Skip to content

A tool for bandwidth measurements on NVIDIA GPUs.

active 2023-09-112026-08-07 (UTC)

Partial coverage19,963 / 25,676 hourly files (78%) · 2 absent upstream · 5,710 failed, retryable2023-09-082026-08-12 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
487
Pushes
6
Pull requests
16
Issues
36
Stars
311
Forks
34

Activity over time

Daily event counts in the loaded window

Line chart, 1062 days from 2023-09-11 to 2026-08-07. Pushes: 6 total, peak 2 in a day. Pull requests: 16 total, peak 2 in a day. Issues: 36 total, peak 4 in a day. Comments: 60 total, peak 7 in a day. Stars: 311 total, peak 11 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
esitaridi272413
deepakcu170011
ramasubramanianrahul11451
pzcyyj6006
AtoshDustosh3001
chtcvl3002
yaxinchen6662002
CHENfd5152002
Artem-B2011
nmazzilli32001
vitduck2001
seng1e2002
kaizilla72001
henryli0012011
Fanoid2001
SongyuanZhao1000
EvenChang1001
gongysh20041001
MarkoRamius1001
shawn99771000

Recent activity

Latest issues, pull requests and releases

  • Issue#70matthewpopescu2026-08-07 21:31
    one_to_all and all_to_one memcpy measurement operations are launched before "warmup" iterations are finished
  • Issue comment#61esitaridi2026-04-23 22:02
    cuda 13.2 compat issues
  • Pull request#65Artem-B2026-04-13 23:41
  • Issue comment#64Artem-B2026-04-09 18:06
    simpleCopyKernel fails to advance the pointers.
  • Releaseesitaridi2026-04-08 18:11
    v0.9
  • Pull request#63Doondi-Ashlesh2026-04-07 17:43
  • Issue#61ismaws2026-03-26 02:34
    cuda 13.2 compat issues
  • Issue#60kscweave2026-03-25 03:25
    Typo in the README
  • Issue#57emmett20202026-02-17 02:56
    Do we really need introduce Boost dependency?
  • Issue comment#56esitaridi2026-01-27 05:37
    nvbandwidth tests failing with cuda 13.1 on aws p4 instance type
  • Issue comment#56nmazzilli32026-01-27 03:51
    nvbandwidth tests failing with cuda 13.1 on aws p4 instance type
  • Issue#56nmazzilli32026-01-27 00:02
    nvbandwidth tests failing with cuda 13.1 on aws p4 instance type
  • Issue#55shawn99772026-01-23 02:13
    NVML_ERROR: [Not Found] in expression nvmlDeviceGetHandleByUUID(s.str().c_str(), &device) in void setOptimalCpuAffinity(int)() : /home/jetson/project/nvbandwidth/inline_common.h:142
  • Issue comment#51deepakcu2026-01-21 05:59
    Skip boost static libs on Azure Linux
  • Pull request#54shadychan2026-01-15 09:23
  • Issue comment#51deepakcu2026-01-12 07:22
    Skip boost static libs on Azure Linux
  • Issue comment#51henryli0012026-01-09 22:29
    Skip boost static libs on Azure Linux
  • Issue comment#53deepakcu2025-11-25 05:17
    Asymmetric bandwidth in bidirectional Host-Device transfers: H2D drops to ~11 GB/s while D2H maintains ~25 GB/s
  • Issue comment#34jinpuwang2025-11-17 11:40
    ./nvbandwidth -t device_to_device_memcpy_read_ce failed due to unknown issue
  • Issue comment#51deepakcu2025-11-15 18:41
    Skip boost static libs on Azure Linux
  • Pull request#52sonic742025-11-15 04:44
  • Pull request#51henryli0012025-11-12 00:51
  • Pull request#49xxrjun2025-09-26 03:35
  • Issue comment#47sammiaobizlink2025-09-23 16:52
    [CUDA_ERROR_INVALID_CONTEXT] with device_to_host_memcpy_ce on native Ubuntu (works fine in WSL2)
  • Issue comment#44pzcyyj2025-09-19 01:59
    multinode benchmarks fail to start with error - [CUDA_ERROR_INVALID_VALUE] invalid argument in expression cuMemSetAccess

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 311 stars here means stars gained during the window, not the repo's star count.