NVIDIA / NVIDIA/AMGX

An issue from Fraunhofer-Chalmers Centre: Unit tests

Open
#7 11 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement tracking internally
Dominant language
Cuda
Stars
692
Forks
197
PR merge metrics
No merged PRs in 30d

Description

Which unit tests are relevant and up to date?
We have extensively test on different GPUs, CUDA 8 vs 9.0 and quite a few keep failing under Linux. Besides the missing matrices, which we have filled in from either Matrix Market or proper use of generate_poisson, there are some tests that seem basic and keep failing. Find attached the logs of unit tests on an array of GPUs and configurations.

gtx970_9.0.log
gtx970.log
gtx1080ti.log
gtx660ti.log

All unit tests were compiled with CUDA 8.0, except gtx_970_9.0.log. The tests performed on 660ti were done after we have changed some (possibly too strict? check #8) hardcoded CUDA_ARCH>=3.5 to CUDA_ARCH>=3.0.

We will be more than happy to fix them, as long as we have a list of the tests that are supposed to work, because some seem very outdated or poorly maintained.

Contributor guide

No contributing guide indexed for this repository

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.

Research direction

Review the attached gtx970_9.0.log, gtx970.log, gtx1080ti.log, and gtx660ti.log files alongside the current unit-test suite. Compare failures across CUDA 8/9.0, Linux, GPU configurations, and the CUDA_ARCH change noted with #8. Done means documenting which tests are supported and distinguishing outdated or poorly maintained tests from real failures.

Written by the indexing model from the issue text.

Assessment

Tech stack
linux
Domain
testing-qa
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.