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]