
218
+ Pygments==2.19.1
219
+ pylibcudf @ file:///rapids/pylibcudf-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=401b8bfd022d06f84e120779ca7c3d01a28b68762aaa8f7873b10d6b1975cf6a
220
+ pylibcugraph @ file:///rapids/pylibcugraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=061c7389a88c815b54fd5d3a3584fb87bf693e5602f5be335260d0ba575ba557
221
+ pylibcugraphops @ file:///rapids/pylibcugraphops-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=a25ef4b22166c80718d5c7e5bddb755e81a2fe340fa5becd0ac36ec3edfb208a
222
+ pylibraft @ file:///rapids/pylibraft-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=f2ecd23fe51f9c2caa87bbbe85cb32fcaff4ecfb52ceb9a9991bc50f82782c2e
223
+ pylibwholegraph @ file:///rapids/pylibwholegraph-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=b82913d1e38bda76d8c272fbd8cc9072d987209d3d47cc93a464a37c2925aca2
224
+ pynvjitlink @ file:///rapids/pynvjitlink-0.3.0-cp312-cp312-linux_x86_64.whl#sha256=066caf1314730643d3b48c9847bd65789defdd6a7c78993176eb3d8bbe3f4919
225
+ pynvml @ file:///rapids/pynvml-12.0.0-py3-none-any.whl#sha256=fdff84b62a27dbe98e08e1a647eb77342bef1aebe0878bcd15e99a83fcbecb9e
226
+ pyparsing==3.2.3
227
+ pytest==8.1.1
228
+ pytest-flakefinder==1.1.0
229
+ pytest-rerunfailures==15.1
230
+ pytest-shard==0.1.2
231
+ pytest-xdist==3.7.0
232
+ python-dateutil==2.9.0.post0
233
+ python-hostlist==2.2.1
234
+ python-json-logger==3.3.0
235
+ pytorch-triton @ file:///tmp/dist/pytorch_triton-3.3.0%2Bgit96316ce52.nvinternal-cp312-cp312-linux_x86_64.whl#sha256=6d6921772b9dfd0243c055e0d054de42a113e92cabcfdd1d3e80a5ad9c08cf7d
236
+ pytz @ file:///rapids/pytz-2023.4-py2.py3-none-any.whl#sha256=f90ef520d95e7c46951105338d918664ebfd6f1d995bd7d153127ce90efafa6a
237
+ PyYAML==6.0.2
238
+ pyzmq==26.4.0
239
+ raft-dask @ file:///rapids/raft_dask-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=36cf37ba6ce35454c4729c2cb597a360f22780eefb11d514a16a07137da5eb6c
240
+ rapids-dask-dependency @ file:///rapids/rapids_dask_dependency-25.4.0a0-py3-none-any.whl#sha256=484faca626e7760112a7d8ad3426f5bdd720b609006d8dca87c2607076a348bd
241
+ rapids-logger @ file:///rapids/rapids_logger-0.1.18-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl#sha256=3617b2d86291ac6cb957542895f20fc9737ace29773ca90337e6fb04c73bd120
242
+ referencing==0.36.2
243
+ regex==2024.11.6
244
+ requests==2.32.3
245
+ rfc3339-validator==0.1.4
246
+ rfc3986-validator==0.1.1
247
+ rich==14.0.0
248
+ rmm @ file:///rapids/rmm-25.4.0-cp312-cp312-linux_x86_64.whl#sha256=1824bcab3836296666980a997ddcb9c4bca24dfb873012cfa0a7ecea583ca712
249
+ rpds-py==0.25.1
250
+ safetensors==0.5.3
251
+ scikit-learn==1.6.1
252
+ scipy==1.15.3
253
+ Send2Trash==1.8.3
254
+ setuptools==78.1.1
255
+ shellingham==1.5.4
256
+ six==1.16.0
257
+ smart-open==7.1.0
258
+ sniffio==1.3.1
259
+ sortedcontainers==2.4.0
260
+ soundfile==0.13.1
261
+ soupsieve==2.7
262
+ soxr==0.5.0.post1
263
+ spacy==3.7.5
264
+ spacy-legacy==3.0.12
265
+ spacy-loggers==1.0.5
266
+ srsly==2.5.1
267
+ stack-data==0.6.3
268
+ sympy==1.14.0
269
+ tabulate==0.9.0
270
+ tbb==2021.13.1
271
+ tblib @ file:///rapids/tblib-3.1.0-py3-none-any.whl#sha256=670bb4582578134b3d81a84afa1b016128b429f3d48e6cbbaecc9d15675e984e
272
+ tensorboard==2.16.2
273
+ tensorboard-data-server==0.7.2
274
+ tensorrt @ file:///workspace/TensorRT-10.11.0.33/python/tensorrt-10.11.0.33-cp312-none-linux_x86_64.whl#sha256=c2f18f72758a3d54a71901495c8fda9e7094a8bcdd2a5b6b58967fbf05733b3a
275
+ terminado==0.18.1
276
+ thinc==8.2.5
277
+ threadpoolctl==3.6.0
278
+ thriftpy2 @ file:///rapids/thriftpy2-0.5.2-cp312-cp312-linux_x86_64.whl#sha256=c0c7d939bdc5ccf7739154959938445d5fdf72f1d6e12bd563947a36f67e2147
279
+ tinycss2==1.4.0
280
+ toolz @ file:///rapids/toolz-1.0.0-py3-none-any.whl#sha256=292c8f1c4e7516bf9086f8850935c799a874039c8bcf959d47b600e4c44a6236
281
+ torch @ file:///opt/transfer/torch-2.8.0a0%2B5228986c39.nv25.6-29946474-cp312-cp312-linux_x86_64.whl#sha256=7a3c9ff5288d8102a31296b3d62491f3f47c8820a5c9aaefc7dc0fe20e1e8ad5
282
+ torch_tensorrt @ file:///opt/pytorch/torch_tensorrt/dist/torch_tensorrt-2.8.0a0-cp312-cp312-linux_x86_64.whl#sha256=26f4d7a04e7e6e017a5b917b8b98b83720bb2bb0195f48d25f52d567d286bfe1
283
+ torchao @ file:///opt/pytorch/ao
284
+ torchprofile==0.0.4
285
+ torchvision @ file:///opt/transfer/torchvision-0.22.0a0%2B95f10a4e-29946474-cp312-cp312-manylinux_2_28_x86_64.whl#sha256=746a3c674cc61680a86f4a14b374b21af1b5c21f629c42765b40c9494f6b234d
286
+ tornado==6.5.1
287
+ tqdm==4.67.1
288
+ traitlets==5.14.3
289
+ transformer_engine @ file:///pvcdata/aiinfra/b-aghosh/wheels/transformer_engine-2.10.0.dev0%2B10f6128d-cp312-cp312-linux_x86_64.whl#sha256=35cf5f4b83b2baad864e12e512bb422e0929409ac05610812fa521d4942d0189
290
+ treelite @ file:///rapids/treelite-4.4.1-py3-none-linux_x86_64.whl#sha256=174741724205540a1a5b37da81483aeff55d89548f42b3461cf07acac7ae96eb
291
+ typer==0.16.0
292
+ types-dataclasses==0.6.6
293
+ types-python-dateutil==2.9.0.20250516
294
+ typing-inspection==0.4.1
295
+ typing_extensions==4.14.0
296
+ tzdata @ file:///rapids/tzdata-2025.2-py2.py3-none-any.whl#sha256=1a403fada01ff9221ca8044d701868fa132215d84beb92242d9acd2147f667a8
297
+ ucx-py @ file:///rapids/ucx_py-0.43.0-cp312-cp312-linux_x86_64.whl#sha256=4a452fd8d8c983283ada3d49086e3657dc2ee94fbad08efdd708d15ab6e0b73a
298
+ ucxx @ file:///rapids/ucxx-0.43.0-cp312-cp312-linux_x86_64.whl#sha256=f1b4e55cb45c66fe88a2d8f9c2b87751165f3cfec21fbdc9c47c4ddff75e8a5f
299
+ -e git+https://github.com/Dots-Infra/UltraEP.git@94cab099b44fffa99a82fea99e7c12d89cf65e4f#egg=ultra_ep
300
+ uri-template==1.3.0
301
+ urllib3==2.0.7
302
+ wasabi==1.1.3
303
+ wcwidth==0.2.13
304
+ weasel==0.4.1
305
+ webcolors==24.11.1
306
+ webencodings==0.5.1
307
+ websocket-client==1.8.0
308
+ Werkzeug==3.1.3
309
+ wheel==0.45.1
310
+ wrapt==1.17.2
311
+ xdoctest==1.0.2
312
+ xgboost @ file:///rapids/xgboost-2.1.4-py3-none-manylinux_2_39_x86_64.whl#sha256=25eeca27f4048fc491940b3a0b4b5782ee48c33cb3aec059cef71f8da542930b
313
+ xxhash==4.0.1
314
+ yarl @ file:///rapids/yarl-1.20.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl#sha256=bf099e2432131093cc611623e0b0bcc399b8cddd9a91eded8bfb50402ec35018
315
+ zarr @ file:///rapids/zarr-2.18.7-py3-none-any.whl#sha256=ac3dc4033e9ae4e9d7b5e27c97ea3eaf1003cc0a07f010bd83d5134bf8c4b223
316
+ zict @ file:///rapids/zict-3.0.0-py2.py3-none-any.whl#sha256=5796e36bd0e0cc8cf0fbc1ace6a68912611c1dbd74750a3f3026b9b9d6a327ae
317
+ zipp==3.22.0
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/provenance.txt CHANGED
@@ -27,3 +27,64 @@ command : run_ladder.sh --phase tierbtrace --gbs 8 --out /workspace/granul
27
  7, 120 MHz, 1965 MHz, 1000.00 W, 29
28
 
29
  NVLink topology:
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
27
  7, 120 MHz, 1965 MHz, 1000.00 W, 29
28
 
29
  NVLink topology:
30
+ GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 GPU7 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 NIC8 CPU Affinity NUMA Affinity GPU NUMA ID
31
+ GPU0 X NV18 NV18 NV18 NV18 NV18 NV18 NV18 PIX NODE SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A
32
+ GPU1 NV18 X NV18 NV18 NV18 NV18 NV18 NV18 NODE PIX SYS SYS SYS SYS SYS SYS NODE 0-27 0 N/A
33
+ GPU2 NV18 NV18 X NV18 NV18 NV18 NV18 NV18 SYS SYS PIX NODE SYS SYS SYS SYS SYS 28-55 1 N/A
