sglang 0.4.1.post2__py3-none-any.whl → 0.4.1.post4__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- sglang/bench_one_batch.py +2 -0
- sglang/srt/layers/attention/__init__.py +14 -5
- sglang/srt/layers/attention/double_sparsity_backend.py +0 -52
- sglang/srt/layers/attention/flashinfer_backend.py +211 -81
- sglang/srt/layers/attention/torch_native_backend.py +1 -38
- sglang/srt/layers/attention/triton_backend.py +20 -11
- sglang/srt/layers/attention/triton_ops/decode_attention.py +4 -0
- sglang/srt/layers/logits_processor.py +167 -212
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=14336,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=14336,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=1792,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=1792,device_name=NVIDIA_A100-SXM4-80GB.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=3072,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=3072,device_name=NVIDIA_H100_80GB_HBM3,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=3072,device_name=NVIDIA_H100_80GB_HBM3.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=3584,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=3584,device_name=NVIDIA_A100-SXM4-80GB.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=1,N=7168,device_name=NVIDIA_A100-SXM4-80GB.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=1344,device_name=NVIDIA_A100-SXM4-40GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=1344,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=1344,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=14336,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=14336,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=1792,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=1792,device_name=NVIDIA_A100-SXM4-80GB.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=2688,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=2688,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=3072,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=3072,device_name=NVIDIA_H100_80GB_HBM3,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=3200,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +130 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=3584,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=3584,device_name=NVIDIA_A100-SXM4-80GB.json +218 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=6400,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +130 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=7168,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=7168,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=int8_w8a16.