All issues

Search results

[BUG] replay_increment fails mid-write after cache flush expansion

bugSomething isn't workingSomething isn't working

Status: Open.

#494 In NVIDIA/recsys-examples;

· honlzl opened on Sep 25, 2026

[BUG] SM120 dense HSTU smem overflow

bugSomething isn't workingSomething isn't working

Status: Open.

#492 In NVIDIA/recsys-examples;

· Clebrate opened on Sep 21, 2026

Split key_value_table.py into multiple files to improve maintainability.

Status: Open.

#483 In NVIDIA/recsys-examples;

· jiashuy opened on Sep 10, 2026

[PERF][RFC] Explore a shared-memory AOTI + FlexKV integration for Triton

Status: Open.

#475 In NVIDIA/recsys-examples;

· shijieliu opened on Sep 1, 2026

[FEA] Add KV-cache-aware Triton orchestration and move offload/reap off the inference critical path`

Status: Open.

#474 In NVIDIA/recsys-examples;

· shijieliu opened on Sep 1, 2026

[FEA] Add CUDA Graph replay to the HSTU KV-cache AOTI compute path`

Status: Open.

#473 In NVIDIA/recsys-examples;

· shijieliu opened on Sep 1, 2026

[BUG] FlexKV SSD-to-host transfer bandwidth is lower than expected

bugSomething isn't workingSomething isn't working

Status: Open.

#472 In NVIDIA/recsys-examples;

· geoffreyQiu opened on Sep 1, 2026

[FEA] Overlap SSD-to-host and host-to-device KV-cache transfers in FlexKV

Status: Open.

#471 In NVIDIA/recsys-examples;

· geoffreyQiu opened on Sep 1, 2026

[FEA] Support configurable table-fusion indexing and NVE layer types for exportable inference embeddings

Status: Open.

#470 In NVIDIA/recsys-examples;

· geoffreyQiu opened on Sep 1, 2026

Development Roadmap(July/Aug/Sep)

Status: Open.

#462 In NVIDIA/recsys-examples;

· shijieliu opened on Aug 10, 2026

[FEA] Optimize online request scheduling for HSTU inference

Status: Open.

#461 In NVIDIA/recsys-examples;

· shijieliu opened on Aug 10, 2026

[FEA] Add an online serving benchmark for HSTU inference

Status: Open.

#460 In NVIDIA/recsys-examples;

· shijieliu opened on Aug 10, 2026