34
+ GPU3 NV18 NV18 NV18 X NV18 NV18 NV18 NV18 SYS SYS NODE PIX SYS SYS SYS SYS SYS 28-55 1 N/A
35
+ GPU4 NV18 NV18 NV18 NV18 X NV18 NV18 NV18 SYS SYS SYS SYS PIX NODE SYS SYS SYS 56-83 2 N/A
36
+ GPU5 NV18 NV18 NV18 NV18 NV18 X NV18 NV18 SYS SYS SYS SYS NODE PIX SYS SYS SYS 56-83 2 N/A
37
+ GPU6 NV18 NV18 NV18 NV18 NV18 NV18 X NV18 SYS SYS SYS SYS SYS SYS PIX NODE SYS 84-111 3 N/A
38
+ GPU7 NV18 NV18 NV18 NV18 NV18 NV18 NV18 X SYS SYS SYS SYS SYS SYS NODE PIX SYS 84-111 3 N/A
39
+ NIC0 PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS SYS NODE
40
+ NIC1 NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS SYS NODE
41
+ NIC2 SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS SYS SYS
42
+ NIC3 SYS SYS NODE PIX SYS SYS SYS SYS SYS SYS NODE X SYS SYS SYS SYS SYS
43
+ NIC4 SYS SYS SYS SYS PIX NODE SYS SYS SYS SYS SYS SYS X NODE SYS SYS SYS
44
+ unavailable
45
+
46
+ ---- toolchain -----------------------------------------------
47
+ nvcc : release 12.9, V12.9.86
48
+ gcc : gcc (Ubuntu 13.3.0-6ubuntu2~24.04) 13.3.0
49
+ python : 3.12.3
50
+ torch : 2.8.0a0+5228986c39.nv25.06 cuda=12.9 nccl=2.27.3
51
+ torch arch : ['sm_75', 'sm_80', 'sm_86', 'sm_90', 'sm_100', 'sm_120', 'compute_120']
52
+ transformer_engine: 2.10.0.dev0+10f6128d [/usr/local/lib/python3.12/dist-packages/transformer_engine/__init__.py]
53
+ deep_ep : 1.2.1+34152ae [/workspace/DeepEP/deep_ep/__init__.py]
54
+ ultra_ep : 1.0.0 [/workspace/UltraEP/ultra_ep/__init__.py]
55
+ moonep : 0.0.1 [/workspace/MoonEP/moonep/__init__.py]
56
+ megatron : 0.14.0rc5 [None]
57
+ flash_attn : 2.7.4.post1 [/usr/local/lib/python3.12/dist-packages/flash_attn/__init__.py]
58
+ numpy : 1.26.4 [/usr/local/lib/python3.12/dist-packages/numpy/__init__.py]
59
+ huggingface_hub: 1.32.0 [/usr/local/lib/python3.12/dist-packages/huggingface_hub/__init__.py]
60
+
61
+ ---- repositories --------------------------------------------
62
+ Megatron-MoE-Echo bracing-regions-through-kernels 9a23679390d7
63
+ Megatron-UltraEP ultraep-better-memory-dump-support c3a8a8a7b69d
64
+ UltraEP main 94cab099b44f
65
+ DeepEP HEAD 34152ae28f80 (+3 untracked: build output)
66
+ MoonEP megatron-ddp-symmetric-params d8bc6667fad7
67
+ blackwell_run_scripts granularity-echo b8bd10097bb5
68
+
69
+ ---- environment ---------------------------------------------
70
+ CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0
71
+ CUDA_DRIVER_VERSION=575.57.08
72
+ CUDA_HOME=/usr/local/cuda
73
+ CUDA_MODULE_LOADING=LAZY
74
+ CUDA_VERSION=12.9.1.010
75
+ HF_REMOTE_ROOT=granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4
76
+ NCCL_VERSION=2.27.3
77
+ NVSHMEM_VERSION=3.2.5
78
+ PYTORCH_BUILD_NUMBER=0
79
+ PYTORCH_BUILD_VERSION=2.8.0a0+5228986
80
+ PYTORCH_HOME=/opt/pytorch/pytorch
81
+ PYTORCH_VERSION=2.8.0a0+5228986
82
+ TORCH_ALLOW_TF32_CUBLAS_OVERRIDE=1
83
+ TORCH_CUDA_ARCH_LIST=7.5 8.0 8.6 9.0 10.0 12.0+PTX
84
+ TORCH_NCCL_USE_COMM_NONBLOCKING=0
85
+
86
+ ---- shared library resolution -------------------------------
87
+ libnvshmem_host.so: not in ldconfig
88
+ libnccl.so: /lib/x86_64-linux-gnu/libnccl.so.2
89
+ lrwxrwxrwx 1 root root 20 Sep 21 04:57 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so -> libnvshmem_host.so.3
90
+ -rw-r--r-- 1 root root 179045744 Sep 21 04:54 /usr/local/lib/python3.12/dist-packages/nvidia/nvshmem/lib/libnvshmem_host.so.3
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875.nsys-rep ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10c45408628ad43efc4035650d3c6060cdf30c47810a324c577cf3c332d278f8
3
+ size 3670203
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/run.env ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ backend=echo
2
+ config=DEEPSEEK_V3_L4
3
+ family=tierB
4
+ source_model=DeepSeek-V3
5
+ standard_num_layers=61
6
+ base_rung=
7
+ repo=/workspace/Megatron-MoE-Echo
8
+ branch=bracing-regions-through-kernels
9
+ commit=9a236793
10
+ num_layers=4
11
+ hidden=7168
12
+ ffn_hidden=18432
13
+ moe_ffn_hidden=2048
14
+ num_experts=256
15
+ topk=8
16
+ experts_per_rank=32
17
+ expert_params=44040192
18
+ heads=128
19
+ kv_groups=128
20
+ kv_channels=128
21
+ qk_layernorm=1
22
+ shared_ffn=2048
23
+ shared_expert_overlap=0
24
+ moe_token_dispatcher=flex
25
+ mla=1
26
+ q_lora_rank=1536
27
+ kv_lora_rank=512
28
+ qk_head_dim=128
29
+ qk_pos_emb_head_dim=64
30
+ v_head_dim=128
31
+ dense_first_layers=3
32
+ moe_layers=1
33
+ rotary_base=10000
34
+ rotary_percent=1.0
35
+ rotary_scaling=40
36
+ mscale=1.0
37
+ mscale_all_dim=1.0
38
+ qkv_bias=0
39
+ router_balancing=seq_aux_loss
40
+ aux_loss_coeff=1e-4
41
+ router_pre_softmax=0
42
+ router_fusion=0
43
+ router_dtype=fp32
44
+ router_groups=8
45
+ router_group_topk=4
46
+ router_score_function=sigmoid
47
+ router_expert_bias=1
48
+ router_bias_update_rate=1e-3
49
+ router_topk_scaling_factor=2.5
50
+ shared_expert_gate=0
51
+ apply_layernorm_1p=0
52
+ attention_output_gate=0
53
+ linear_attention_type=
54
+ linear_attention_freq=
55
+ linear_conv_kernel_dim=
56
+ linear_key_head_dim=
57
+ linear_value_head_dim=
58
+ linear_num_key_heads=
59
+ linear_num_value_heads=
60
+ mtp_num_layers=0
61
+ mtp_loss_scaling_factor=0.1
62
+ no_weight_decay_cond_type=
63
+ max_position_embeddings=4096
64
+ init_method_std=0.02
65
+ device_initiated_grouped_gemm=1
66
+ norm_epsilon=1e-6
67
+ seq_length=4096
68
+ micro_batch_size=1
69
+ global_batch_size=8
70
+ num_microbatches=1
71
+ assignments_per_rank=32768
72
+ mean_tokens_per_expert=1024
73
+ slot_policy=matched
74
+ slots_per_rank=32
75
+ capacity_mode=static
76
+ capacity=1.13671875
77
+ cuda_graph=1
78
+ probe_mode=0
79
+ async_grad_reduce=0
80
+ branch_role=timing
81
+ memory_dump_mode=none
82
+ memory_snapshot_path=
83
+ cuda_graph_warmup_steps=3
84
+ recompute=none
85
+ force_balance=0
86
+ train_iters=10
87
+ ep_size=8
88
+ echo_dispatch_overlap=1
89
+ echo_dispatch_recompute=0
90
+ ultraep_quota_min_tokens=
91
+ ultraep_weight_sync_join=
92
+ profile_enable=1
93
+ kernel_trace=1
94
+ fp8_mode=mxfp8
95
+ fp8_recipe=mxfp8
96
+ training_seed=1234
97
+ learning_rate=1e-4
98
+ min_learning_rate=1.0e-5
99
+ lr_warmup_iters=5
100
+ state_dump_enable=0
101
+ state_dump_dir=
102
+ moonep_assert_no_pending_reduce=0
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/train.log ADDED
@@ -0,0 +1,1047 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [07:17:00] WARN: bracing-regions-through-kernels lacks --moe-router-fusion; DEEPSEEK_V3_L4 runs unfused (timing only, routing identical)
2
+ ==============================================================
3
+ DEEPSEEK_V3_L4 / echo (slots/rank=32 policy=matched)
4
+ E=256 K=8 H'=2048 m=32
5
+ L=4 seq=4096 gbs=8 (1 microbatches)
6
+ cap_mode=static cap=1.13671875 cg=1 recompute=none
7
+ memory_dump=none snapshots=/workspace/granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/memory_snapshots/snapshot.pickle
8
+ repo=/workspace/Megatron-MoE-Echo @ 9a236793
9
+ ==============================================================
10
+ W0921 07:17:01.942000 416120 torch/distributed/run.py:766]
11
+ W0921 07:17:01.942000 416120 torch/distributed/run.py:766] *****************************************
12
+ W0921 07:17:01.942000 416120 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.
