High Performance LLM Inference Operator Library
active 2026-01-22 → 2026-08-07 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 198 days from 2026-01-22 to 2026-08-07. Pushes: 11 total, peak 3 in a day. Pull requests: 9 total, peak 2 in a day. Issues: 13 total, peak 4 in a day. Comments: 14 total, peak 3 in a day. Stars: 245 total, peak 40 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| reed-lau | 28 | 11 | 0 | 11 |
| oraluben | 2 | 0 | 2 | 0 |
| Echo-Nie | 2 | 0 | 0 | 0 |
| xueyangcs | 2 | 0 | 2 | 0 |
| weishengying | 2 | 0 | 2 | 0 |
| WingEdge777 | 1 | 0 | 0 | 1 |
| XzenithAI | 1 | 0 | 0 | 0 |
| qibaoyuan | 1 | 0 | 0 | 0 |
| TheoCheng100 | 1 | 0 | 1 | 0 |
| Nekofish-L | 1 | 0 | 0 | 0 |
| shaochangxu | 1 | 0 | 1 | 0 |
| JiaxiangZheng | 1 | 0 | 0 | 0 |
| ZelinMa557 | 1 | 0 | 0 | 1 |
| K11OntheBoat | 1 | 0 | 0 | 0 |
| userstarwind | 1 | 0 | 0 | 1 |
| Ageliss | 1 | 0 | 0 | 0 |
| bcacdwk | 1 | 0 | 0 | 0 |
| xiaoqi35 | 1 | 0 | 0 | 0 |
| lhtin | 1 | 0 | 1 | 0 |
| qqbbiu | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#49reed-lau2026-06-09 04:57add fused fp8 moe kernel for low-latency llm inference
- Issue comment#41ZelinMa5572026-05-09 09:44feat: bf16 fused moe kernel
- Pull request#39xueyangcs2026-04-13 01:56
- Issue comment#41reed-lau2026-04-10 03:18feat: bf16 fused moe kernel
- Pull request#40xueyangcs2026-04-03 08:19
- Pull request#36TheoCheng1002026-04-01 07:36
- Issue comment#32reed-lau2026-03-11 03:38是否可以添加一个bench目录验证算子性能？
- Issue#30XzenithAI2026-02-20 03:05🤖 AI Prompt Pack ($5)
- Issue#28NiuMa-12342026-02-11 07:35[Performance] GroupedGEMM seems not to be faster than DeepGEMM
- Issue comment#26userstarwind2026-02-10 03:23[bug] test_attention_decode_bf16 can't pass with block_size=32
- Issue comment#26WingEdge7772026-02-09 11:30[bug] test_attention_decode_bf16 can't pass with block_size=32
- Issue#15Echo-Nie2026-02-05 07:02test_attention_decode_bf16 fails
- Issue comment#18reed-lau2026-02-05 03:49能不能列个表，attention kernel支持哪些GQA shape？
- Issue#24qibaoyuan2026-02-05 02:47vllm integrations
- Pull request#22lhtin2026-02-04 08:31
- Issue#19K11OntheBoat2026-02-02 06:50test_fused_moe_blockwise.py 改了一下shape，单测过不了
- Issue#17xiaoqi352026-02-02 06:04A question: these attention interface support MLA?
- Issue comment#15reed-lau2026-01-29 07:25[Error] test_attention_decode_bf16 fails
- Issue#15Echo-Nie2026-01-29 03:17[Error] test_attention_decode_bf16 fails
- Issue comment#14reed-lau2026-01-28 10:55[Performance] Decode performance lower than DeepGEMM in small-M scenario
- Pull request#13weishengying2026-01-28 10:48
- Pull request#13weishengying2026-01-28 05:19
- Issue#12Ageliss2026-01-27 16:37[Feature Req] Any plan to integrate into SGLang?
- Issue comment#2reed-lau2026-01-27 11:17test case failed for attention decode on H20
- Issue#11JiaxiangZheng2026-01-27 09:09[Feature Request] Support for additional head_dim configurations (e.g., 64, 96)
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 245 stars here means stars gained during the window, not the repo's star count.