NVIDIA / NVIDIA/CUDALibrarySamples

[BUG] Your precompiled binaries include a fellow named Jenkins's file structure, so it looks for files on his system that don't exist locally.

Open
#325 3 comments 0 reactions 1 assignee View on GitHub

@naveenaero is already working on this.

Since Apr 29, 2026.

bug nvCOMP
Dominant language
Cuda
Stars
2.5k
Forks
478
PR merge metrics
No merged PRs in 30d

Description

Describe the bug
I downloaded the binary package from the page instructed on this GitHub, and tried to run the benchmark that I was told I did NOT have to recompile, as it was already included. Well, I tried to run it, and I got:

bin/benchmark_allgather -f ~/tmp/Screenshot_20240205_193240.png
For cascaded, RLEs 1 deltas 0 bitpack 1
Loading data - 2355800 elements, chunk size:1177900, total chunks:2, number of gpus being used:2
API call failure "cudaSetDevice(i)" with 101 at /home/jenkins/agent/workspace/nvcomp-release-artifacts/benchmarks/benchmark_allgather.cpp:128
terminate called without an active exception
[1]    2501533 IOT instruction (core dumped)  bin/benchmark_allgather -f ~/tmp/Screenshot_20240205_193240.png

So, I ran grep -ir jenkins . from the root directory of the repository:

grep: ./lib/libnvcomp.so: binary file matches
grep: ./bin/benchmark_cascaded_chunked: binary file matches
grep: ./bin/benchmark_lz4_synth: binary file matches
grep: ./bin/benchmark_deflate_chunked: binary file matches
grep: ./bin/benchmark_lz4_chunked: binary file matches
grep: ./bin/benchmark_bitcomp_chunked: binary file matches
grep: ./bin/benchmark_zstd_chunked: binary file matches
grep: ./bin/benchmark_allgather: binary file matches
grep: ./bin/benchmark_ans_chunked: binary file matches
grep: ./bin/benchmark_snappy_chunked: binary file matches
grep: ./bin/benchmark_gdeflate_chunked: binary file matches
grep: ./bin/benchmark_hlif: binary file matches
grep: ./bin/benchmark_snappy_synth: binary file matches

So nothing will work because it's looking for files in Jenkins's system where it was compiled.

Steps/Code to reproduce bug
Download the tarball for Linux and Cuda 12.x
Extract it a directory, I called mine nvcomp
Try to run a benchmark, get error.

Expected behavior
The program works as expected

Environment details (please complete the following information):

  • Environment location: Linux bare-metal
  • Method of nvCOMP install: Standard download of binary tarball and extraction.

Additional context
Add any other context about the problem here.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.