NVIDIA/nvbench

Issues

• Check for CUDA errors after running benchmark

  #88 opened 2 years ago by alliepiper
  4

• Custom CUPTI metrics

  #121 opened a year ago by mmcloughlin
  1

• Tracking `nvbench` benchmark results

  #132 opened a year ago by epompeii
  5

• Add example showing how to parse additional arguments

  #87 opened 2 years ago by alliepiper
  6

• Async benchmarks always deadlock

  #136 opened 3 months ago by gevtushenko
  1

• Getting a lot of errors when building tests and examples on Windows

  #156 opened 4 months ago by gabswb
  0

• compile nvbench_demo error when in c++20 (c++17 works well)

  #155 opened 5 months ago by condy0919
  0

• Investigate refined benchmark sample stopping criterion

  #150 opened 5 months ago by jrhemstad
  0

• Smart selection of the entropy step

  #154 opened 5 months ago by gevtushenko
  0

• Associate parameters with stopping criterion to allow stateful CLI parsing and improved diagnostics

  #153 opened 5 months ago by alliepiper
  0

• vcpkg upstream

  #152 opened 5 months ago by endingly
  0

• Consider `is_finished` customization

  #147 opened 8 months ago by gevtushenko
  0

• Bazel build rules

  #120 opened a year ago by mmcloughlin
  6

• How can the output metrics be interpreted

  #115 opened 2 years ago by seanajohnston
  4

• Cmake Cannot Locate `git_revision.cuh` file in `nvbench/details`

  #144 opened 8 months ago by nv-rliu
  2

• [FEA] Add `nvbench::exec_tag::host` to support CPU-only benchmarks

  #95 opened 2 years ago by sleeepyjack
  1

• [QST] How can I choose `enum_type_list` axis values at runtime?

  #137 opened a year ago by GregoryKimball
  2

• Add option for reporting min/max/median runtime

  #135 opened a year ago by sleeepyjack
  0

• Zero Documentation on CMAKE_CUDA_ARCHITECTURES values

  #128 opened a year ago by Qubitium
  1

• Add workaround for corrupted nsys GPU utilization data

  #104 opened a year ago by GregoryKimball
  3

• Project assumes CXX compiler and and CUDACXXHOST compiler are the same

  #126 opened a year ago by theHamsta
  0

• Automatic cuda graph testing

  #124 opened a year ago by luitjens
  0

• Is there anyway to build offline?

  #112 opened a year ago by ax7e
  8

• [QST] Is there a CLI command to record GPU time for all iterations?

  #105 opened 2 years ago by GregoryKimball
  1

• Document jsonbin output format

  #122 opened a year ago by alliepiper
  0

• Implement and test NVRTC support

  #119 opened a year ago by alliepiper
  1

• Compact output format

  #118 opened a year ago by alliepiper
  0

• Extend `nvbench::..._t` aliases to include CUDA specific types

  #117 opened a year ago by alliepiper
  1

• Invalid JSON and CSV output when running with skipped tests

  #103 opened a year ago by GregoryKimball
  2

• CUDA 11.4 can not compile

  #111 opened 2 years ago by ax7e
  3

• Use publicly visible CI runs

  #110 opened 2 years ago by aaronmondal
  1

• Use squash merging by default

  #109 opened 2 years ago by aaronmondal
  0

• Passing method with multiple template arguments to NVBENCH_BENCH macro

  #101 opened 2 years ago by hibagus
  9

• [FEA] The `--profile` flag should disable CUPTI metrics

  #99 opened 2 years ago by jrhemstad
  11

• [BUG] Memory allocations from flushing L2 can lead to significant delays between benchmark executions

  #100 opened 2 years ago by jrhemstad
  0

• Specify the kernel name or nvtx range for benchmarking.

  #92 opened 2 years ago by neoblizz
  9

• How to properly extend more metric collections of CUPTI

  #91 opened 2 years ago by 4mod3
  3

• [QST] Is there any way to pass filenames to nvbench via command line?

  #86 opened 2 years ago by annielytical
  3

• Reasonable enum defaults don't work

  #84 opened 2 years ago by pauleonix
  2

• Add optional types to type axes

  #82 opened 2 years ago by pauleonix
  4

• Add benchmark groups/sets/tags

  #81 opened 2 years ago by pauleonix
  3

• Add hidden benchmarks

  #83 opened 2 years ago by pauleonix
  0

• Feature Request: Setup and Teardown for `benchmark_base`

  #77 opened 2 years ago by codecircuit
  2

• [FEA] Add Global Set-Up and Tear-Down

  #78 opened 2 years ago by karthikeyann
  0

• Unused parameter warnings

  #75 opened 2 years ago by harrism
  1

• Build fails with fmt 8.1.1

  #73 opened 2 years ago by S-o-T
  0

• `nvbench_compare.py` coloring should be optional and off by default.

  #71 opened 2 years ago by alliepiper
  0

• Export raw samples as binary blobs

  #67 opened 2 years ago by alliepiper
  0

• Add a `zip_axis` / `tuple_axis` that iterates through multiple parameter axes in lockstep

  #68 opened 2 years ago by alliepiper
  6

• What's the right way to include nvbench to an existing cmake project?

  #66 opened 3 years ago by neoblizz
  4