Skip to content

MB-72262: Probe GPU devices for health #69

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
CascadingRadium merged 2 commits into from
Jun 5, 2026
Merged

MB-72262: Probe GPU devices for health #69

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fc3c920
Probe GPU devices for health
CascadingRadium Jun 3, 2026
11119bb
fixup API
CascadingRadium Jun 3, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 11 additions & 7 deletions gpu.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -214,14 +214,18 @@ type gpuSnapshotStore struct {
func newGPUSnapshotStore() *gpuSnapshotStore {
snapshots := make([]*gpuSnapshot, gpuCount)
for device := 0; device < gpuCount; device++ {
// get total free memory for the GPU device.
cDev := C.int(device)
totMemory := uint64(0)
var freeBytes C.size_t
if c := C.faiss_gpu_free_memory(
C.int(device),
&freeBytes,
); c == 0 {
totMemory = uint64(freeBytes)
// first probe if the device is healthy and can be used
var probeResult C.int
if c := C.faiss_probe_gpu(cDev, &probeResult); c == 0 && probeResult == 0 {
var freeBytes C.size_t
if c := C.faiss_gpu_free_memory(
Comment thread
CascadingRadium marked this conversation as resolved.
cDev,
&freeBytes,
); c == 0 {
totMemory = uint64(freeBytes)
}
}
Comment thread
CascadingRadium marked this conversation as resolved.
// if we fail to get the free memory for the GPU,
// we still create a snapshot with 0 total and free memory,
Expand Down