Search results

SGLang paged allocator free path should move indices to CPU before torch.unique

Status: Open.

#382 In ovg-project/kvcached;

· shipiyouniao opened on Jul 5, 2026

FTensor map/unmap should tolerate idempotent page operations

Status: Open.

#381 In ovg-project/kvcached;

· shipiyouniao opened on Jul 5, 2026

Explicit KVCACHED_IPC_NAME should not be auto-suffixed

Status: Open.

#378 In ovg-project/kvcached;

· shipiyouniao opened on Jul 2, 2026

Discussion: stable extension points for production kvcached integrations

feature requested

Status: Open.

#375 In ovg-project/kvcached;

· shipiyouniao opened on Jul 1, 2026

TP KV tensor map failures can be hidden behind generic wait timeouts

Status: Open.

#372 In ovg-project/kvcached;

· shipiyouniao opened on Jul 1, 2026

Page-prealloc deadlock: a co-located kvcached engine hangs in Init C++ PageAllocator and never finishes startup

bugSomething isn't workingSomething isn't working

Status: Open.

#371 In ovg-project/kvcached;

· jiarong0907 opened on Jun 30, 2026

Support Gemma4 model

feature requested

Status: Open.

#370 In ovg-project/kvcached;

· jiarong0907 opened on Jun 30, 2026

Feature request: support DeepSeek-V4 KV cache pooling in vLLM and SGLang

feature requested

Status: Open.

#369 In ovg-project/kvcached;

· shipiyouniao opened on Jun 30, 2026

KVCacheManager does not roll back partial allocation when alloc_page fails

Status: Open.

#364 In ovg-project/kvcached;

· shipiyouniao opened on Jun 24, 2026

vLLM prefix cache blocks keep stale block_hash after eviction

Status: Open.

#362 In ovg-project/kvcached;

· shipiyouniao opened on Jun 24, 2026

Idle instance pins GBs beyond MAX_CACHED_TOKENS due to cached-block fragmentation across 2MB pages, starving co-located instances

Status: Open.

#359 In ovg-project/kvcached;

· cui36 opened on Jun 10, 2026

vLLM ≥ 0.21/0.22 support

feature requested

Status: Open.

#358 In ovg-project/kvcached;

· cui36 opened on Jun 10, 2026