[BUG]: catch2 binary runs into frequent time outs

[miscco]

Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this bug and that I agree to the Code of Conduct

Type of Bug

Something else

Component

CUB

Describe the bug

Our catch2 test binary has grown over time.

We are now running into timeouts during CI which frequently require manual intervention / rerunning of jobs.

We should consider splitting the catch2 binary into one per test to lessen the burden on the test infrastructure. There is a build configuration for that, so it should not be too invasive

How to Reproduce

Look at Ci with a red x

Expected behavior

Our tests should not run into timeouts

Reproduction link

No response

Operating System

No response

nvidia-smi output

No response

NVCC version

No response

[jrhemstad]

I haven't seen the timeouts. Can you link to a CI run where that is happening?

[miscco]

Nope, it seems the history of previous runs is not preserved (I usually just go on rerun failed)

[jrhemstad]

The history is preserved. Can you link me to a job that you reran?

[miscco]

Ah I got it https://github.com/NVIDIA/cccl/actions/runs/5888238667/attempts/1

[alliepiper]

We can hardcode a higher timeout in cmake. Is this still an issue?

[miscco]

It hasnt been that prominent recently

lets close and reopen if it becomes a problem again