13
+ W0921 07:17:01.942000 416120 torch/distributed/run.py:766] *****************************************
14
+ 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
15
+ WARNING: overriding default arguments for tokenizer_type:GPT2BPETokenizer with tokenizer_type:GPT2BPETokenizer
16
+ Number of virtual stages per pipeline stage: None
17
+ accumulate and all-reduce gradients in fp32 for bfloat16 data type.
18
+ using torch.bfloat16 for parameters ...
19
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
20
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
21
+ ------------------------ arguments ------------------------
22
+ account_for_embedding_in_pipeline_split ......... False
23
+ account_for_loss_in_pipeline_split .............. False
24
+ accumulate_allreduce_grads_in_fp32 .............. True
25
+ adam_beta1 ...................................... 0.9
26
+ adam_beta2 ...................................... 0.999
27
+ adam_eps ........................................ 1e-08
28
+ add_bias_linear ................................. False
29
+ add_position_embedding .......................... False
30
+ add_qkv_bias .................................... False
31
+ adlr_autoresume ................................. False
32
+ adlr_autoresume_interval ........................ 1000
33
+ align_grad_reduce ............................... True
34
+ align_param_gather .............................. False
35
+ app_tag_run_name ................................ None
36
+ app_tag_run_version ............................. 0.0.0
37
+ apply_layernorm_1p .............................. False
38
+ apply_query_key_layer_scaling ................... False
39
+ apply_residual_connection_post_layernorm ........ False
40
+ apply_rope_fusion ............................... False
41
+ async_save ...................................... None
42
+ async_tensor_model_parallel_allreduce ........... True
43
+ attention_backend ............................... AttnBackend.auto
44
+ attention_dropout ............................... 0.0
45
+ attention_softmax_in_fp32 ....................... False
46
+ auto_detect_ckpt_format ......................... False
47
+ barrier_with_L1_time ............................ True
48
+ bert_binary_head ................................ True
49
+ bert_embedder_type .............................. megatron
50
+ bert_load ....................................... None
51
+ bf16 ............................................ True
52
+ bias_dropout_fusion ............................. True
53
+ bias_gelu_fusion ................................ False
54
+ bias_swiglu_fusion .............................. True
55
+ biencoder_projection_dim ........................ 0
56
+ biencoder_shared_query_context_model ............ False
57
+ block_data_path ................................. None
58
+ calc_ft_timeouts ................................ False
59
+ calculate_per_token_loss ........................ False
60
+ check_for_large_grads ........................... False
61
+ check_for_nan_in_loss_and_grad .................. False
62
+ check_for_spiky_loss ............................ False
63
+ check_weight_hash_across_dp_replicas_interval ... None
64
+ ckpt_assume_constant_structure .................. False
65
+ ckpt_convert_format ............................. None
66
+ ckpt_convert_save ............................... None
67
+ ckpt_convert_update_legacy_dist_opt_format ...... False
68
+ ckpt_format ..................................... torch_dist
69
+ ckpt_fully_parallel_load ........................ False
70
+ ckpt_fully_parallel_save ........................ True
71
+ ckpt_fully_parallel_save_deprecated ............. False
72
+ ckpt_step ....................................... None
73
+ classes_fraction ................................ 1.0
74
+ clip_grad ....................................... 1.0
75
+ clone_scatter_output_in_embedding ............... True
76
+ config_logger_dir ...............................
77
+ consumed_train_samples .......................... 0
78
+ consumed_valid_samples .......................... 0
79
+ context_parallel_size ........................... 1
80
+ cp_comm_type .................................... ['p2p']
81
+ create_attention_mask_in_dataloader ............. True
82
+ cross_entropy_fusion_impl ....................... native
83
+ cross_entropy_loss_fusion ....................... False
84
+ cuda_graph_scope ................................ full_iteration
85
+ cuda_graph_warmup_steps ......................... 3
86
+ data_args_path .................................. None
87
+ data_cache_path ................................. None
88
+ data_parallel_random_init ....................... False
89
+ data_parallel_sharding_strategy ................. no_shard
90
+ data_parallel_size .............................. 8
91
+ data_path ....................................... ['/workspace/data/wikitext/wikitext103_gpt2_text_document']
92
+ data_per_class_fraction ......................... 1.0
93
+ data_sharding ................................... True
94
+ dataloader_type ................................. single
95
+ ddp_average_in_collective ....................... False
96
+ ddp_bucket_size ................................. None
97
+ ddp_num_buckets ................................. None
98
+ ddp_pad_buckets_for_high_nccl_busbw ............. False
99
+ decoder_first_pipeline_num_layers ............... None
100
+ decoder_last_pipeline_num_layers ................ None
101
+ decoder_num_layers .............................. None
102
+ decoder_seq_length .............................. None
103
+ decoupled_lr .................................... None
104
+ decoupled_min_lr ................................ None
105
+ decrease_batch_size_if_needed ................... False
106
+ defer_embedding_wgrad_compute ................... False
107
+ delay_wgrad_compute ............................. False
108
+ deprecated_use_mcore_models ..................... True
109
+ deterministic_mode .............................. False
110
+ dino_bottleneck_size ............................ 256
111
+ dino_freeze_last_layer .......................... 1
112
+ dino_head_hidden_size ........................... 2048
113
+ dino_local_crops_number ......................... 10
114
+ dino_local_img_size ............................. 96
115
+ dino_norm_last_layer ............................ False
116
+ dino_teacher_temp ............................... 0.07
117
+ dino_warmup_teacher_temp ........................ 0.04
118
+ dino_warmup_teacher_temp_epochs ................. 30
119
+ disable_bf16_reduced_precision_matmul ........... False
120
+ disable_mamba_mem_eff_path ...................... False
121
+ disable_straggler_on_startup .................... False
122
+ dist_ckpt_format_deprecated ..................... None
123
+ dist_ckpt_strictness ............................ assume_ok_unexpected
124
+ distribute_saved_activations .................... False
125
+ distributed_backend ............................. nccl
126
+ distributed_timeout_minutes ..................... 10
127
+ embedding_init_method_std ....................... None
128
+ embedding_path .................................. None
129
+ empty_unused_memory_level ....................... 0
130
+ enable_cuda_graph ............................... True
131
+ enable_experimental ............................. False
132
+ enable_ft_package ............................... False
133
+ enable_gloo_process_groups ...................... True
134
+ enable_msc ...................................... True
135
+ enable_one_logger ............................... True
136
+ encoder_num_layers .............................. 4
137
+ encoder_seq_length .............................. 4096
138
+ end_weight_decay ................................ 0.1
139
+ eod_mask_loss ................................... False
140
+ error_injection_rate ............................ 0
141
+ error_injection_type ............................ transient_error
142
+ eval_interval ................................... 1000000
143
+ eval_iters ...................................... 0
144
+ evidence_data_path .............................. None
145
+ exit_duration_in_mins ........................... None
146
+ exit_interval ................................... None
147
+ exit_on_missing_checkpoint ...................... False
148
+ exit_signal_handler ............................. False
149
+ exp_avg_dtype ................................... torch.float32
150
+ exp_avg_sq_dtype ................................ torch.float32
151
+ expert_model_parallel_size ...................... 8
152
+ expert_tensor_parallel_size ..................... 1
153
+ export_force_local_attention .................... False
154
+ export_kd_cfg ................................... None
155
+ export_kd_teacher_ckpt_format ................... None
156
+ export_kd_teacher_load .......................... None
157
+ export_kv_cache_quant ........................... False
158
+ export_legacy_megatron .......................... False
159
+ export_model_type ............................... GPTModel
160
+ export_moe_apply_probs_on_input ................. False
161
+ export_qk_l2_norm ............................... False
162
+ export_quant_cfg ................................ None
163
+ export_real_quant_cfg ........................... None
164
+ export_te_mcore_model ........................... False
165
+ external_cuda_graph ............................. False
166
+ ffn_hidden_size ................................. 18432
167
+ finetune ........................................ False
168
+ finetune_data_split ............................. train
169
+ finetune_hf_dataset ............................. None
170
+ first_last_layers_bf16 .......................... False
171
+ flash_decode .................................... False
172
+ fp16 ............................................ False
173
+ fp16_lm_cross_entropy ........................... False
174
+ fp32_residual_connection ........................ False
175
+ fp8 ............................................. e4m3
176
+ fp8_amax_compute_algo ........................... most_recent
177
+ fp8_amax_history_len ............................ 1
178
+ fp8_interval .................................... 1
179
+ fp8_margin ...................................... 0
180
+ fp8_param_gather ................................ False
181
+ fp8_recipe ...................................... mxfp8
182
+ fp8_wgrad ....................................... True
183
+ fsdp_double_buffer .............................. False
184
+ global_batch_size ............................... 8
185
+ grad_reduce_in_bf16 ............................. False
186
+ gradient_accumulation_fusion .................... True
187
+ gradient_reduce_div_fusion ...................... True
188
+ group_query_attention ........................... False
189
+ head_lr_mult .................................... 1.0
190
+ heterogeneous_layers_config_encoded_json ........ None
191
+ heterogeneous_layers_config_path ................ None
192
+ hidden_dropout .................................. 0.0
193
+ hidden_size ..................................... 7168
194
