feat(hardware): use NVML to grab the hardware profile during the regi… #314
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
|
|
@@ -3,7 +3,7 @@ before: | |
| - go mod download | ||
| builds: | ||
| - env: | ||
| - CGO_ENABLED=1 | ||
| goos: | ||
| - darwin | ||
| - linux | ||
|
|
||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,158 @@ | ||
| //go:build linux || windows | ||
|
|
||
| package register | ||
|
|
||
| import ( | ||
| "fmt" | ||
|
|
||
| "github.com/NVIDIA/go-nvml/pkg/nvml" | ||
| ) | ||
|
|
||
| // archNames maps NVML compute capability (major version) to GPU architecture name. | ||
| var archNames = map[int]string{ | ||
|
Contributor
There was a problem hiding this comment.
|
||
| 1: "Tesla", | ||
| 2: "Fermi", | ||
| 3: "Kepler", | ||
| 5: "Maxwell", | ||
| 6: "Pascal", | ||
| 7: "Volta/Turing", | ||
| 8: "Ampere", | ||
| 9: "Hopper/Ada Lovelace", | ||
| 10: "Blackwell", | ||
| 12: "Vera Rubin", | ||
| } | ||
|
|
||
| // probeGPUsNVML uses NVML to detect GPUs and interconnects. | ||
| // Returns (nil, nil) if NVML is unavailable (e.g. no driver installed). | ||
| func probeGPUsNVML() ([]GPU, []Interconnect) { | ||
| ret := nvml.Init() | ||
| if ret != nvml.SUCCESS { | ||
| return nil, nil | ||
| } | ||
| defer func() { _ = nvml.Shutdown() }() | ||
|
|
||
| count, ret := nvml.DeviceGetCount() | ||
| if ret != nvml.SUCCESS || count == 0 { | ||
| return nil, nil | ||
| } | ||
|
|
||
| type gpuKey struct { | ||
| model string | ||
| arch string | ||
| mem int64 | ||
| } | ||
| counts := make(map[gpuKey]int32) | ||
| var order []gpuKey | ||
| var interconnects []Interconnect | ||
|
|
||
| for i := 0; i < count; i++ { | ||
| device, ret := nvml.DeviceGetHandleByIndex(i) | ||
| if ret != nvml.SUCCESS { | ||
| continue | ||
| } | ||
|
|
||
| name, ret := device.GetName() | ||
| if ret != nvml.SUCCESS { | ||
| name = "Unknown" | ||
| } | ||
|
|
||
| var memBytes int64 | ||
| memInfo, ret := device.GetMemoryInfo() | ||
| if ret == nvml.SUCCESS { | ||
| memBytes = int64(memInfo.Total) | ||
| } | ||
|
|
||
| arch := "" | ||
| major, minor, ret := device.GetCudaComputeCapability() | ||
| if ret == nvml.SUCCESS { | ||
| if archName, ok := archNames[major]; ok { | ||
| arch = archName | ||
| } else { | ||
| arch = fmt.Sprintf("sm_%d%d", major, minor) | ||
| } | ||
| } | ||
|
|
||
| key := gpuKey{model: name, arch: arch, mem: memBytes} | ||
| if counts[key] == 0 { | ||
| order = append(order, key) | ||
| } | ||
| counts[key]++ | ||
|
|
||
| // Probe NVLink interconnects for this device. | ||
| interconnects = append(interconnects, probeNVLink(device, i)...) | ||
|
|
||
| // Probe PCIe interconnect for this device. | ||
| if ic := probePCIe(device, i); ic != nil { | ||
| interconnects = append(interconnects, *ic) | ||
| } | ||
| } | ||
|
|
||
| gpus := make([]GPU, 0, len(order)) | ||
| for _, key := range order { | ||
| mem := key.mem | ||
| g := GPU{ | ||
| Model: key.model, | ||
| Architecture: key.arch, | ||
| Count: counts[key], | ||
| } | ||
| if mem > 0 { | ||
| g.MemoryBytes = &mem | ||
| } | ||
| gpus = append(gpus, g) | ||
| } | ||
|
|
||
| return gpus, interconnects | ||
| } | ||
|
|
||
| // probeNVLink checks NVLink connections for a device. | ||
| func probeNVLink(device nvml.Device, deviceIdx int) []Interconnect { | ||
| var ics []Interconnect | ||
| activeLinks := 0 | ||
|
|
||
| // NVLink link count varies by architecture; try up to 18 links. | ||
| var nvlinkVersion uint32 | ||
| for link := 0; link < 18; link++ { | ||
| state, ret := device.GetNvLinkState(link) | ||
| if ret != nvml.SUCCESS { | ||
| break | ||
| } | ||
| if state == nvml.FEATURE_ENABLED { | ||
| activeLinks++ | ||
| if nvlinkVersion == 0 { | ||
| ver, ret := device.GetNvLinkVersion(link) | ||
| if ret == nvml.SUCCESS { | ||
| nvlinkVersion = uint32(ver) | ||
|
Check failure on line 124 in pkg/cmd/register/gpu_nvml.go
|
||
| } | ||
| } | ||
| } | ||
| } | ||
|
|
||
| if activeLinks > 0 { | ||
| ics = append(ics, Interconnect{ | ||
| Type: "NVLink", | ||
| Device: fmt.Sprintf("GPU %d", deviceIdx), | ||
| ActiveLinks: activeLinks, | ||
| Version: nvlinkVersion, | ||
| }) | ||
| } | ||
|
|
||
| return ics | ||
| } | ||
|
|
||
| // probePCIe reads PCIe generation and width for a device. | ||
| func probePCIe(device nvml.Device, deviceIdx int) *Interconnect { | ||
| gen, ret := device.GetCurrPcieLinkGeneration() | ||
| if ret != nvml.SUCCESS { | ||
| return nil | ||
| } | ||
| width, ret := device.GetCurrPcieLinkWidth() | ||
| if ret != nvml.SUCCESS { | ||
| return nil | ||
| } | ||
| return &Interconnect{ | ||
| Type: "PCIe", | ||
| Device: fmt.Sprintf("GPU %d", deviceIdx), | ||
| Generation: gen, | ||
| Width: width, | ||
| } | ||
| } | ||
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We now always need CGO