3 people 11d03eaeef runtime: Add flashinfer rmsnorm + quant fusion support SM90, SM100, SM120- #32994 (#33471) ...

Signed-off-by: Devashish Lal <devcode@fb.com>
Co-authored-by: Devashish Lal <devcode@fb.com>
Co-authored-by: Xiaoyu Zhang <1182563586@qq.com>

2026-08-09 20:15:03 +08:00

..

asr

model: support qwen3-asr (#22073)

2026-04-07 13:27:05 +08:00

bench_attention_sink

[Kernel] Introduce sglang.kernels namespace and migrate scattered triton_ops kernels (RFC #29630, Phase 2) (#30044)

2026-07-10 21:41:08 +08:00

bench_in_batch_prefix

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

bench_linear_attention

[Kernel] RFC #29630 finale: retire sglang.jit_kernel into sglang.kernels (#32072)

2026-07-23 08:35:09 +08:00

bench_pynccl_allocator

Refactor: decouple segment tracking from comm registration (#21392)

2026-05-06 17:07:58 +08:00

bench_rope

Remove smoke wording from tests and comments (#23355)

2026-04-28 12:05:27 +08:00

benchmark_batch

Super tiny update tokenizer benchmark (#16429)

2026-01-05 09:14:52 +08:00

benchmark_vllm_060

Revert "fix some typos" (#6244)

2025-05-12 12:53:26 -07:00

blog_v0_2

Add Lychee Doc Links Check to Local and CI (#19742)

2026-03-24 13:48:26 -07:00

boolq

[NPU]TP Communications compression For Qwen3 models for NPU (#20520)

2026-05-02 14:29:11 +03:00

ceval

fix ceval (#10504)

2025-09-24 02:35:25 +08:00

deepseek_v3

[misc] Use --cuda-graph-max-bs-decode in tests, examples, and docs (#29591)

2026-06-28 18:38:28 -07:00

dspy

Organize Benchmark (#381)

2024-05-05 16:14:17 +08:00

fla

[Kernel] Relocate vendored fla and mamba kernel trees to sglang.kernels (RFC #29630, Phase 2.5, 7/7) (#30795)

2026-07-15 12:52:15 +08:00

generative_agents

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

gpt_oss

[Fix] Add speculative_draft_model_revision to server_args (#5255)

2025-09-05 19:45:46 +08:00

gsm8k

[misc] Improve benchmark determinism and dataset API coverage (#33255)

2026-08-02 01:39:50 -07:00

hellaswag

Simplify frontend language (#9029)

2025-08-10 10:59:30 -07:00

hf3fs

[HiCache] Optimize L2 mem allocation when cache miss in L3 (#19320)

2026-07-16 15:00:53 -07:00

hicache

[Kernel] RFC #29630 finale: retire sglang.jit_kernel into sglang.kernels (#32072)

2026-07-23 08:35:09 +08:00

io

API Perf: Replace pydantic per-element validation with C loop validation (#26355)

2026-05-27 02:04:07 -07:00

json_decode_regex

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

json_jump_forward

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

json_schema

Tiny move files to utils folder (#11166)

2025-10-03 22:40:06 +08:00

kernels

runtime: Add flashinfer rmsnorm + quant fusion support SM90, SM100, SM120- #32994 (#33471)

2026-08-09 20:15:03 +08:00

line_retrieval

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

llava_bench

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

llm_judge

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

long_json_decode

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

lora

benchmark/lora: make number of LoRA adapters configurable (#25363)

2026-05-20 18:23:13 -07:00

mmlu

[New Model] Gemma 4 (#21952)

2026-04-06 20:24:44 -07:00

mmmu

Remove legacy Sphinx docs/ and finish the Mintlify cutover (#28964)

2026-07-13 15:06:08 -07:00

mtbench

[Misc] Auto download question file for benchmark/mtbench (#17019)

2026-01-13 10:34:29 -05:00

multi_chain_reasoning

Bug: fixed multi_chain_reasoning test (#16192)

2026-01-12 09:06:41 -08:00

multi_document_qa

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

multi_turn_chat

Tiny move files to utils folder (#11166)

2025-10-03 22:40:06 +08:00

ocr

Add Accuracy Benchmark for OCR models (#25364)

2026-07-06 16:18:43 +08:00

prefill_only

Add overlap scheduling for embeddings code path (#14032)

2025-12-24 18:24:18 -08:00

react

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

2025-05-11 14:32:49 -07:00

reasoning_benchmark

refactor(benchmark): prevents variable shadowing (#17607)

2026-01-22 17:00:11 -05:00

scheduler

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

tip_suggestion

update pre-commit config (#18860)

2026-02-16 00:18:31 +08:00

tree_of_thought_deep

Revert "fix some typos" (#6244)

2025-05-12 12:53:26 -07:00

tree_of_thought_v0

Revert "fix some typos" (#6244)

2025-05-12 12:53:26 -07:00

bench_adaptive_speculative.py

[SPEC][1/N] feat: add adaptive speculative_num_steps for EAGLE topk=1 (#21599)

2026-04-20 14:25:04 -07:00