+ hierarchical_context_parallel_sizes ............. None
195
+ high_priority_stream_groups ..................... []
196
+ hybrid_attention_ratio .......................... 0.0
197
+ hybrid_mlp_ratio ................................ 0.0
198
+ hybrid_override_pattern ......................... None
199
+ hysteresis ...................................... 2
200
+ ict_head_size ................................... None
201
+ ict_load ........................................ None
202
+ img_h ........................................... 224
203
+ img_w ........................................... 224
204
+ indexer_batch_size .............................. 128
205
+ indexer_log_interval ............................ 1000
206
+ inference_batch_times_seqlen_threshold .......... -1
207
+ inference_dynamic_batching ...................... False
208
+ inference_dynamic_batching_buffer_guaranteed_fraction 0.2
209
+ inference_dynamic_batching_buffer_overflow_factor None
210
+ inference_dynamic_batching_buffer_size_gb ....... 40.0
211
+ inference_dynamic_batching_chunk_size ........... 256
212
+ inference_dynamic_batching_max_requests_override None
213
+ inference_dynamic_batching_max_tokens_override .. None
214
+ inference_dynamic_batching_num_cuda_graphs ...... 16
215
+ inference_max_batch_size ........................ 8
216
+ inference_max_seq_length ........................ 2560
217
+ inference_rng_tracker ........................... False
218
+ init_method_std ................................. 0.02
219
+ init_method_xavier_uniform ...................... False
220
+ init_model_with_meta_device ..................... False
221
+ initial_loss_scale .............................. 4294967296
222
+ inprocess_active_world_size ..................... 8
223
+ inprocess_barrier_timeout ....................... 120
224
+ inprocess_completion_timeout .................... 120
225
+ inprocess_empty_cuda_cache ...................... False
226
+ inprocess_granularity ........................... node
227
+ inprocess_hard_timeout .......................... 90
228
+ inprocess_heartbeat_interval .................... 30
229
+ inprocess_heartbeat_timeout ..................... 60
230
+ inprocess_last_call_wait ........................ 1
231
+ inprocess_max_iterations ........................ None
232
+ inprocess_monitor_process_interval .............. 1.0
233
+ inprocess_monitor_thread_interval ............... 1.0
234
+ inprocess_progress_watchdog_interval ............ 1.0
235
+ inprocess_restart ............................... False
236
+ inprocess_soft_timeout .......................... 60
237
+ inprocess_termination_grace_time ................ 1
238
+ is_hybrid_model ................................. False
239
+ iter_per_epoch .................................. 1250
240
+ iterations_to_skip .............................. []
241
+ keep_fp8_transpose_cache_when_using_custom_fsdp . False
242
+ kitchen_config_file ............................. None
243
+ kitchen_recipe_number ........................... None
244
+ kv_channels ..................................... 128
245
+ kv_lora_rank .................................... 512
246
+ lazy_mpu_init ................................... None
247
+ load ............................................ None
248
+ load_main_params_from_ckpt ...................... None
249
+ load_model_opt_format ........................... False
250
+ local_rank ...................................... 0
251
+ log_energy ...................................... False
252
+ log_interval .................................... 1
253
+ log_loss_scale_to_tensorboard ................... True
254
+ log_memory_to_tensorboard ....................... False
255
+ log_num_zeros_in_grad ........................... False
256
+ log_params_norm ................................. False
257
+ log_progress .................................... False
258
+ log_straggler ................................... False
259
+ log_throughput .................................. True
260
+ log_timers_to_tensorboard ....................... False
261
+ log_validation_ppl_to_tensorboard ............... False
262
+ log_world_size_to_tensorboard ................... False
263
+ logging_level ................................... None
264
+ loss_scale ...................................... None
265
+ loss_scale_window ............................... 1000
266
+ lr .............................................. 0.0001
267
+ lr_decay_iters .................................. 8
268
+ lr_decay_samples ................................ None
269
+ lr_decay_style .................................. cosine
270
+ lr_warmup_fraction .............................. None
271
+ lr_warmup_init .................................. 0.0
272
+ lr_warmup_iters ................................. 5
273
+ lr_warmup_samples ............................... 0
274
+ lr_wsd_decay_iters .............................. None
275
+ lr_wsd_decay_samples ............................ None
276
+ lr_wsd_decay_style .............................. exponential
277
+ main_grads_dtype ................................ torch.float32
278
+ main_params_dtype ............................... torch.float32
279
+ make_vocab_size_divisible_by .................... 128
280
+ mamba_head_dim .................................. 64
281
+ mamba_num_groups ................................ 8
282
+ mamba_num_heads ................................. None
283
+ mamba_state_dim ................................. 128
284
+ manual_gc ....................................... False
285
+ manual_gc_eval .................................. True
286
+ manual_gc_interval .............................. 0
287
+ mask_factor ..................................... 1.0
288
+ mask_prob ....................................... 0.15
289
+ mask_type ....................................... random
290
+ masked_softmax_fusion ........................... False
291
+ max_position_embeddings ......................... 4096
292
+ max_tokens_to_oom ............................... 12000
293
+ memory_snapshot_path ............................ snapshot.pickle
294
+ merge_file ...................................... /workspace/data/wikitext/gpt2-merges.txt
295
+ micro_batch_size ................................ 1
296
+ microbatch_group_size_per_vp_stage .............. None
297
+ mid_level_dataset_surplus ....................... 0.005
298
+ min_loss_scale .................................. 1.0
299
+ min_lr .......................................... 1e-05
300
+ mlp_chunks_for_prefill .......................... 1
301
+ mmap_bin_files .................................. True
302
+ mock_data ....................................... False
303
+ moe_apply_probs_on_input ........................ False
304
+ moe_aux_loss_coeff .............................. 0.0001
305
+ moe_deepep_num_sms .............................. 20
306
+ moe_echo_enable_random_offloading ............... False
307
+ moe_echo_expert_dispatch_overlap ................ True
308
+ moe_echo_expert_dispatcher_type ................. hybridep
309
+ moe_echo_recompute_expert_dispatch .............. False
310
+ moe_enable_deepep ............................... False
311
+ moe_enable_echo ................................. True
312
+ moe_enable_hybridep ............................. True
313
+ moe_expert_capacity_factor ...................... None
314
+ moe_extended_tp ................................. False
315
+ moe_ffn_hidden_size ............................. 2048
316
+ moe_grouped_gemm ................................ True
317
+ moe_hybridep_num_sms ............................ 16
318
+ moe_input_jitter_eps ............................ None
319
+ moe_layer_freq .................................. [0, 0, 0, 1]
320
+ moe_layer_recompute ............................. False
321
+ moe_num_echo_experts ............................ 256
322
+ moe_pad_expert_input_to_capacity ................ False
323
+ moe_per_layer_logging ........................... False
324
+ moe_permute_fusion .............................. False
325
+ moe_received_token_capacity ..................... 1.13671875
326
+ moe_router_bias_update_rate ..................... 0.001
327
+ moe_router_dtype ................................ fp32
328
+ moe_router_enable_expert_bias ................... True
329
+ moe_router_force_load_balancing ................. False
330
+ moe_router_group_topk ........................... 4
331
+ moe_router_load_balancing_type .................. seq_aux_loss
332
+ moe_router_num_groups ........................... 8
333
+ moe_router_padding_for_fp8 ...................... False
334
+ moe_router_pre_softmax .......................... False
335
+ moe_router_score_function ....................... sigmoid
336
+ moe_router_topk ................................. 8
337
+ moe_router_topk_scaling_factor .................. 2.5
338
+ moe_shared_expert_intermediate_size ............. 2048
339
+ moe_shared_expert_overlap ....................... False
340
+ moe_token_dispatcher_type ....................... flex
341
+ moe_token_drop_policy ........................... probs
342
+ moe_upcycling_granularity ....................... 1
343
+ moe_use_device_initiated_grouped_gemm ........... True
344
+ moe_use_legacy_grouped_gemm ..................... False
345
+ moe_use_upcycling ............................... False
346
+ moe_z_loss_coeff ................................ None
347
+ mrope_section ................................... None
348
+ mscale .......................................... 1.0
349
+ mscale_all_dim .................................. 1.0
350
+ mtp_loss_scaling_factor ......................... 0.1
351
+ mtp_num_layers .................................. None
352
+ multi_latent_attention .......................... True
353
+ nccl_all_reduce_for_prefill ..................... False
354
+ nccl_communicator_config_path ................... None
355
+ nccl_ub ......................................... False
356
+ no_load_optim ................................... None
357
+ no_load_rng ..................................... None
358
+ no_persist_layer_norm ........................... False
359
+ no_rope_freq .................................... None
360
+ no_save_optim ................................... None
361
+ no_save_rng ..................................... None
362
+ non_persistent_ckpt_type ........................ None
363
+ non_persistent_global_ckpt_dir .................. None
364
+ non_persistent_local_ckpt_algo .................. fully_parallel
365
+ non_persistent_local_ckpt_dir ................... None
366
+ non_persistent_save_interval .................... None
367
+ norm_epsilon .................................... 1e-06
368
+ normalization ................................... RMSNorm
369
+ num_attention_heads ............................. 128
370
+ num_channels .................................... 3
371
+ num_classes ..................................... 1000
372
+ num_dataset_builder_threads ..................... 1
373
+ num_distributed_optimizer_instances ............. 1
374
+ num_experts ..................................... 256
375
+ num_layers ...................................... 4
376
+ num_layers_at_end_in_bf16 ....................... 1
377
