NVIDIA/nvbench
1
0
Fork 0
mirror of https://github.com/NVIDIA/nvbench.git synced 2026-03-14 20:27:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

Do not use blocking kernel in warmup run of measure_cold (#241)

Browse Source 
See https://github.com/NVIDIA/nvbench/issues/240
This commit is contained in:
Oleksandr Pavlyk
2025-07-03 23:22:12 -05:00
committed by GitHub
parent 53bf11a27d
commit b8c664d22e
1 changed files with 9 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
nvbench/detail/measure_cold.cuh
View File

@@ -135,6 +135,11 @@ struct measure_cold_base::kernel_launch_timer
, m_disable_blocking_kernel{measure.m_disable_blocking_kernel}
{}
explicit kernel_launch_timer(measure_cold_base &measure, bool disable_blocking_kernel)
: m_measure{measure}
, m_disable_blocking_kernel{disable_blocking_kernel}
{}
__forceinline__ void start()
{
m_measure.flush_device_l2();
@@ -206,7 +211,10 @@ private:
return;
}
kernel_launch_timer timer(*this);
// disable use of blocking kernel for warm-up run
// see https://github.com/NVIDIA/nvbench/issues/240
constexpr bool disable_blocking_kernel = true;
kernel_launch_timer timer(*this, disable_blocking_kernel);
this->launch_kernel(timer);
this->check_skip_time(m_cuda_timer.get_duration());