mirror of
https://github.com/NVIDIA/nvbench.git
synced 2026-03-14 20:27:24 +00:00
178dd0eb68
Previously, convergence was tested by waiting for the relative stdev
of cuda timings ("noise") to drop below a certain percentage
(`max_noise`).
This assumed that all benchmarks would eventually see their noise drop
to some threshold, but this is not the case. In practice, many benchmarks
never converge to the default 0.5% relative stdev and instead will always
run to the 15s timeout -- even if the means have converged in a second
or two.
Added a new check that tests when the noise itself stabilizes and ends
the benchmark, even if noise > max_noise.
After testing, this patch alone significantly reduces the runtime of the
Thrust+CUB benchmark suite (from 30 hours to 5 hours) and produces similar
timing results.
The parameters used to tune this feature are not exposed -- if this
approach works long-term and there's a strong motivation to let users
tweak them, then we can worry about names/APIs/CLI/docs later.
41 lines
1.0 KiB
CMake
41 lines
1.0 KiB
CMake
set(test_srcs
|
|
axes_metadata.cu
|
|
benchmark.cu
|
|
create.cu
|
|
cuda_timer.cu
|
|
cpu_timer.cu
|
|
enum_type_list.cu
|
|
float64_axis.cu
|
|
int64_axis.cu
|
|
named_values.cu
|
|
option_parser.cu
|
|
range.cu
|
|
ring_buffer.cu
|
|
runner.cu
|
|
state.cu
|
|
state_generator.cu
|
|
string_axis.cu
|
|
type_axis.cu
|
|
type_list.cu
|
|
)
|
|
|
|
# Metatarget for all examples:
|
|
add_custom_target(nvbench.test.all)
|
|
add_dependencies(nvbench.all nvbench.test.all)
|
|
|
|
foreach(test_src IN LISTS test_srcs)
|
|
get_filename_component(test_name "${test_src}" NAME_WLE)
|
|
string(PREPEND test_name "nvbench.test.")
|
|
add_executable(${test_name} "${test_src}")
|
|
target_include_directories(${test_name} PRIVATE "${CMAKE_CURRENT_LIST_DIR}")
|
|
target_link_libraries(${test_name} PRIVATE nvbench::nvbench fmt)
|
|
set_target_properties(${test_name} PROPERTIES COMPILE_FEATURES cuda_std_17)
|
|
nvbench_config_target(${test_name})
|
|
add_test(NAME ${test_name} COMMAND "$<TARGET_FILE:${test_name}>")
|
|
|
|
add_dependencies(nvbench.test.all ${test_name})
|
|
endforeach()
|
|
|
|
add_subdirectory(cmake)
|
|
add_subdirectory(device)
|