+ num_layers_at_start_in_bf16 ..................... 1
378
+ num_layers_per_virtual_pipeline_stage ........... None
379
+ num_query_groups ................................ 1
380
+ num_virtual_stages_per_pipeline_rank ............ None
381
+ num_workers ..................................... 2
382
+ object_storage_cache_path ....................... None
383
+ one_logger_async ................................ False
384
+ one_logger_project .............................. megatron-lm
385
+ one_logger_run_name ............................. None
386
+ onnx_safe ....................................... None
387
+ openai_gelu ..................................... False
388
+ optimizer ....................................... adam
389
+ optimizer_cpu_offload ........................... False
390
+ optimizer_offload_fraction ...................... 1.0
391
+ output_bert_embeddings .......................... False
392
+ overlap_cpu_optimizer_d2h_h2d ................... False
393
+ overlap_grad_reduce ............................. True
394
+ overlap_moe_expert_parallel_comm ................ False
395
+ overlap_p2p_comm ................................ False
396
+ overlap_p2p_comm_warmup_flush ................... False
397
+ overlap_param_gather ............................ True
398
+ overlap_param_gather_with_optimizer_step ........ False
399
+ override_opt_param_scheduler .................... False
400
+ params_dtype .................................... torch.bfloat16
401
+ patch_dim ....................................... 16
402
+ per_split_data_args_path ........................ None
403
+ perform_initialization .......................... True
404
+ pin_cpu_grads ................................... True
405
+ pin_cpu_params .................................. True
406
+ pipeline_model_parallel_comm_backend ............ None
407
+ pipeline_model_parallel_layout .................. None
408
+ pipeline_model_parallel_size .................... 1
409
+ position_embedding_type ......................... rope
410
+ pretrained_checkpoint ........................... None
411
+ profile ......................................... True
412
+ profile_ranks ................................... [0, 1, 2, 3, 4, 5, 6, 7]
413
+ profile_step_end ................................ 8
414
+ profile_step_start .............................. 6
415
+ q_lora_rank ..................................... 1536
416
+ qk_head_dim ..................................... 128
417
+ qk_l2_norm ...................................... False
418
+ qk_layernorm .................................... True
419
+ qk_pos_emb_head_dim ............................. 64
420
+ query_in_block_prob ............................. 0.1
421
+ rampup_batch_size ............................... None
422
+ rank ............................................ 0
423
+ recompute_granularity ........................... None
424
+ recompute_method ................................ None
425
+ recompute_modules ............................... None
426
+ recompute_num_layers ............................ None
427
+ record_memory_history ........................... False
428
+ relative_attention_max_distance ................. 128
429
+ relative_attention_num_buckets .................. 32
430
+ replication ..................................... False
431
+ replication_factor .............................. 2
432
+ replication_jump ................................ None
433
+ rerun_mode ...................................... validate_results
434
+ reset_attention_mask ............................ False
435
+ reset_position_ids .............................. False
436
+ result_rejected_tracker_filename ................ None
437
+ retriever_report_topk_accuracies ................ []
438
+ retriever_score_scaling ......................... False
439
+ retriever_seq_length ............................ 256
440
+ retro_add_retriever ............................. False
441
+ retro_attention_gate ............................ 1
442
+ retro_cyclic_train_iters ........................ None
443
+ retro_encoder_attention_dropout ................. 0.1
444
+ retro_encoder_hidden_dropout .................... 0.1
445
+ retro_encoder_layers ............................ 2
446
+ retro_num_neighbors ............................. 2
447
+ retro_num_retrieved_chunks ...................... 2
448
+ retro_project_dir ............................... None
449
+ retro_verify_neighbor_count ..................... True
450
+ reuse_grad_buf_for_mxfp8_param_ag ............... False
451
+ rope_scaling_factor ............................. 8.0
452
+ rope_type ....................................... None
453
+ rotary_base ..................................... 10000
454
+ rotary_interleaved .............................. False
455
+ rotary_percent .................................. 1.0
456
+ rotary_scaling_factor ........................... 40.0
457
+ rotary_seq_len_interpolation_factor ............. None
458
+ run_workload_inspector_server ................... False
459
+ sample_rate ..................................... 1.0
460
+ save ............................................ None
461
+ save_interval ................................... None
462
+ save_retain_interval ............................ None
463
+ scatter_gather_tensors_in_pipeline .............. True
464
+ seed ............................................ 1234
465
+ seq_length ...................................... 4096
466
+ sequence_parallel ............................... False
467
+ sft ............................................. False
468
+ sft_tokenizer_prompt_format ..................... nemotron-h-aligned
469
+ sgd_momentum .................................... 0.9
470
+ short_seq_prob .................................. 0.1
471
+ skip_train ...................................... False
472
+ skipped_train_samples ........................... 0
473
+ spec ............................................ None
474
+ split ........................................... 100,0,0
475
+ squared_relu .................................... False
476
+ start_weight_decay .............................. 0.1
477
+ straggler_ctrlr_port ............................ 65535
478
+ straggler_minmax_count .......................... 1
479
+ suggested_communication_unit_size ............... None
480
+ swiglu .......................................... True
481
+ swin_backbone_type .............................. tiny
482
+ symmetric_ar_type ............................... None
483
+ te_rng_tracker .................................. True
484
+ tensor_model_parallel_size ...................... 1
485
+ tensorboard_dir ................................. None
486
+ tensorboard_log_interval ........................ 1
487
+ tensorboard_queue_size .......................... 1000
488
+ test_data_path .................................. None
489
+ test_mode ....................................... False
490
+ tiktoken_num_special_tokens ..................... 1000
491
+ tiktoken_pattern ................................ None
492
+ tiktoken_special_tokens ......................... None
493
+ timing_log_level ................................ 0
494
+ timing_log_option ............................... minmax
495
+ titles_data_path ................................ None
496
+ tokenizer_model ................................. None
497
+ tokenizer_type .................................. GPT2BPETokenizer
498
+ torch_fsdp2_reshard_after_forward ............... True
499
+ tp_comm_bootstrap_backend ....................... nccl
500
+ tp_comm_bulk_dgrad .............................. True
501
+ tp_comm_bulk_wgrad .............................. True
502
+ tp_comm_overlap ................................. False
503
+ tp_comm_overlap_ag .............................. True
504
+ tp_comm_overlap_cfg ............................. None
505
+ tp_comm_overlap_rs .............................. True
506
+ tp_comm_overlap_rs_dgrad ........................ False
507
+ tp_comm_split_ag ................................ True
508
+ tp_comm_split_rs ................................ True
509
+ train_data_path ................................. None
510
+ train_iters ..................................... 10
511
+ train_samples ................................... None
512
+ train_sync_interval ............................. None
513
+ transformer_impl ................................ transformer_engine
514
+ transformer_pipeline_model_parallel_size ........ 1
515
+ untie_embeddings_and_output_weights ............. True
516
+ use_checkpoint_args ............................. False
517
+ use_checkpoint_opt_param_scheduler .............. False
518
+ use_cpu_initialization .......................... None
519
+ use_custom_fsdp ................................. False
520
+ use_dist_ckpt ................................... True
521
+ use_dist_ckpt_deprecated ........................ False
522
+ use_distributed_optimizer ....................... True
523
+ use_flash_attn .................................. False
524
+ use_legacy_models ............................... False
525
+ use_mp_args_from_checkpoint_args ................ False
526
+ use_one_sent_docs ............................... False
527
+ use_persistent_ckpt_worker ...................... False
528
+ use_precision_aware_optimizer ................... False
529
+ use_pytorch_profiler ............................ False
530
+ use_ring_exchange_p2p ........................... False
531
+ use_rope_scaling ................................ False
532
+ use_rotary_position_embeddings .................. False
533
+ use_sharp ....................................... False
534
+ use_tokenizer_model_from_checkpoint_args ........ True
535
+ use_torch_fsdp2 ................................. False
536
+ use_torch_optimizer_for_cpu_offload ............. False
537
+ use_tp_pp_dp_mapping ............................ False
538
+ v_head_dim ...................................... 128
539
+ valid_data_path ................................. None
540
+ variable_seq_lengths ............................ False
541
+ virtual_pipeline_model_parallel_size ............ None
542
+ vision_backbone_type ............................ vit
543
+ vision_pretraining .............................. False
544
+ vision_pretraining_type ......................... classify
545
+ vocab_extra_ids ................................. 0
546
+ vocab_file ...................................... /workspace/data/wikitext/gpt2-vocab.json
547
+ vocab_size ...................................... None
548
+ wandb_exp_name ..................................
549
+ wandb_project ...................................
550
+ wandb_save_dir ..................................
551
+ weight_decay .................................... 0.1
552
+ weight_decay_incr_style ......................... constant
553
+ wgrad_deferral_limit ............................ 0
554
+ world_size ...................................... 8
555
+ yaml_cfg ........................................ None
556
+ -------------------- end of arguments ---------------------
557
+ INFO:megatron.core.num_microbatches_calculator:setting number of microbatches to constant 1
558
+ > building GPT2BPETokenizer tokenizer ...
