relu / bench

62.9 kB

Ctrl+K

1 contributor

History: 2 commits

superexpai

v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude)

a6e5c43 verified 3 days ago

build_icx.sh

1.39 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
cpu_relu.cpp

3.11 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
cpu_relu2.cpp

2.37 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
cpu_relu3.cpp

2.24 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
int8opt.cu

3.56 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
relu_dtype.cu

5.99 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
relu_icx.cu

3.46 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
relu_icx_modes.cu

4.39 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
spec_dtypes_h100.py

1.75 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
spec_fp32_h100.py

2.34 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
spec_pair_h100.py

1.51 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
spec_small_h100.py

1.88 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
spec_targeted_h100.py

1.89 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
sweep.cu

7.08 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sweep2.cu

4.08 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sweep_h100.cu

5.65 kB
v2: H100 (sm_90) retune — Hopper sub-16MB launch branch, H100 benchmarks/docs, version bump (MK, powered by Claude) 3 days ago
sycl_concurrent.cpp

2.44 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sycl_igpu.cpp

1.56 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sycl_igpu_modes.cpp

2.22 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sycl_igpu_opt.cpp

2.73 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago
sycl_relu.cpp

1.3 kB
Optimized relu: cpu/cuda/xpu, 1.2-1.85x faster on RTX 4090, benchmarked vs upstream and torch.relu about 1 month ago