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=16,N=800,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +130 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=256,N=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=256,N=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_A800-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=1280,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=2560,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=2560,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=2560,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=320,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=320,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=320,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=320,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_A800-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_GeForce_RTX_4090,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=64,N=640,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=14336,device_name=AMD_Instinct_MI300X.json +200 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=14336,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +138 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=14336,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=14336,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=AMD_Instinct_MI300X.json +200 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=NVIDIA_A100-SXM4-40GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=1792,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=2048,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=2048,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=2048,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=AMD_Instinct_MI300X.json +200 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_A100-SXM4-40GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_GeForce_RTX_4090,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=3584,device_name=NVIDIA_L40S.json +173 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8.json +178 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=4096,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=AMD_Instinct_MI300X.json +200 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=NVIDIA_A100-SXM4-80GB.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=NVIDIA_H100_80GB_HBM3.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=7168,device_name=NVIDIA_H200.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=8192,device_name=AMD_Instinct_MI300X,dtype=fp8_w8a8.json +175 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=8192,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/E=8,N=8192,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/fused_moe.py +187 -29
- sglang/srt/layers/moe/fused_moe_triton/layer.py +14 -6
- sglang/srt/layers/quantization/configs/N=1536,K=1536,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=1536,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=1536,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=2048,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=2048,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=2304,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=2304,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=24576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