559
+ > padded vocab (size: 50257) with 47 dummy tokens (new size: 50304)
560
+ WARNING:megatron.core.rerun_state_machine:RerunStateMachine initialized in mode RerunMode.VALIDATE_RESULTS
561
+ > initializing torch distributed ...
562
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
563
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
564
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
565
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
566
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
567
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
568
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
569
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
570
+ 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
571
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
572
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
573
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
574
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
575
+ /workspace/Megatron-MoE-Echo/megatron/training/arguments.py:831: UserWarning: Disabling sequence parallelism because tensor model parallelism is disabled
576
+ warnings.warn("Disabling sequence parallelism because tensor model parallelism is disabled")
577
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
578
+ [Gloo] Rank [Gloo] Rank 3 is connected to 7 peer ranks. 4Expected number of connected peer ranks is : is connected to 77 peer ranks. Expected number of connected peer ranks is : 7
579
+
580
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
581
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
582
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
583
+ [Gloo] Rank [Gloo] Rank 7 is connected to 76 peer ranks. is connected to Expected number of connected peer ranks is : 77 peer ranks. Expected number of connected peer ranks is : 7
584
+
585
+ [Gloo] Rank 5 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
586
+ [Gloo] Rank 1 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
587
+ [Gloo] Rank 2 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
588
+ [Gloo] Rank 0 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
589
+ [Gloo] Rank 3 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
590
+ [Gloo] Rank 4 is connected to 7 peer ranks. Expected number of connected peer ranks is : 7
591
+ [Gloo] Rank 7[Gloo] Rank is connected to 7 peer ranks. Expected number of connected peer ranks is : 67 is connected to 7 peer ranks.
592
+ Expected number of connected peer ranks is : 7
593
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
594
+ > initialized tensor model parallel with size 1
595
+ > initialized pipeline model parallel with size 1
596
+ > setting random seeds to 1234 ...
597
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
598
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
599
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
600
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
601
+ [Gloo] Rank 0 is connected to 0 peer ranks. Expected number of connected peer ranks is : 0
602
+ [Gloo] Rank 0 is connected to 0[Gloo] Rank peer ranks. Expected number of connected peer ranks is : 00 is connected to 0
603
+ peer ranks. Expected number of connected peer ranks is : 0
604
+ [rank5]:[W921 07:17:14.545290678 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.
605
+ [rank6]:[W921 07:17:14.545418035 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.
606
+ [rank7]:[W921 07:17:14.545428855 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.
607
+ [rank3]:[W921 07:17:14.545617179 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.
608
+ [rank1]:[W921 07:17:14.545626763 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.
609
+ [rank4]:[W921 07:17:14.545761259 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.
610
+ [rank2]:[W921 07:17:14.545840052 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.
611
+ > compiling dataset index builder ...
612
+ make: Entering directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
613
+ make: Nothing to be done for 'default'.
614
+ make: Leaving directory '/workspace/Megatron-MoE-Echo/megatron/core/datasets'
615
+ >>> done with dataset index builder. Compilation time: 0.425 seconds
616
+ WARNING: constraints for invoking optimized fused softmax kernel are not met. We default back to unfused kernel invocations.
617
+ > compiling and loading fused kernels ...
618
+ [rank0]:[W921 07:17:15.145937093 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.
619
+ >>> done with compiling and loading fused kernels. Compilation time: 3.826 seconds
620
+ time to initialize megatron (seconds): 8.944
621
+ [after megatron is initialized] datetime: 2026-09-21 07:17:21
622
+ building GPT model ...
623
+ /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.
624
+ warnings.warn(
625
+ /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.
626
+ warnings.warn(
627
+ /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.
628
+ warnings.warn(
629
+ /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.
630
+ warnings.warn(
631
+ /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.
632
+ warnings.warn(
633
+ /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.
634
+ warnings.warn(
635
+ /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.
636
+ warnings.warn(
637
+ /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.
638
+ warnings.warn(
639
+ > number of parameters on (tensor, pipeline) model parallel rank (0, 0): 4113898496
640
+ 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)
641
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 17
642
+ Params for bucket 1 (360579072 elements, 360579072 padded size):
643
+ module.output_layer.weight
644
+ Params for bucket 2 (44047360 elements, 44047360 padded size):
645
+ module.decoder.final_layernorm.weight
646
+ module.decoder.layers.3.mlp.shared_experts.linear_fc2.weight
647
+ module.decoder.layers.3.mlp.shared_experts.linear_fc1.weight
648
+ Params for bucket 3 (60497408 elements, 60497408 padded size):
649
+ module.decoder.layers.3.pre_mlp_layernorm.weight
650
+ module.decoder.layers.3.self_attention.linear_kv_up_proj.weight
651
+ module.decoder.layers.3.mlp.router.weight
652
+ module.decoder.layers.3.self_attention.linear_kv_down_proj.weight
653
+ module.decoder.layers.3.self_attention.linear_kv_up_proj.layer_norm_weight
654
+ module.decoder.layers.3.self_attention.linear_q_up_proj.weight
655
+ Params for bucket 4 (128452096 elements, 128452096 padded size):
656
+ module.decoder.layers.3.self_attention.linear_q_up_proj.layer_norm_weight
657
+ module.decoder.layers.3.self_attention.linear_proj.weight
658
+ module.decoder.layers.3.self_attention.linear_q_down_proj.weight
659
+ Params for bucket 5 (132127744 elements, 132127744 padded size):
660
+ module.decoder.layers.3.input_layernorm.weight
661
+ module.decoder.layers.2.mlp.linear_fc2.weight
662
+ Params for bucket 6 (264241152 elements, 264241152 padded size):
663
+ module.decoder.layers.2.mlp.linear_fc1.weight
664
+ Params for bucket 7 (58662400 elements, 58662400 padded size):
665
+ module.decoder.layers.2.self_attention.linear_kv_up_proj.weight
666
+ module.decoder.layers.2.self_attention.linear_kv_down_proj.weight
667
+ module.decoder.layers.2.self_attention.linear_q_up_proj.weight
668
+ module.decoder.layers.2.mlp.linear_fc1.layer_norm_weight
669
+ module.decoder.layers.2.self_attention.linear_kv_up_proj.layer_norm_weight
670
+ Params for bucket 8 (128452096 elements, 128452096 padded size):
671
+ module.decoder.layers.2.self_attention.linear_q_up_proj.layer_norm_weight
672
+ module.decoder.layers.2.self_attention.linear_proj.weight
673
+ module.decoder.layers.2.self_attention.linear_q_down_proj.weight
674
+ Params for bucket 9 (132127744 elements, 132127744 padded size):
675
+ module.decoder.layers.2.input_layernorm.weight
676
+ module.decoder.layers.1.mlp.linear_fc2.weight
677
+ Params for bucket 10 (264241152 elements, 264241152 padded size):
678
+ module.decoder.layers.1.mlp.linear_fc1.weight
679
+ Params for bucket 11 (58662400 elements, 58662400 padded size):
680
+ module.decoder.layers.1.self_attention.linear_q_up_proj.weight
681
+ module.decoder.layers.1.self_attention.linear_kv_down_proj.weight
682
+ module.decoder.layers.1.mlp.linear_fc1.layer_norm_weight
683
+ module.decoder.layers.1.self_attention.linear_kv_up_proj.layer_norm_weight
684
+ module.decoder.layers.1.self_attention.linear_kv_up_proj.weight
685
+ Params for bucket 12 (128452096 elements, 128452096 padded size):
686
+ module.decoder.layers.1.self_attention.linear_q_up_proj.layer_norm_weight
687
+ module.decoder.layers.1.self_attention.linear_q_down_proj.weight
688
+ module.decoder.layers.1.self_attention.linear_proj.weight
689
+ Params for bucket 13 (132127744 elements, 132127744 padded size):
690
+ module.decoder.layers.1.input_layernorm.weight
691
+ module.decoder.layers.0.mlp.linear_fc2.weight
692
+ Params for bucket 14 (264241152 elements, 264241152 padded size):
693
+ module.decoder.layers.0.mlp.linear_fc1.weight
694
+ Params for bucket 15 (58662400 elements, 58662400 padded size):
695
+ module.decoder.layers.0.self_attention.linear_kv_up_proj.weight
696
+ module.decoder.layers.0.self_attention.linear_q_up_proj.weight
697
+ module.decoder.layers.0.self_attention.linear_kv_up_proj.layer_norm_weight
698
+ module.decoder.layers.0.self_attention.linear_kv_down_proj.weight
699
+ module.decoder.layers.0.mlp.linear_fc1.layer_norm_weight
700
+ Params for bucket 16 (128452096 elements, 128452096 padded size):
701
+ module.decoder.layers.0.self_attention.linear_q_up_proj.layer_norm_weight
702
+ module.decoder.layers.0.self_attention.linear_q_down_proj.weight
703
+ module.decoder.layers.0.self_attention.linear_proj.weight
704
+ Params for bucket 17 (360586240 elements, 360586240 padded size):
705
+ module.decoder.layers.0.input_layernorm.weight
706
+ module.embedding.word_embeddings.weight
707
+ INFO:megatron.core.distributed.param_and_grad_buffer:Number of buckets for gradient all-reduce / reduce-scatter: 27
708
+ Params for bucket 1 (44040192 elements, 44040192 padded size):
709
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight31
710
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight29
711
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight30
712
+ Params for bucket 2 (44040192 elements, 44040192 padded size):
713
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight28
714
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight26
715
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight27
716
+ Params for bucket 3 (44040192 elements, 44040192 padded size):
717
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight25
718
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight23
719
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight24
720
+ Params for bucket 4 (44040192 elements, 44040192 padded size):
721
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight22
722
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight20
723
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight21
724
+ Params for bucket 5 (44040192 elements, 44040192 padded size):
725
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight19
726
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight17
727
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight18
728
+ Params for bucket 6 (44040192 elements, 44040192 padded size):
729
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight16
730
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight14
731
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight15
732
+ Params for bucket 7 (44040192 elements, 44040192 padded size):
733
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight13
734
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight12
735
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight11
736
+ Params for bucket 8 (44040192 elements, 44040192 padded size):
737
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight10
738
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight9
739
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight8
740
+ Params for bucket 9 (44040192 elements, 44040192 padded size):
741
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight7
742
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight6
743
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight5
744
+ Params for bucket 10 (44040192 elements, 44040192 padded size):
745
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight4
746
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight3
747
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight2
748
+ Params for bucket 11 (58720256 elements, 58720256 padded size):
749
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight1
750
+ module.decoder.layers.3.mlp.experts.linear_fc2.weight0
751
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight31
752
+ Params for bucket 12 (58720256 elements, 58720256 padded size):
753
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight30
754
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight29
755
+ Params for bucket 13 (58720256 elements, 58720256 padded size):
756
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight28
757
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight27
758
+ Params for bucket 14 (58720256 elements, 58720256 padded size):
759
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight26
760
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight25
761
+ Params for bucket 15 (58720256 elements, 58720256 padded size):
762
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight24
763
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight23
764
+ Params for bucket 16 (58720256 elements, 58720256 padded size):
765
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight22
766
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight21
767
+ Params for bucket 17 (58720256 elements, 58720256 padded size):
768
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight20
769
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight19
770
+ Params for bucket 18 (58720256 elements, 58720256 padded size):
771
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight18
772
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight17
773
+ Params for bucket 19 (58720256 elements, 58720256 padded size):
774
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight16
775
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight15
776
+ Params for bucket 20 (58720256 elements, 58720256 padded size):
777
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight14
778
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight13
779
+ Params for bucket 21 (58720256 elements, 58720256 padded size):
780
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight12
781
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight11
782
+ Params for bucket 22 (58720256 elements, 58720256 padded size):
783
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight10
784
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight9
785
+ Params for bucket 23 (58720256 elements, 58720256 padded size):
786
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight8
787
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight7
788
+ Params for bucket 24 (58720256 elements, 58720256 padded size):
789
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight6
790
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight5
791
+ Params for bucket 25 (58720256 elements, 58720256 padded size):
792
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight4
793
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight3
794
+ Params for bucket 26 (58720256 elements, 58720256 padded size):
795
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight2
796
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight1
797
+ Params for bucket 27 (29360128 elements, 29360128 padded size):
798
+ module.decoder.layers.3.mlp.experts.linear_fc1.weight0
799
+ 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=<megatron.core.timers.Timers object at 0x7947b3bffa10>, config_logger_dir='')
800
+ INFO:megatron.core.optimizer_param_scheduler:> learning rate decay style: cosine
801
+ [after model, optimizer, and learning rate scheduler are built] datetime: 2026-09-21 07:17:21
802
+ > building train, validation, and test datasets ...
803
+ > datasets target sizes (minimum size):
804
+ train: 80
805
+ validation: 0
806
+ test: 0
807
+ INFO:megatron.core.datasets.blended_megatron_dataset_config:Let split_matrix = [(0, 1.0), None, None]
808
+ > building train, validation, and test datasets for GPT ...
809
+ INFO:megatron.core.datasets.blended_megatron_dataset_builder:Building GPTDataset splits with sizes=(80, 0, 0) and config=GPTDatasetConfig(random_seed=1234, sequence_length=4096, 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=<megatron.training.tokenizer.tokenizer._GPT2BPETokenizer object at 0x7947b3ebadb0>, 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)
810
+ INFO:megatron.core.datasets.indexed_dataset:Load the _IndexReader from /workspace/data/wikitext/wikitext103_gpt2_text_document.idx
811
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence lengths
812
+ INFO:megatron.core.datasets.indexed_dataset: Extract the sequence pointers
813
+ INFO:megatron.core.datasets.indexed_dataset: Extract the document indices
814
+ INFO:megatron.core.datasets.indexed_dataset:> total number of sequences: 1165029
815
+ INFO:megatron.core.datasets.indexed_dataset:> total number of documents: 1801350
816
+ INFO:megatron.core.datasets.gpt_dataset:Build and save the GPTDataset train indices
817
+ INFO:megatron.core.datasets.gpt_dataset:> total number of samples: 29073
818
+ INFO:megatron.core.datasets.gpt_dataset:> total number of epochs: 1
819
+ > finished creating GPT datasets ...
820
+ [after dataloaders are built] datetime: 2026-09-21 07:17:21
821
+ done with setup ...
822
+ (min, max) time across ranks (ms):
823
+ model-and-optimizer-setup ......................: (333.87, 335.67)
824
+ train/valid/test-data-iterators-setup ..........: (157.20, 166.65)
825
+ training ...
826
+ Overwriting rerun_state_machine.current_iteration from -1 to 0...
827
+ [before the start of training step] datetime: 2026-09-21 07:17:21
828
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
829
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
830
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
831
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
832
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
833
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
834
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
835
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
836
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
837
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
838
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
839
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
840
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
841
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
842
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
843
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
844
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
845
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
846
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
847
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
848
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
849
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
850
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
851
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
852
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
853
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
854
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
855
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
856
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
857
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
858
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
859
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
860
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
861
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
862
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
863
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
864
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
865
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
866
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
867
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
868
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
869
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
870
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
871
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
872
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
873
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
874
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
875
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
876
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
877
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
878
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
879
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
880
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
881
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
882
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
883
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
884
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
885
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
886
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
887
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
888
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
889
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
890
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
891
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
892
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
893
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
894
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
895
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
896
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
897
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
898
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
899
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
900
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
901
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
902
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
903
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
904
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
905
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
906
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
907
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
908
+ [2026-09-21 07:18:08] iteration 1/ 10 | consumed samples: 8 | elapsed time per iteration (ms): 46160.0 | throughput per GPU (TFLOP/s/GPU): 1.6 | learning rate: 2.000000E-05 | global batch size: 8 | lm loss: 1.226155E+01 | loss scale: 1.0 | grad norm: 15.006 | number of skipped iterations: 0 | number of nan iterations: 0 |
909
+ Number of parameters in transformer block in billions: 13.26
910
+ Number of parameters in embedding layers in billions: 0.72
911
+ Total number of parameters in billions: 13.98
912
+ Number of parameters in most loaded shard in billions: 13.9771
913
+ Theoretical memory footprints: weight and optimizer=99972.23 MB
914
+ [Rank 0] (after 1 iterations) memory (MB) | allocated: 52423.1884765625 | max allocated: 70232.15869140625 | reserved: 84018.0 | max reserved: 84018.0
915
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
916
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
917
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
918
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
919
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
920
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
921
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
922
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
923
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
924
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
925
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
926
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
927
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
928
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
929
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
930
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
931
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
932
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
933
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
934
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
935
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
936
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
937
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
938
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
939
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
940
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
941
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
942
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
943
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
944
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
945
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
946
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
947
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
948
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
949
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
950
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
951
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
952
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
953
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:334: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
954
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
955
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:336: UserWarning: The use_host_meta is deprecated, it will be removed in the future.
956
+ warnings.warn("The use_host_meta is deprecated, it will be removed in the future.")
957
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
958
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
959
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
960
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
961
+ /workspace/DeepEP/deep_ep/hybrid_ep_buffer.py:482: UserWarning: The num_dispatched_tokens is deprecated, it will be removed in the future.
962
+ warnings.warn("The num_dispatched_tokens is deprecated, it will be removed in the future.")
963
+ [2026-09-21 07:18:08] iteration 2/ 10 | consumed samples: 16 | elapsed time per iteration (ms): 209.9 | throughput per GPU (TFLOP/s/GPU): 354.7 | learning rate: 4.000000E-05 | global batch size: 8 | lm loss: 1.227771E+01 | loss scale: 1.0 | grad norm: 14.855 | number of skipped iterations: 0 | number of nan iterations: 0 |
964
+ [2026-09-21 07:18:08] iteration 3/ 10 | consumed samples: 24 | elapsed time per iteration (ms): 207.9 | throughput per GPU (TFLOP/s/GPU): 358.2 | learning rate: 6.000000E-05 | global batch size: 8 | lm loss: 1.070347E+01 | loss scale: 1.0 | grad norm: 18.770 | number of skipped iterations: 0 | number of nan iterations: 0 |
965
+ Capture CUDA graph for training!!!
966
+ CUDA graph capture done!!!
