-
Notifications
You must be signed in to change notification settings - Fork 133
Pull requests: ovg-project/kvcached
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix: compare resize target against current allocator capacity
#466
opened Aug 26, 2026 by
jeff3071
Contributor
Loading…
fix: split page create from map to avoid a remap race on Volta
#465
opened Aug 26, 2026 by
andrewleech
Loading…
fix(vllm): defer physical unmap until async batches complete
#464
opened Aug 23, 2026 by
shipiyouniao
Contributor
Loading…
fix(sglang): Patch MambaSlotAllocator on sglang 0.5.13+
#459
opened Aug 21, 2026 by
jeff3071
Contributor
Loading…
docs: explain revisioned memory limit integration
#458
opened Aug 21, 2026 by
shipiyouniao
Contributor
Loading…
perf(kv-cache): cache get_avail_physical_pages in available_size to skip per-alloc cudaMemGetInfo
#456
opened Aug 19, 2026 by
SuperMarioYL
Contributor
Loading…
fix(kv-cache): size pages via get_block_range in _get_num_alloced_blocks
#452
opened Aug 18, 2026 by
SuperMarioYL
Contributor
Loading…
Add page-granular CPU offloading foundation
#450
opened Aug 17, 2026 by
Lanoxia
Contributor
Loading…
Add manually triggered model compatibility matrix
#440
opened Aug 10, 2026 by
Lanoxia
Contributor
Loading…
Fix vLLM padded KV page allocation and strides
#426
opened Aug 5, 2026 by
shipiyouniao
Contributor
Loading…
Add automated vLLM and SGLang upstream synchronization
#423
opened Aug 4, 2026 by
Lanoxia
Contributor
Loading…
Export kvcached metrics through SGLang's native endpoint
#421
opened Aug 3, 2026 by
shipiyouniao
Contributor
Loading…
Add portable self-hosted GPU CI profiles
gpu-ci
#420
opened Jul 31, 2026 by
Lanoxia
Contributor
Loading…
fix: make KV tensor VMM batches transactional
#418
opened Jul 30, 2026 by
shipiyouniao
Contributor
Loading…
feat(vllm): avoid CUDA context in EngineCore
#416
opened Jul 30, 2026 by
shipiyouniao
Contributor
•
Draft
fix(sglang): keep KV pool ownership local to each TP worker
#415
opened Jul 30, 2026 by
shipiyouniao
Contributor
Loading…
fix: bind worker IPC listener to its CUDA device
#413
opened Jul 28, 2026 by
shipiyouniao
Contributor
Loading…
fix(controller): don't fail the losing racer of a concurrent wakeup_model() call
#409
opened Jul 23, 2026 by
AmirF194
Contributor
Loading…
feat(sglang): pool DeepSeek-V4 SWA KV cache
#400
opened Jul 21, 2026 by
shipiyouniao
Contributor
•
Draft
fix(sglang): use storage dtype for FP8 KV pools
#388
opened Jul 16, 2026 by
shipiyouniao
Contributor
Loading…
Fix vLLM Gemma4 support by adding v2 module fallback and patch target fallback
#386
opened Jul 14, 2026 by
mahendrarathore1742
Contributor
Loading…
fix(sglang): bound paged free-index reduction cost
#384
opened Jul 5, 2026 by
shipiyouniao
Contributor
Loading…
feat(sglang): account DeepSeek-V4 runtime reservations
#376
opened Jul 1, 2026 by
shipiyouniao
Contributor
•
Draft
Previous Next
ProTip!
Type g p on any issue or pull request to go back to the pull request listing page.