=24576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=256,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=3072,K=1536,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=3072,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=3072,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=32768,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=32768,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=36864,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=36864,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=4096,K=512,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=4096,K=512,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=4608,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=4608,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=512,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=576,K=7168,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=576,K=7168,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=1024,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=1024,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=1152,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=1152,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=16384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=16384,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=18432,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=18432,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=2048,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=2048,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=2304,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=2304,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/configs/N=7168,K=256,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json +146 -0
- sglang/srt/layers/quantization/fp8.py +2 -2
- sglang/srt/layers/sampler.py +57 -21
- sglang/srt/layers/torchao_utils.py +17 -3
- sglang/srt/managers/detokenizer_manager.py +2 -0
- sglang/srt/managers/io_struct.py +12 -3
- sglang/srt/managers/schedule_batch.py +26 -2
- sglang/srt/managers/schedule_policy.py +159 -90
- sglang/srt/managers/scheduler.py +71 -27
- sglang/srt/managers/tokenizer_manager.py +29 -20
- sglang/srt/managers/tp_worker.py +16 -4
- sglang/srt/managers/tp_worker_overlap_thread.py +3 -4
- sglang/srt/model_executor/cuda_graph_runner.py +118 -73
- sglang/srt/model_executor/forward_batch_info.py +33 -8
- sglang/srt/model_executor/model_runner.py +63 -61
- sglang/srt/models/deepseek_v2.py +34 -7
- sglang/srt/models/grok.py +97 -26
- sglang/srt/openai_api/adapter.py +0 -17
- sglang/srt/openai_api/protocol.py +3 -3
- sglang/srt/sampling/sampling_batch_info.py +21 -0
- sglang/srt/sampling/sampling_params.py +9 -1
- sglang/srt/server.py +9 -5
- sglang/srt/server_args.py +109 -51
- sglang/srt/speculative/build_eagle_tree.py +347 -0
- sglang/srt/speculative/eagle_utils.py +618 -0
- sglang/srt/speculative/eagle_worker.py +170 -0
- sglang/srt/speculative/spec_info.py +5 -0