967
+ [2026-09-21 07:18:09] iteration 4/ 10 | consumed samples: 32 | elapsed time per iteration (ms): 1164.8 | throughput per GPU (TFLOP/s/GPU): 63.9 | learning rate: 8.000000E-05 | global batch size: 8 | lm loss: 1.559743E+01 | loss scale: 1.0 | grad norm: 97.632 | number of skipped iterations: 0 | number of nan iterations: 0 |
968
+ [2026-09-21 07:18:09] iteration 5/ 10 | consumed samples: 40 | elapsed time per iteration (ms): 197.6 | throughput per GPU (TFLOP/s/GPU): 376.8 | learning rate: 1.000000E-04 | global batch size: 8 | lm loss: 1.454625E+01 | loss scale: 1.0 | grad norm: 99.162 | number of skipped iterations: 0 | number of nan iterations: 0 |
969
+ [2026-09-21 07:18:09] iteration 6/ 10 | consumed samples: 48 | elapsed time per iteration (ms): 198.4 | throughput per GPU (TFLOP/s/GPU): 375.3 | learning rate: 7.750000E-05 | global batch size: 8 | lm loss: 1.293264E+01 | loss scale: 1.0 | grad norm: 16.053 | number of skipped iterations: 0 | number of nan iterations: 0 |
970
+ [2026-09-21 07:18:13] iteration 7/ 10 | consumed samples: 56 | elapsed time per iteration (ms): 3031.1 | throughput per GPU (TFLOP/s/GPU): 24.6 | learning rate: 3.250000E-05 | global batch size: 8 | lm loss: 1.693121E+01 | loss scale: 1.0 | grad norm: 100.179 | number of skipped iterations: 0 | number of nan iterations: 0 |
971
+ [2026-09-21 07:18:13] iteration 8/ 10 | consumed samples: 64 | elapsed time per iteration (ms): 203.7 | throughput per GPU (TFLOP/s/GPU): 365.6 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 1.094621E+01 | loss scale: 1.0 | grad norm: 18.627 | number of skipped iterations: 0 | number of nan iterations: 0 |
972
+ Capture range started in the application.
973
+ Capture range ended in the application.
974
+ Generating '/tmp/nsys-report-5c35.qdstrm'
975
+
976
+
977
+ [2026-09-21 07:18:52] iteration 9/ 10 | consumed samples: 72 | elapsed time per iteration (ms): 38776.4 | throughput per GPU (TFLOP/s/GPU): 1.9 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 1.050186E+01 | loss scale: 1.0 | grad norm: 16.710 | number of skipped iterations: 0 | number of nan iterations: 0 |
978
+ [2026-09-21 07:18:52] iteration 10/ 10 | consumed samples: 80 | elapsed time per iteration (ms): 203.0 | throughput per GPU (TFLOP/s/GPU): 366.9 | learning rate: 1.000000E-05 | global batch size: 8 | lm loss: 1.016786E+01 | loss scale: 1.0 | grad norm: 13.111 | number of skipped iterations: 0 | number of nan iterations: 0 |
979
+ [after training is done] datetime: 2026-09-21 07:18:52
980
+ [rank0]:[W921 07:18:53.652406132 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())
981
+ [rank2]:[W921 07:18:54.972982294 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
982
+ pure virtual method called
983
+ terminate called without an active exception
984
+ [rank4]:[W921 07:18:54.414692121 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
985
+ pure virtual method called
986
+ terminate called without an active exception
987
+ [rank0]:[W921 07:18:54.431228566 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
988
+ pure virtual method called
989
+ terminate called without an active exception
990
+ [rank7]:[W921 07:18:54.477713302 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
991
+ pure virtual method called
992
+ terminate called without an active exception
993
+ [rank6]:[W921 07:18:54.516678557 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
994
+ pure virtual method called
995
+ terminate called without an active exception
996
+ [rank3]:[W921 07:18:54.524053134 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
997
+ pure virtual method called
998
+ terminate called without an active exception
999
+ [rank5]:[W921 07:18:54.526257941 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1000
+ pure virtual method called
1001
+ terminate called without an active exception
1002
+ [rank1]:[W921 07:18:54.679379149 observer.cpp:116] Warning: Leaked callback handle: 1 (function operator())
1003
+ pure virtual method called
1004
+ terminate called without an active exception
1005
+ W0921 07:18:54.746000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416250 closing signal SIGTERM
1006
+ W0921 07:18:54.747000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416251 closing signal SIGTERM
1007
+ W0921 07:18:54.747000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416253 closing signal SIGTERM
1008
+ W0921 07:18:54.748000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416254 closing signal SIGTERM
1009
+ W0921 07:18:54.748000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416255 closing signal SIGTERM
1010
+ W0921 07:18:54.748000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416256 closing signal SIGTERM
1011
+ W0921 07:18:54.749000 416120 torch/distributed/elastic/multiprocessing/api.py:900] Sending process 416257 closing signal SIGTERM
1012
+ Collection stop is not allowed in this state.
1013
+ E0921 07:18:57.431000 416120 torch/distributed/elastic/multiprocessing/api.py:874] failed (exitcode: -6) local_rank: 2 (pid: 416252) of binary: /usr/bin/python
1014
+ Traceback (most recent call last):
1015
+ File "/usr/local/bin/torchrun", line 8, in <module>
1016
+ sys.exit(main())
1017
+ ^^^^^^
1018
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 355, in wrapper
1019
+ return f(*args, **kwargs)
1020
+ ^^^^^^^^^^^^^^^^^^
1021
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 892, in main
1022
+ run(args)
1023
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 883, in run
1024
+ elastic_launch(
1025
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 139, in __call__
1026
+ return launch_agent(self._config, self._entrypoint, list(args))
1027
+ ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1028
+ File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 270, in launch_agent
1029
+ raise ChildFailedError(
1030
+ torch.distributed.elastic.multiprocessing.errors.ChildFailedError:
1031
+ ============================================================
1032
+ pretrain_gpt.py FAILED
1033
+ ------------------------------------------------------------
1034
+ Failures:
1035
+ <NO_OTHER_FAILURES>
1036
+ ------------------------------------------------------------
1037
+ Root Cause (first observed failure):
1038
+ [0]:
1039
+ time : 2026-09-21_07:18:54
1040
+ host : sp-aiinfra-interactive-aghosh-8-mb99k-8xb200-moe-cg-0
1041
+ rank : 2 (local_rank: 2)
1042
+ exitcode : -6 (pid: 416252)
1043
+ error_file: <N/A>
1044
+ traceback : Signal 6 (SIGABRT) received by PID 416252
1045
+ ============================================================
1046
+ Generated:
1047
+ /workspace/granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875.nsys-rep
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/verify_trace_ranks.log ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ report : /workspace/granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875.nsys-rep
2
+ kernels : 32024
3
+ devices : 8 of 8 -> [0, 1, 2, 3, 4, 5, 6, 7]
4
+ processes: 1
5
+ markers : 800 echo_trace__ kernels on devices [0, 1, 2, 3, 4, 5, 6, 7]
6
+ PASS: all ranks present and backend markers emitted
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/results.tsv ADDED
@@ -0,0 +1 @@
 
 
1
+ config backend policy slots seq gbs cap recompute rep status iter_ms_mean iter_ms_min peak_mib run_dir
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/trace_extension_build.log ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Could not find an up-to-date installation of `packaging`. License expressions might not be validated. To enforce validation, please install `packaging>=24.2`.
2
+ /usr/local/lib/python3.12/dist-packages/setuptools/config/_apply_pyprojecttoml.py:82: SetuptoolsDeprecationWarning: `project.license` as a TOML table is deprecated
3
+ !!
4
+
5
+ ********************************************************************************
6
+ 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).
7
+
8
+ This deprecation is overdue, please update your project and remove deprecated
9
+ calls to avoid build errors in the future.
10
+
11
+ See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details.
12
+ ********************************************************************************
13
+
14
+ !!
15
+ corresp(dist, value, root_dir)
16
+ /usr/local/lib/python3.12/dist-packages/setuptools/config/_apply_pyprojecttoml.py:61: SetuptoolsDeprecationWarning: License classifiers are deprecated.
17
+ !!
18
+
19
+ ********************************************************************************
20
+ Please consider removing the following classifiers in favor of a SPDX license expression:
21
+
22
+ License :: OSI Approved :: BSD License
23
+
24
+ See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details.
25
+ ********************************************************************************
26
+
27
+ !!
28
+ dist._finalize_license_expression()
29
+ /usr/local/lib/python3.12/dist-packages/setuptools/dist.py:759: SetuptoolsDeprecationWarning: License classifiers are deprecated.
30
+ !!
31
+
32
+ ********************************************************************************
33
+ Please consider removing the following classifiers in favor of a SPDX license expression:
34
+
35
+ License :: OSI Approved :: BSD License
36
+
37
+ See https://packaging.python.org/en/latest/guides/writing-pyproject-toml/#license for details.
38
+ ********************************************************************************
39
+
40
+ !!
41
+ self._finalize_license_expression()
42
+ running build_ext
43
+ building 'megatron.core.datasets.helpers_cpp' extension
44
+ Emitting ninja build file /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/build.ninja...
45
+ Compiling objects...
46
+ Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
47
+ ninja: no work to do.
48
+ 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
49
+ building 'megatron.core.transformer.moe._echo_trace_C' extension
50
+ Emitting ninja build file /workspace/Megatron-MoE-Echo/build/temp.linux-x86_64-cpython-312/build.ninja...
51
+ Compiling objects...
52
+ Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
53
+ [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
54
+ [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
55
+ 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
56
+ copying build/lib.linux-x86_64-cpython-312/megatron/core/datasets/helpers_cpp.cpython-312-x86_64-linux-gnu.so -> megatron/core/datasets
57
+ 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
58
+ 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
granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/trace_results.tsv ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ model backend capacity status report async_overlap_json
2
+ DEEPSEEK_V3_L4 echo 1.13671875 OK /workspace/granularity_next_b200/tierb_models/traces/DEEPSEEK_V3_L4/echo/tierbtrace/tierbtrace_gbs8/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875/nsys/DEEPSEEK_V3_L4__echo__matched__seq4096__gbs8__cap1p13671875.nsys-rep