| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| KernelBench FP4 Matmul | AutoKernel (Triton) | Throughput (TF/s)2,898 | 14 | 25d ago | |
| Synthetic Transformer Shapes Query-Key Q ⊗ K⊤ | BWTA_QK | Latency (µs)5.41 | 9 | 12d ago | |
| Synthetic Transformer Shapes Attention-Value Att ⊗ V | BWTA_Att | Latency (µs)6.38 | 9 | 12d ago | |
| CUDA-LLM kernels task suite Matrix Multiplication | OptiML | Execution Time (s)4.4 | 9 | 1mo ago | |
| CUDA-LLM kernels task suite 1.0 (test) | OptiML | Execution Time4.4 | 9 | 1mo ago | |
| Matrix Multiplication 3x3, 4x4, 5x5 | CHEHAB RL | Circuit Depth4 | 2 | 1mo ago | |
| CUDA-LLM kernels task suite (test) | - | Execution Time (s)- | 0 | 1mo ago |