- sglang/srt/utils.py +15 -2
- sglang/version.py +1 -1
- {sglang-0.4.1.post2.dist-info → sglang-0.4.1.post4.dist-info}/METADATA +9 -8
- sglang-0.4.1.post4.dist-info/RECORD +329 -0
- {sglang-0.4.1.post2.dist-info → sglang-0.4.1.post4.dist-info}/WHEEL +1 -1
- sglang-0.4.1.post2.dist-info/RECORD +0 -197
- {sglang-0.4.1.post2.dist-info → sglang-0.4.1.post4.dist-info}/LICENSE +0 -0
- {sglang-0.4.1.post2.dist-info → sglang-0.4.1.post4.dist-info}/top_level.txt +0 -0
@@ -0,0 +1,146 @@
|
|
1
|
+
{
|
2
|
+
"1": {
|
3
|
+
"BLOCK_SIZE_M": 64,
|
4
|
+
"BLOCK_SIZE_N": 64,
|
5
|
+
"BLOCK_SIZE_K": 128,
|
6
|
+
"GROUP_SIZE_M": 16,
|
7
|
+
"num_warps": 4,
|
8
|
+
"num_stages": 5
|
9
|
+
},
|
10
|
+
"2": {
|
11
|
+
"BLOCK_SIZE_M": 64,
|
12
|
+
"BLOCK_SIZE_N": 64,
|
13
|
+
"BLOCK_SIZE_K": 128,
|
14
|
+
"GROUP_SIZE_M": 1,
|
15
|
+
"num_warps": 4,
|
16
|
+
"num_stages": 5
|
17
|
+
},
|
18
|
+
"4": {
|
19
|
+
"BLOCK_SIZE_M": 64,
|
20
|
+
"BLOCK_SIZE_N": 64,
|
21
|
+
"BLOCK_SIZE_K": 128,
|
22
|
+
"GROUP_SIZE_M": 32,
|
23
|
+
"num_warps": 4,
|
24
|
+
"num_stages": 4
|
25
|
+
},
|
26
|
+
"8": {
|
27
|
+
"BLOCK_SIZE_M": 64,
|
28
|
+
"BLOCK_SIZE_N": 64,
|
29
|
+
"BLOCK_SIZE_K": 128,
|
30
|
+
"GROUP_SIZE_M": 64,
|
31
|
+
"num_warps": 4,
|
32
|
+
"num_stages": 4
|
33
|
+
},
|
34
|
+
"16": {
|
35
|
+
"BLOCK_SIZE_M": 64,
|
36
|
+
"BLOCK_SIZE_N": 64,
|
37
|
+
"BLOCK_SIZE_K": 128,
|
38
|
+
"GROUP_SIZE_M": 32,
|
39
|
+
"num_warps": 4,
|
40
|
+
"num_stages": 4
|
41
|
+
},
|
42
|
+
"24": {
|
43
|
+
"BLOCK_SIZE_M": 64,
|
44
|
+
"BLOCK_SIZE_N": 64,
|
45
|
+
"BLOCK_SIZE_K": 128,
|
46
|
+
"GROUP_SIZE_M": 64,
|
47
|
+
"num_warps": 4,
|
48
|
+
"num_stages": 4
|
49
|
+
},
|
50
|
+
"32": {
|
51
|
+
"BLOCK_SIZE_M": 64,
|
52
|
+
"BLOCK_SIZE_N": 64,
|
53
|
+
"BLOCK_SIZE_K": 128,
|
54
|
+
"GROUP_SIZE_M": 16,
|
55
|
+
"num_warps": 4,
|
56
|
+
"num_stages": 4
|
57
|
+
},
|
58
|
+
"48": {
|
59
|
+
"BLOCK_SIZE_M": 64,
|
60
|
+
"BLOCK_SIZE_N": 64,
|
61
|
+
"BLOCK_SIZE_K": 128,
|
62
|
+
"GROUP_SIZE_M": 64,
|
63
|
+
"num_warps": 4,
|
64
|
+
"num_stages": 4
|
65
|
+
},
|
66
|
+
"64": {
|
67
|
+
"BLOCK_SIZE_M": 64,
|
68
|
+
"BLOCK_SIZE_N": 64,
|
69
|
+
"BLOCK_SIZE_K": 128,
|
70
|
+
"GROUP_SIZE_M": 16,
|
71
|
+
"num_warps": 4,
|
72
|
+
"num_stages": 4
|
73
|
+
},
|
74
|
+
"96": {
|
75
|
+
"BLOCK_SIZE_M": 64,
|
76
|
+
"BLOCK_SIZE_N": 32,
|
77
|
+
"BLOCK_SIZE_K": 128,
|
78
|
+
"GROUP_SIZE_M": 1,
|
79
|
+
"num_warps": 4,
|
80
|
+
"num_stages": 4
|
81
|
+
},
|
82
|
+
"128": {
|
83
|
+
"BLOCK_SIZE_M": 64,
|
84
|
+
"BLOCK_SIZE_N": 32,
|
85
|
+
"BLOCK_SIZE_K": 128,
|
86
|
+
"GROUP_SIZE_M": 1,
|
87
|
+
"num_warps": 4,
|
88
|
+
"num_stages": 3
|
89
|
+
},
|
90
|
+
"256": {
|
91
|
+
"BLOCK_SIZE_M": 64,
|
92
|
+
"BLOCK_SIZE_N": 64,
|
93
|
+
"BLOCK_SIZE_K": 128,
|
94
|
+
"GROUP_SIZE_M": 1,
|
95
|
+
"num_warps": 4,
|
96
|
+
"num_stages": 3
|
97
|
+
},
|
98
|
+
"512": {
|
99
|
+
"BLOCK_SIZE_M": 64,
|
100
|
+
"BLOCK_SIZE_N": 64,
|
101
|
+
"BLOCK_SIZE_K": 128,
|
102
|
+
"GROUP_SIZE_M": 32,
|
103
