oneapi-agent-toolkit

GPU test: matrix_rotate_hecbench

• Result: ok: PASS on NVIDIA RTX A5000 (cuda:0, nvidia): sycl_verify_vs_reference: PASS: 1/1 case(s) match (worst max abs err 0, atol 0); sycl_verify_vs_native: PASS: 1/1 case(s) match (worst max abs err 0, atol 0)
• Device: NVIDIA RTX A5000 (cuda:0), vendor nvidia
• SYCL target flags: -fsycl-targets=nvidia_gpu_sm_86

step	status	summary
reference_build	ok	exit 0: g++ -O2 -std=c++17 /tmp/tmp.KGKnuT86bF/task/matrix_rotate_ref.cpp -o build/reference
native_build	ok	exit 0: /usr/local/cuda/bin/nvcc -O2 -std=c++17 -arch=native /tmp/tmp.KGKnuT86bF/task/matrix_rotate.cu -o build/native
native_verify_vs_reference	ok	PASS: 1/1 case(s) match (worst max abs err 0, atol 0)
sycl_compile	ok	compiled 1 source(s) -> /tmp/tmp.KGKnuT86bF/out/work/build/sycl (0 warning(s))
sycl_verify_vs_reference	ok	PASS: 1/1 case(s) match (worst max abs err 0, atol 0)
sycl_verify_vs_native	ok	PASS: 1/1 case(s) match (worst max abs err 0, atol 0)
sycl_benchmark	ok	10 run(s) each; cuda:0: median 356.49 ms, p95 436.12 ms
native_benchmark	ok	10 run(s) each; none: median 396.19 ms, p95 444.62 ms

Limitations:

• Benchmarks are single-process wall times plus program-reported kernel times; treat them as rough.

This site is open source. Improve this page.