xpra icon
Bug tracker and wiki

Opened 13 months ago

Closed 7 months ago

#2416 closed enhancement (worksforme)

CUDA device selection by name

Reported by: Antoine Martin Owned by: Antoine Martin
Priority: major Milestone: 4.0
Component: encodings Version: 3.0.x
Keywords: Cc:

Description

Follow up for #2415: we should be able to select the list of enabled / disabled devices using their name and / or PCI bus location (ie: 0000:1A:00.0) without doing a full CUDA context initialization.

load_device can get to the device name before calling make_context (which is slow).

Change History (5)

comment:1 Changed 13 months ago by Antoine Martin

Status: newassigned

We can't skip the make_context completely in select_device because driver.mem_get_info() requires a valid context.

Maybe we can have different policies for load balancing, and one may not require memory info to do its job.

Last edited 13 months ago by Antoine Martin (previous) (diff)

comment:2 Changed 11 months ago by Antoine Martin

Updates:

  • r24331: can use device name or PCI id to enable or disable devices
  • r24332 + r24333: update examples (maybe we should ignore the PCI prefix? or allow a pattern match?)

comment:3 Changed 11 months ago by Antoine Martin

r24340 adds a load-balancing attribute and a round-robin option, the memory policy is still the default. (chooses the GPU with the most free memory)

Still needs testing.

comment:4 Changed 10 months ago by Antoine Martin

Seems to have caused a bug: #2489.

See also initial load balancing code in #520.

Last edited 10 months ago by Antoine Martin (previous) (diff)

comment:5 Changed 7 months ago by Antoine Martin

Resolution: worksforme
Status: assignedclosed
Note: See TracTickets for help on using tickets.