|
+
"num_warps": 4,
|
104
|
+
"num_stages": 3
|
105
|
+
},
|
106
|
+
"1024": {
|
107
|
+
"BLOCK_SIZE_M": 64,
|
108
|
+
"BLOCK_SIZE_N": 64,
|
109
|
+
"BLOCK_SIZE_K": 128,
|
110
|
+
"GROUP_SIZE_M": 32,
|
111
|
+
"num_warps": 4,
|
112
|
+
"num_stages": 3
|
113
|
+
},
|
114
|
+
"1536": {
|
115
|
+
"BLOCK_SIZE_M": 64,
|
116
|
+
"BLOCK_SIZE_N": 64,
|
117
|
+
"BLOCK_SIZE_K": 128,
|
118
|
+
"GROUP_SIZE_M": 64,
|
119
|
+
"num_warps": 4,
|
120
|
+
"num_stages": 3
|
121
|
+
},
|
122
|
+
"2048": {
|
123
|
+
"BLOCK_SIZE_M": 64,
|
124
|
+
"BLOCK_SIZE_N": 128,
|
125
|
+
"BLOCK_SIZE_K": 128,
|
126
|
+
"GROUP_SIZE_M": 64,
|
127
|
+
"num_warps": 4,
|
128
|
+
"num_stages": 3
|
129
|
+
},
|
130
|
+
"3072": {
|
131
|
+
"BLOCK_SIZE_M": 64,
|
132
|
+
"BLOCK_SIZE_N": 128,
|
133
|
+
"BLOCK_SIZE_K": 128,
|
134
|
+
"GROUP_SIZE_M": 64,
|
135
|
+
"num_warps": 4,
|
136
|
+
"num_stages": 3
|
137
|
+
},
|
138
|
+
"4096": {
|
139
|
+
"BLOCK_SIZE_M": 64,
|
140
|
+
"BLOCK_SIZE_N": 128,
|
141
|
+
"BLOCK_SIZE_K": 128,
|
142
|
+
"GROUP_SIZE_M": 64,
|
143
|
+
"num_warps": 4,
|
144
|
+
"num_stages": 3
|
145
|
+
}
|
146
|
+
}
|
@@ -0,0 +1,146 @@
|
|
1
|
+
{
|
2
|
+
"1": {
|
3
|
+
"BLOCK_SIZE_M": 64,
|
4
|
+
"BLOCK_SIZE_N": 32,
|
5
|
+
"BLOCK_SIZE_K": 128,
|
6
|
+
"GROUP_SIZE_M": 1,
|
7
|
+
"num_warps": 4,
|
8
|
+
"num_stages": 4
|
9
|
+
},
|
10
|
+
"2": {
|
11
|
+
"BLOCK_SIZE_M": 64,
|
12
|
+
"BLOCK_SIZE_N": 32,
|
13
|
+
"BLOCK_SIZE_K": 128,
|
14
|
+
"GROUP_SIZE_M": 1,
|
15
|
+
"num_warps": 4,
|
16
|
+
"num_stages": 4
|
17
|
+
},
|
18
|
+
"4": {
|
19
|
+
"BLOCK_SIZE_M": 64,
|
20
|
+
"BLOCK_SIZE_N": 32,
|
21
|
+
"BLOCK_SIZE_K": 128,
|
22
|
+
"GROUP_SIZE_M": 1,
|
23
|
+
"num_warps": 4,
|
24
|
+
"num_stages": 5
|
25
|
+
},
|
26
|
+
"8": {
|
27
|
+
"BLOCK_SIZE_M": 64,
|
28
|
+
"BLOCK_SIZE_N": 32,
|
29
|
+
"BLOCK_SIZE_K": 128,
|
30
|
+
"GROUP_SIZE_M": 64,
|
31
|
+
"num_warps": 4,
|
32
|
+
"num_stages": 4
|
33
|
+
},
|
34
|
+
"16": {
|
35
|
+
"BLOCK_SIZE_M": 64,
|
36
|
+
"BLOCK_SIZE_N": 32,
|
37
|
+
"BLOCK_SIZE_K": 128,
|
38
|
+
"GROUP_SIZE_M": 1,
|
39
|
+
"num_warps": 4,
|
40
|
+
"num_stages": 5
|
41
|
+
},
|
42
|
+
"24": {
|
43
|
+
"BLOCK_SIZE_M": 64,
|
44
|
+
"BLOCK_SIZE_N": 32,
|
45
|
+
"BLOCK_SIZE_K": 128,
|
46
|
+
"GROUP_SIZE_M": 1,
|
47
|
+
"num_warps": 4,
|
48
|
+
"num_stages": 5
|
49
|
+
},
|
50
|
+
"32": {
|
51
|
+
"BLOCK_SIZE_M": 64,
|
52
|
+
"BLOCK_SIZE_N": 32,
|
53
|
+
"BLOCK_SIZE_K": 128,
|
54
|
+
"GROUP_SIZE_M": 1,
|
55
|
+
"num_warps": 4,
|
56
|
+
"num_stages": 4
|
57
|
+
},
|
58
|
+
"48": {
|
59
|
+
"BLOCK_SIZE_M": 64,
|
60
|
+
"BLOCK_SIZE_N": 32,
|
61
|
+
"BLOCK_SIZE_K": 128,
|
62
|
+
"GROUP_SIZE_M": 16,
|
63
|
+
"num_warps": 4,
|
64
|
+
"num_stages": 4
|
65
|
+
},
|
66
|
+
"64": {
|
67
|
+
"BLOCK_SIZE_M": 64,
|
68
|
+
"BLOCK_SIZE_N": 32,
|
69
|
+
"BLOCK_SIZE_K": 128,
|
70
|
+
"GROUP_SIZE_M": 1,
|
71
|
+
"num_warps": 4,
|
72
|
+
"num_stages": 5
|
73
|
+
},
|
74
|
+
"96": {
|
75
|
+
"BLOCK_SIZE_M": 64,
|
76
|
+
"BLOCK_SIZE_N": 32,
|
77
|
+
"BLOCK_SIZE_K": 128,
|
78
|
+
"GROUP_SIZE_M": 1,
|
79
|
+
"num_warps": 4,
|
80
|
+
"num_stages": 4
|
81
|
+
},
|
82
|
+
"128": {
|
83
|
+
"BLOCK_SIZE_M": 64,
|
84
|
+
"BLOCK_SIZE_N": 32,
|
85
|
+
"BLOCK_SIZE_K": 128,
|
86
|
+
"GROUP_SIZE_M": 32,
|
87
|
+
"num_warps": 4,
|
88
|
+
"num_stages": 4
|
89
|
+
},
|
90
|
+
"256": {
|
91
|
+
"BLOCK_SIZE_M": 64,
|
92
|
+
"BLOCK_SIZE_N": 32,
|
93
|
+
"BLOCK_SIZE_K": 128,
|
94
|
+
"GROUP_SIZE_M": 16,
|
95
|
+
"num_warps": 4,
|
96
|
+
"num_stages": 4
|
97
|
+
},
|
98
|
+
"512": {
|
99
|
+
"BLOCK_SIZE_M": 64,
|
100
|
+
"BLOCK_SIZE_N": 32,
|
101
|
+
"BLOCK_SIZE_K": 128,
|
102
|
+
"GROUP_SIZE_M": 32,
|
103
|
+
"num_warps": 4,
|
104
|
+
"num_stages": 4
|
105
|
+
},
|
106
|
+
"1024": {
|
107
|
+
"BLOCK_SIZE_M": 64,
|
108
|
+
"BLOCK_SIZE_N": 64,
|
109
|
+
"BLOCK_SIZE_K": 128,
|
110
|
+
"GROUP_SIZE_M": 16,
|
111
|
+
"num_warps": 4,
|
112
|
+
"num_stages": 4
|
113
|
+
},
|
114
|
+
"1536": {
|
115
|
+
"BLOCK_SIZE_M": 64,
|
116
|
+
"BLOCK_SIZE_N": 32,
|
117
|
+
"BLOCK_SIZE_K": 128,
|
118
|
+
"GROUP_SIZE_M": 64,
|
119
|
+
"num_warps": 4,
|
120
|
+
"num_stages": 3
|
121
|
+
},
|
122
|
+
"2048": {
|
123
|
+
"BLOCK_SIZE_M": 64,
|
124
|
+
"BLOCK_SIZE_N": 64,
|
125
|
+
"BLOCK_SIZE_K": 128,
|
126
|
+
"GROUP_SIZE_M": 32,
|
127
|
+
"num_warps": 4,
|
128
|
+
"num_stages": 5
|
129
|
+
},
|
130
|
+
"3072": {
|
131
|
+
"BLOCK_SIZE_M": 64,
|
132
|
+
"BLOCK_SIZE_N": 64,
|
133
|
+
"BLOCK_SIZE_K": 128,
|
134
|
+
"GROUP_SIZE_M": 1,
|
135
|
+
"num_warps": 4,
|
136
|
+
"num_stages": 3
|
137
|
+
},
|
138
|
+
"4096": {
|
139
|
+
"BLOCK_SIZE_M": 64,
|
140
|
+
"BLOCK_SIZE_N": 128,
|
141
|
+
"BLOCK_SIZE_K": 128,
|
142
|
+
"GROUP_SIZE_M": 64,
|
143
|
+
"num_warps": 4,
|
144
|
+
"num_stages": 3
|
145
|
+
}
|
146
|
+
}
|
@@ -0,0 +1,146 @@
|
|
1
|
+
{
|
2
|
+
"1": {
|
3
|
+
"BLOCK_SIZE_M": 64,
|
4
|
+
"BLOCK_SIZE_N": 32,
|
5
|
+
"BLOCK_SIZE_K": 128,
|
6
|
+
"GROUP_SIZE_M": 32,
|
7
|
+
"num_warps": 4,
|
8
|
+
"num_stages": 5
|
9
|
+
},
|
10
|
+
"2": {
|
11
|
+
"BLOCK_SIZE_M": 64,
|
12
|
+
"BLOCK_SIZE_N": 32,
|
13
|
+
"BLOCK_SIZE_K": 128,
|
14
|
+
"GROUP_SIZE_M": 32,
|
15
|
+
"num_warps": 4,
|
16
|
+
"num_stages": 4
|
17
|
+
},
|
18
|
+
"4": {
|
19
|
+
"BLOCK_SIZE_M": 64,
|
20
|
+
"BLOCK_SIZE_N": 32,
|
21
|
+
"BLOCK_SIZE_K": 128,
|
22
|
+
"GROUP_SIZE_M": 1,
|
23
|
+
"num_warps": 4,
|
24
|
+
"num_stages": 4
|
25
|
+
},
|
26
|
+
"8": {
|
27
|
+
"BLOCK_SIZE_M": 64,
|
28
|
+
"BLOCK_SIZE_N": 32,
|
29
|
+
"BLOCK_SIZE_K": 128,
|
30
|
+
"GROUP_SIZE_M": 32,
|
31
|
+
"num_warps": 4,
|
32
|
+
"num_stages": 4
|
33
|
+
},
|
34
|
+
"16": {
|
35
|
+
"BLOCK_SIZE_M": 64,
|
36
|
+
"BLOCK_SIZE_N": 32,
|
37
|
+
"BLOCK_SIZE_K": 128,
|
38
|
+
"GROUP_SIZE_M": 64,
|
39
|
+
"num_warps": 4,
|
40
|
+
"num_stages": 4
|
41
|
+
},
|
42
|
+
"24": {
|
43
|
+
"BLOCK_SIZE_M": 64,
|
44
|
+
"BLOCK_SIZE_N": 32,
|
45
|
+
"BLOCK_SIZE_K": 128,
|
46
|
+
"GROUP_SIZE_M": 1,
|
47
|
+
"num_warps": 4,
|
48
|
+
"num_stages": 4
|
49
|
+
},
|
50
|
+
"32": {
|
51
|
+
"BLOCK_SIZE_M": 64,
|
52
|
+
