NVIDIA/nvbench
1
0
Fork 0
mirror of https://github.com/NVIDIA/nvbench.git synced 2026-05-13 17:55:39 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add decorators for registering benchmarks and adding axis

Browse Source 
cuda.bench.register(fn) continues returning Benchmark, and supports
legacy use.

New signature added:
   cuda.bench.register():
      Returns a decorator

```
@bench.register()
@bench.axis.float64("Duration (s)", [7e-5, 1e-4, 5e-4])
@bench.option.min_samples(120)
def single_float64_axis(state: bench.State):
   ...
```
This commit is contained in:
Oleksandr Pavlyk
2026-05-04 08:21:41 -05:00
parent 9ea77bccaa
commit e07f87910a
15 changed files with 585 additions and 79 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
python/examples/skip.py
View File

@@ -1,4 +1,4 @@
# Copyright 2025 NVIDIA Corporation
# Copyright 2025-2026 NVIDIA Corporation
#
# Licensed under the Apache License, Version 2.0 with the LLVM exception
# (the "License"); you may not use this file except in compliance with
@@ -57,6 +57,9 @@ __global__ void sleep_kernel(double seconds) {
return mod.get_kernel("sleep_kernel")
@bench.register()
@bench.axis.float64("Duration", [1e-4 + k * 0.25e-3 for k in range(5)])
@bench.axis.string("Kramble", ["Foo", "Bar", "Baz"])
def runtime_skip(state: bench.State):
duration = state.get_float64("Duration")
kramble = state.get_string("Kramble")
@@ -82,8 +85,4 @@ def runtime_skip(state: bench.State):
if __name__ == "__main__":
b = bench.register(runtime_skip)
b.add_float64_axis("Duration", [1e-4 + k * 0.25e-3 for k in range(5)])
b.add_string_axis("Kramble", ["Foo", "Bar", "Baz"])
bench.run_all_benchmarks(sys.argv)