My commit of Feb 7 caused work fetch to project P
to be deferred for up to 5 min if an upload to P is active,
even if some instances are idle.
This was to deal with a case where the idleness was caused
by a jobs-in-progress limit by P,
and work requests lead to long backoff.
However, this can cause instances to be idle unnecessarily.
I changed things so that, if instances are idle,
a work fetch can happen even during upload.
But only one such fetch will be done.
Previously, if a project specified a limit on GPU jobs in progress,
it would be enforced across GPU types.
This could lead to starvation for hosts with multiple GPU types.
E.g. the limit is 10, and a host has 10 NVIDIA jobs and no AMD jobs.
Fix this by enforcing limits separately for each GPU type.
- the exported team list wasn't parsing due to non-ASCII characters
in some team descriptions.
htmlspecialchars() wasn't doing the job.
Use DOMDocument instead.
- If SHOW_NONVALIDATE_TEAMS is set (as on the BOINC-wide teams site)
show teams whose founders email addresses are not validated
(and hence won't be exported).
- Fix a crash bug which prevented "Launch another BOINC Manager" from working
- Fix compiler warnings of possible buffer overflow due to incorrect calls to strncat()
- Adjust directories for per-user files on Macintosh
On some hosts, gpu_active_frac may be much less than active_frac
(i.e., GPUs may be available much less than CPUs).
Use gpu_active_frac in the following places:
- scheduler: in estimating the elapsed time of jobs,
to decide whether they can meet deadline
- scheduler: in computing the effective speed of a (host, app version),
when deciding what size class it belongs to
- size_census: in computing effective speed of (host, app versions)
(Previously, we were just using active_frac in all these cases)
- gpu_active_frac is the fraction of time GPU use is allowed
while the client is running.
Previously the client reported it but we weren't storing it in the DB.
We may need it in the future for batch scheduling logic.
- fix a crashing bug in scheduler
- client: minor message tweak
MGR: Remove what are now over active asserts complaining about pConfig being NULL. Apparently we knew already knew that pConfig could be NULL in certain circumstances and followed up with a check to return from the function if pConfig was NULL, that code was added in 2007. The assert check wasn't really needed after that point.
Charlie Fenton