"BLOCK_SIZE_N": 32,
|
53
|
+
"BLOCK_SIZE_K": 128,
|
54
|
+
"GROUP_SIZE_M": 64,
|
55
|
+
"num_warps": 4,
|
56
|
+
"num_stages": 4
|
57
|
+
},
|
58
|
+
"48": {
|
59
|
+
"BLOCK_SIZE_M": 64,
|
60
|
+
"BLOCK_SIZE_N": 32,
|
61
|
+
"BLOCK_SIZE_K": 128,
|
62
|
+
"GROUP_SIZE_M": 64,
|
63
|
+
"num_warps": 4,
|
64
|
+
"num_stages": 4
|
65
|
+
},
|
66
|
+
"64": {
|
67
|
+
"BLOCK_SIZE_M": 64,
|
68
|
+
"BLOCK_SIZE_N": 32,
|
69
|
+
"BLOCK_SIZE_K": 128,
|
70
|
+
"GROUP_SIZE_M": 64,
|
71
|
+
"num_warps": 4,
|
72
|
+
"num_stages": 4
|
73
|
+
},
|
74
|
+
"96": {
|
75
|
+
"BLOCK_SIZE_M": 64,
|
76
|
+
"BLOCK_SIZE_N": 32,
|
77
|
+
"BLOCK_SIZE_K": 128,
|
78
|
+
"GROUP_SIZE_M": 1,
|
79
|
+
"num_warps": 4,
|
80
|
+
"num_stages": 5
|
81
|
+
},
|
82
|
+
"128": {
|
83
|
+
"BLOCK_SIZE_M": 64,
|
84
|
+
"BLOCK_SIZE_N": 32,
|
85
|
+
"BLOCK_SIZE_K": 128,
|
86
|
+
"GROUP_SIZE_M": 16,
|
87
|
+
"num_warps": 4,
|
88
|
+
"num_stages": 4
|
89
|
+
},
|
90
|
+
"256": {
|
91
|
+
"BLOCK_SIZE_M": 64,
|
92
|
+
"BLOCK_SIZE_N": 32,
|
93
|
+
"BLOCK_SIZE_K": 128,
|
94
|
+
"GROUP_SIZE_M": 16,
|
95
|
+
"num_warps": 4,
|
96
|
+
"num_stages": 4
|
97
|
+
},
|
98
|
+
"512": {
|
99
|
+
"BLOCK_SIZE_M": 64,
|
100
|
+
"BLOCK_SIZE_N": 64,
|
101
|
+
"BLOCK_SIZE_K": 128,
|
102
|
+
"GROUP_SIZE_M": 16,
|
103
|
+
"num_warps": 4,
|
104
|
+
"num_stages": 4
|
105
|
+
},
|
106
|
+
"1024": {
|
107
|
+
"BLOCK_SIZE_M": 64,
|
108
|
+
"BLOCK_SIZE_N": 32,
|
109
|
+
"BLOCK_SIZE_K": 128,
|
110
|
+
"GROUP_SIZE_M": 32,
|
111
|
+
"num_warps": 4,
|
112
|
+
"num_stages": 4
|
113
|
+
},
|
114
|
+
"1536": {
|
115
|
+
"BLOCK_SIZE_M": 64,
|
116
|
+
"BLOCK_SIZE_N": 64,
|
117
|
+
"BLOCK_SIZE_K": 128,
|
118
|
+
"GROUP_SIZE_M": 1,
|
119
|
+
"num_warps": 4,
|
120
|
+
"num_stages": 4
|
121
|
+
},
|
122
|
+
"2048": {
|
123
|
+
"BLOCK_SIZE_M": 64,
|
124
|
+
"BLOCK_SIZE_N": 64,
|
125
|
+
"BLOCK_SIZE_K": 128,
|
126
|
+
"GROUP_SIZE_M": 64,
|
127
|
+
"num_warps": 4,
|
128
|
+
"num_stages": 3
|
129
|
+
},
|
130
|
+
"3072": {
|
131
|
+
"BLOCK_SIZE_M": 64,
|
132
|
+
"BLOCK_SIZE_N": 128,
|
133
|
+
"BLOCK_SIZE_K": 128,
|
134
|
+
"GROUP_SIZE_M": 32,
|
135
|
+
"num_warps": 4,
|
136
|
+
"num_stages": 3
|
137
|
+
},
|
138
|
+
"4096": {
|
139
|
+
"BLOCK_SIZE_M": 64,
|
140
|
+
"BLOCK_SIZE_N": 64,
|
141
|
+
"BLOCK_SIZE_K": 128,
|
142
|
+
"GROUP_SIZE_M": 16,
|
143
|
+
"num_warps": 4,
|
144
|
+
"num_stages": 3
|
145
|
+
}
|
146
|
+
}
|
@@ -0,0 +1,146 @@
|
|
1
|
+
{
|
2
|
+
"1": {
|
3
|
+
"BLOCK_SIZE_M": 64,
|
4
|
+
"BLOCK_SIZE_N": 32,
|
5
|
+
"BLOCK_SIZE_K": 128,
|
6
|
+
"GROUP_SIZE_M": 1,
|
7
|
+
"num_warps": 4,
|
8
|
+
"num_stages": 4
|
9
|
+
},
|
10
|
+
"2": {
|
11
|
+
"BLOCK_SIZE_M": 64,
|
12
|
+
"BLOCK_SIZE_N": 32,
|
13
|
+
"BLOCK_SIZE_K": 128,
|
14
|
+
"GROUP_SIZE_M": 16,
|
15
|
+
"num_warps": 4,
|
16
|
+
"num_stages": 4
|
17
|
+
},
|
18
|
+
"4": {
|
19
|
+
"BLOCK_SIZE_M": 64,
|
20
|
+
"BLOCK_SIZE_N": 32,
|
21
|
+
"BLOCK_SIZE_K": 128,
|
22
|
+
"GROUP_SIZE_M": 16,
|
23
|
+
"num_warps": 4,
|
24
|
+
"num_stages": 4
|
25
|
+
},
|
26
|
+
"8": {
|
27
