Skip to content

Conversation

@djw8605
Copy link
Collaborator

@djw8605 djw8605 commented Mar 26, 2022

The CPUs and GPUs are floats, for some reason. Convert to integer
which makes it easier to parse with python classad bindings (for example
in the gratia probe)

The gratia probe reads MATCH_EXP_JOB_GLIDEIN_Cpus.

I am open to opinions.

The CPUs and GPUs are floats, for some reason.  Convert to integer
which makes it easier to parse with python classad bindings (for example
in the gratia probe)
@timtheisen
Copy link
Member

So, in the near future we will be allocating fractional GPUs. And, allocating fractional CPUs in on the road map. What is the difficulty in parsing with the Python ClassAd bindings? Is there a bug there that we need to fix?

@tannenba
Copy link

Seems like a problem that the probe relies on MATCH_EXP_JOB_GLIDEIN_Cpus, which is not ensured to be there. The job should explicitly request to have attributes from the matching ad (slot ad) inserted into the job ad. As for converting to int, the patch looks simple enough, but we are wondering.... how does it become easier to parse with ClassAd bindings? Or perhaps you meant easier to parse without bindings? Or easier to work with? Thanks.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants