Repository navigation

feat(nvsnap): import gpushare, checkpoint support for GPU memory shared between processes - #2300

Open

balajinvda wants to merge 58 commits into

nvsnap/helm-artifactsfrom

nvsnap/gpushare-import

Commits on Oct 4, 2026
- refactor(nvsnap): retire the legacy CRIU engine and the preload injection stack
  balajinvda
  committed
  c187fd3
Commits on Oct 5, 2026
- Merge remote-tracking branch 'origin/nvsnap/helm-artifacts' into nvsnap/criu-v2-only
  balajinvda
  committed
  8d2e05e
- Merge remote-tracking branch 'origin/nvsnap/helm-artifacts' into nvsnap/criu-v2-only
  balajinvda
  committed
  f7ea840
Commits on Oct 6, 2026
- feat(nvsnap): import gpushare, checkpoint support for GPU memory shared between processes
  
  balajinvda
  and
  claude
  committed
  6c76c8b
- fix(nvsnap): address gpushare review findings
  
  balajinvda
  and
  claude
  committed
  4c35167
- docs(nvsnap): state gpushare store trust in terms of writers, not tenants
  
  balajinvda
  and
  claude
  committed
  191b2f3
- Merge remote-tracking branch 'origin/nvsnap/helm-artifacts' into nvsnap/criu-v2-only
  balajinvda
  committed
  cc1141e
- build(nvsnap): regenerate BUILD files, drop the lz4 v4 repo import
  balajinvda
  committed
  f085f98
- Merge remote-tracking branch 'origin/nvsnap/criu-v2-only' into nvsnap/gpushare-import
  balajinvda
  committed
  641fe48
- ci(nvsnap): compile the gpushare C code for amd64 and arm64 (#2304)
  balajinvda
  authored
  52113b0
- fix(nvsnap): roll back a failed gpushare release exactly
  
  balajinvda
  and
  claude
  committed
  fbb9e9b
- fix(nvsnap): gate gpushare memory operations, harden suspend tool
  
  balajinvda
  and
  claude
  committed
  45a3b10
- build(nvsnap): base image v0.0.24 with the gpushare review fixes
  
  balajinvda
  and
  claude
  committed
  725f9a6
- Merge nvsnap/helm-artifacts into nvsnap/gpushare-import
  balajinvda
  committed
  33e13c2
- fix(nvsnap): point export fds the app keeps at /dev/null on release
  
  balajinvda
  and
  claude
  committed
  e5d8d1a
Commits on Oct 7, 2026
- feat(nvsnap): checkpoint pods with an IMEX channel
  
  balajinvda
  and
  claude
  committed
  04b3927
- feat(nvsnap): checkpoint memory shared across nodes over NVLink
  
  balajinvda
  and
  claude
  committed
  0c4c0c1
- fix(nvsnap): fabric imports across nodes after a restore
  
  balajinvda
  and
  claude
  committed
  2ab21e4
- feat(nvsnap): checkpoint and restore tensor-parallel workloads through gpushare (#2318)
  balajinvda
  authored
  477cccc
- feat(nvsnap): drive multi-node gpushare suspend and resume from the agent
  balajinvda
  committed
  d0e005d
- build(nvsnap): base image v0.0.25 with the multi-node gpushare fixes
  balajinvda
  committed
  1066fbf
Commits on Oct 8, 2026
Commits on Oct 9, 2026
Commits on Oct 10, 2026

Commits

Commits on Oct 4, 2026

Commits on Oct 5, 2026

Commits on Oct 6, 2026

Commits on Oct 7, 2026

Commits on Oct 8, 2026

Commits on Oct 9, 2026

Commits on Oct 10, 2026