Skip to content
PR Tests

[ROCm] Optimize kgemm_4bit_inference_naive for ROCm, use it for batch sizes other than 1 #228

Sign in to view logs

[ROCm] Optimize kgemm_4bit_inference_naive for ROCm, use it for batch sizes other than 1

[ROCm] Optimize kgemm_4bit_inference_naive for ROCm, use it for batch sizes other than 1 #228

Run time

Learn about OS pricing on GitHub Actions
Job Run time
CPU (macos, 2.3.1) / build
20s
CPU (linux-aarch64, 2.11.0) / build
13s
CPU (linux-x64, 2.3.1) / build
18s
CPU (macos, 2.11.0) / build
18s
CPU (linux-x64, 2.11.0) / build
15s
CPU (linux-x64, icelake, 2.11.0) / build
22s
CPU (linux-aarch64, 2.5.1) / build
25s
CPU (linux-x64, icelake, 2.3.1) / build
18s
CUDA (linux-x64, L40S, 11.8.0) / build
3m 4s
CUDA (linux-x64, A10, 13.0.2) / build
3m 10s
CUDA (linux-x64, L40S, 13.0.2) / build
3m 51s
CUDA (linux-x64, A10, 12.8.1) / build
2m 59s
CUDA (linux-x64, L40S, 12.8.1) / build
3m 47s
CUDA (linux-x64, T4, 11.8.0) / build
4m 5s
CUDA (linux-x64, A10, 11.8.0) / build
2m 39s
CUDA (windows, T4, 11.8.0, 2.7.1, https://download.pytorch.org/whl/cu118) / build
3m 45s
CUDA (linux-x64, T4, 12.8.1) / build
6m 2s
CUDA (linux-x64, T4, 13.0.2) / build
2m 37s
CPU (linux-aarch64, 2.11.0) / test
15m 33s
CPU (linux-x64, 2.11.0) / test
26m 50s
CPU (linux-x64, icelake, 2.3.1) / test
18m 22s
CPU (macos, 2.11.0) / test
22m 18s
CPU (linux-x64, 2.3.1) / test
22m 48s
CPU (macos, 2.3.1) / test
23m 46s
CPU (linux-aarch64, 2.5.1) / test
15m 14s
CPU (linux-x64, icelake, 2.11.0) / test
19m 6s
CUDA (linux-x64, T4, 13.0.2) / test
9m 25s
CUDA (linux-x64, A10, 11.8.0) / test
4m 34s
CUDA (linux-x64, A10, 12.8.1) / test
7m 56s
CUDA (linux-x64, L40S, 11.8.0) / test
3m 49s
CUDA (linux-x64, A10, 13.0.2) / test
7m 30s
CUDA (windows, T4, 11.8.0, 2.7.1, https://download.pytorch.org/whl/cu118) / test
29m 41s
CUDA (linux-x64, L40S, 12.8.1) / test
6m 42s
CUDA (linux-x64, L40S, 13.0.2) / test
6m 25s
CUDA (linux-x64, T4, 11.8.0) / test
5m 47s
CUDA (linux-x64, T4, 12.8.1) / test
9m 39s
4h 53m 53s
⚡