| schema_version = "1.1" | |
| [task] | |
| name = "mle-bench/mixed-precision-gemm-split" | |
| description = "INT8 activations break on transformers because a handful of input channels carry values 20x larger than everything else. The fix that actually ships — LLM.int8(), Atom, and every outlier-aware W8A8 stack — is to split the contraction axis: the few high-magnitude channels stay bf16, all the rest go to int8 with per-token and per-channel scales, and the two partial products are summed. Run as two library calls it costs two extra full-size passes over the output; fused it is one mainloop with two operand types and one epilogue." | |
| authors = [] | |
| keywords = ["mle", "kernel-generation", "int8", "mixed-precision", "llm-int8", "outlier", "quantization", "gemm", "low-precision", "split-k"] | |
| [metadata] | |
| suite = "mle-bench" | |
| group = "kernel-generation" | |
| level = "1.0" | |
| difficulty = "hard" | |
| category = "mle" | |
| tags = [ "mle", "kernel-generation", "kernels", "gpu", "real-world",] | |
| [verifier] | |
| timeout_sec = 1800.0 | |
| [agent] | |
| timeout_sec = 14400.0 | |
| # GPU request: honored by Modal/GKE/Daytona. On the local docker backend set environment.override_gpus: 0 | |
| # in the job config and attach a GPU via configs/gpu_overlay_nvidia.yaml (see RUNNING.md §5). | |
| # Profiling: `ncu` needs GPU performance counters — the runner must add the SYS_ADMIN capability | |
| # (docker `--cap-add SYS_ADMIN`) or the host must set NVreg_RestrictProfilingToAdminUsers=0. | |
| # Without it ncu exits with ERR_NVGPUCTRPERM. `nsys` works without any extra capability. | |
| [environment] | |
| build_timeout_sec = 3600.0 | |
| cpus = 8 | |
| memory_mb = 65536 | |
| storage_mb = 40960 | |
| gpus = 1 | |
| network_mode = "public" | |
| mcp_servers = [] | |
| [verifier.env] | |
| [environment.env] | |
| [solution.env] | |