|
+
"BLOCK_SIZE_M": 64,
|
28
|
+
"BLOCK_SIZE_N": 32,
|
29
|
+
"BLOCK_SIZE_K": 128,
|
30
|
+
"GROUP_SIZE_M": 16,
|
31
|
+
"num_warps": 4,
|
32
|
+
"num_stages": 4
|
33
|
+
},
|
34
|
+
"16": {
|
35
|
+
"BLOCK_SIZE_M": 64,
|
36
|
+
"BLOCK_SIZE_N": 32,
|
37
|
+
"BLOCK_SIZE_K": 128,
|
38
|
+
"GROUP_SIZE_M": 1,
|
39
|
+
"num_warps": 4,
|
40
|
+
"num_stages": 4
|
41
|
+
},
|
42
|
+
"24": {
|
43
|
+
"BLOCK_SIZE_M": 64,
|
44
|
+
"BLOCK_SIZE_N": 32,
|
45
|
+
"BLOCK_SIZE_K": 128,
|
46
|
+
"GROUP_SIZE_M": 32,
|
47
|
+
"num_warps": 4,
|
48
|
+
"num_stages": 4
|
49
|
+
},
|
50
|
+
"32": {
|
51
|
+
"BLOCK_SIZE_M": 64,
|
52
|
+
"BLOCK_SIZE_N": 32,
|
53
|
+
"BLOCK_SIZE_K": 128,
|
54
|
+
"GROUP_SIZE_M": 1,
|
55
|
+
"num_warps": 4,
|
56
|
+
"num_stages": 4
|
57
|
+
},
|
58
|
+
"48": {
|
59
|
+
"BLOCK_SIZE_M": 64,
|
60
|
+
"BLOCK_SIZE_N": 32,
|
61
|
+
"BLOCK_SIZE_K": 128,
|
62
|
+
"GROUP_SIZE_M": 1,
|
63
|
+
"num_warps": 4,
|
64
|
+
"num_stages": 4
|
65
|
+
},
|
66
|
+
"64": {
|
67
|
+
"BLOCK_SIZE_M": 64,
|
68
|
+
"BLOCK_SIZE_N": 32,
|
69
|
+
"BLOCK_SIZE_K": 128,
|
70
|
+
"GROUP_SIZE_M": 32,
|
71
|
+
"num_warps": 4,
|
72
|
+
"num_stages": 4
|
73
|
+
},
|
74
|
+
"96": {
|
75
|
+
"BLOCK_SIZE_M": 64,
|
76
|
+
"BLOCK_SIZE_N": 32,
|
77
|
+
"BLOCK_SIZE_K": 128,
|
78
|
+
"GROUP_SIZE_M": 1,
|
79
|
+
"num_warps": 4,
|
80
|
+
"num_stages": 5
|
81
|
+
},
|
82
|
+
"128": {
|
83
|
+
"BLOCK_SIZE_M": 64,
|
84
|
+
"BLOCK_SIZE_N": 32,
|
85
|
+
"BLOCK_SIZE_K": 128,
|
86
|
+
"GROUP_SIZE_M": 64,
|
87
|
+
"num_warps": 4,
|
88
|
+
"num_stages": 4
|
89
|
+
},
|
90
|
+
"256": {
|
91
|
+
"BLOCK_SIZE_M": 64,
|
92
|
+
"BLOCK_SIZE_N": 32,
|
93
|
+
"BLOCK_SIZE_K": 128,
|
94
|
+
"GROUP_SIZE_M": 32,
|
95
|
+
"num_warps": 4,
|
96
|
+
"num_stages": 4
|
97
|
+
},
|
98
|
+
"512": {
|
99
|
+
"BLOCK_SIZE_M": 64,
|
100
|
+
"BLOCK_SIZE_N": 64,
|
101
|
+
"BLOCK_SIZE_K": 128,
|
102
|
+
"GROUP_SIZE_M": 16,
|
103
|
+
"num_warps": 4,
|
104
|
+
"num_stages": 4
|
105
|
+
},
|
106
|
+
"1024": {
|
107
|
+
"BLOCK_SIZE_M": 64,
|
108
|
+
"BLOCK_SIZE_N": 32,
|
109
|
+
"BLOCK_SIZE_K": 128,
|
110
|
+
"GROUP_SIZE_M": 64,
|
111
|
+
"num_warps": 4,
|
112
|
+
"num_stages": 4
|
113
|
+
},
|
114
|
+
"1536": {
|
115
|
+
"BLOCK_SIZE_M": 64,
|
116
|
+
"BLOCK_SIZE_N": 64,
|
117
|
+
"BLOCK_SIZE_K": 128,
|
118
|
+
"GROUP_SIZE_M": 32,
|
119
|
+
"num_warps": 4,
|
120
|
+
"num_stages": 5
|
121
|
+
},
|
122
|
+
"2048": {
|
123
|
+
"BLOCK_SIZE_M": 64,
|
124
|
+
"BLOCK_SIZE_N": 64,
|
125
|
+
"BLOCK_SIZE_K": 128,
|
126
|
+
"GROUP_SIZE_M": 64,
|
127
|
+
"num_warps": 4,
|
128
|
+
"num_stages": 3
|
129
|
+
},
|
130
|
+
"3072": {
|
131
|
+
"BLOCK_SIZE_M": 64,
|
132
|
+
"BLOCK_SIZE_N": 128,
|
133
|
+
"BLOCK_SIZE_K": 128,
|
134
|
+
"GROUP_SIZE_M": 64,
|
135
|
+
"num_warps": 4,
|
136
|
+
"num_stages": 3
|
137
|
+
},
|
138
|
+
"4096": {
|
139
|
+
"BLOCK_SIZE_M": 64,
|
140
|
+
"BLOCK_SIZE_N": 64,
|
141
|
+
"BLOCK_SIZE_K": 128,
|
142
|
+
"GROUP_SIZE_M": 16,
|
143
|
+
"num_warps": 4,
|
144
|
+
"num_stages": 3
|
145
|
+
}
|
146
|
+
}
|