minke.yu
71c6d8bc55
G9 P0 fix: /query_cache_hit 绕过 _tokenize_one_request 的 ReqState 依赖
...
_tokenize_one_request._create_tokenized_object 会读 rid_to_state[obj.rid]
(正常生成路径在 tokenize 前注册 ReqState),纯查询无 rid → KeyError: None。
改为:dsv41 chat encoding 直接渲染的 prompt_ids 直接用;否则对渲染后
text 调 _tokenize_texts(与真实路径同一分词入口)。tok8 环境实测 200。
2026-10-11 15:38:35 +08:00
minke.yu
baba18bc7d
hicache L3 prefetch: io_aux 传输并行化(按 op 分 lane)
...
build-sglang-image / build (push) Failing after 48m34s
打点实测单条 prefetch_io_aux_func 线程串行传输是所有请求的瓶颈:
mc8 下每请求 xfer 排队 0->1.9s 线性增长(每请求 5 批 x ~68ms 固定开销,
介质无关)。本提交把传输层拆成 N 条并行 lane(默认 4,
SGLANG_HICACHE_PREFETCH_IO_THREADS 可调):
- dispatch_prefetch() 用 crc32(request_id) % N 分 lane(内置 hash 对 str
每进程加盐,跨 rank 不一致会导致 gloo 集合通信错位,禁用)
- 同一 op 的所有 chunk 保持在同一 lane,保序 reduce 语义不变
- 每 lane 独立 io 队列 + io_aux 线程 + sync 队列 + sync 线程 +
独立 gloo completion 组(集合通信按 lane 过滤后的分发顺序保序)
- query 阶段的 hit_reduce 仍在单 prefetch_thread,不动
- 生产点(hiradix/unified/hybrid pp-ticket)改调 dispatch_prefetch
- 模拟器与两个单测文件同步适配
注意:SGLANG_HICACHE_PREFETCH_IO_THREADS 必须在实例内所有
TP/CP/PP rank 保持一致,否则 lane 组上的集合通信会错位挂死。
2026-10-10 19:04:49 +08:00
minke.yu
7d5759b2d9
hicache: STORAGE_BATCH_SIZE 可env化(SGLANG_HICACHE_STORAGE_BATCH_SIZE,默认128不变)
...
build-sglang-image / build (push) Failing after 18m40s
打点显示每批存储 IO 有 ~30-70ms 固定开销(与传输介质无关),
调大批次可摊薄。env 化便于不动代码做 128/512/1024 对照。
2026-10-10 17:55:14 +08:00
minke.yu
84fc632088
mooncake_store: local_hostname 加 rank 后缀,让 master 的 preferred-segment 分配落到本进程段
...
build-sglang-image / build (push) Failing after 1s
两个 rank 进程共用默认 localhost 时,master 按名字匹配无法区分两个段,
副本会落到对方进程段;mooncake 的 local memcpy 快路径要求同进程 endpoint
(跨进程虚拟地址无效),于是静默回退 loopback TCP(实测 ~0.78GB/s)。
唯一化后自写自读可命中 memcpy(干净集群探针实测 lo 流量归零)。
注意:memcpy 线程池 kDefaultMemcpyWorkers=1(mooncake 0.3.13 硬编码),
单流 ~1GB/s,进一步提升需改 mooncake 并重编译。
2026-10-10 16:42:00 +08:00
minke.yu
5f2ffa70b8
hicache: SGLANG_HICACHE_TIMING 打点 L3 预取全链路分阶段耗时
...
build-sglang-image / build (push) Failing after 1s
env 开关(默认关),开后在以下阶段输出 HC_TIMING 日志行:
query_wait/query(含 batch_exists 逐批)/hit_count all_reduce/
xfer_wait/xfer(含 batch_get 逐批)/ack_reduce/ack_drain 排队/
graft(insert_host)/prefetch_e2e/loadback。
目的:定位 L3 命中单请求 ~3s(≈重算速度)的时间分布。
真实 mooncake store 与 dummy store 实测吞吐一致,排除后端实现因素。
2026-10-10 15:29:01 +08:00
minke.yu
fa8ad18b5c
手工移植上游 #38348:hicache SWA prefetch fail-soft 降级
...
build-sglang-image / build (push) Failing after 1s
SWA 池 L3 取回短缺时不再整批丢弃 prefetch 结果:保留已取回的 Full KV
前缀,仅丢弃失败的 SWA 池(释放其 host buffer、从 comp_xfers 摘除),
load-back 时经既有 SWARebuild/RecoverSWAWithLockedFull 路径从 Full KV
重建窗口。Mamba 等不可重建池短缺时仍走原 all-or-nothing 整批丢弃。
配套改动(同 PR):
- swa.py validator:SWA 双层缺失但节点已 backuped(Full KV 在 host)时
放行作为匹配边界;LOAD_BACK 收集遇双层缺失空洞时 break 而非 assert。
- prefetch_from_storage:用完整匹配前缀重算 last_hash,修复 anchor 节点
无 host_value(异步 backup 未完成/降级嫁接)时哈希链锚错位。
- 适配差异:本树有 sidecar_hash_values 重构,belief 自愈沿用
invalidation_hashes + trimmed_pages 语义;ongoing_prefetch 以
CacheRequestHandle 为键,comp_xfers 改为显式参数传入检查函数。
- 单测同步更新为 fail-soft 语义(TP 一致性降级、失败池释放无泄漏、
LOAD_BACK 锚定空洞停止收集)。
2026-10-10 07:23:15 +08:00
minke.yu
7f1edbba80
cherry-pick 上游 #40429:FlashMLA sched meta 超 48KiB smem 时跳过预计算(encoder/decoder replay 的大 batch replay pass 会触发 smem 溢出崩溃)
build-sglang-image / build (push) Failing after 1s
2026-10-10 01:20:15 +08:00
minke.yu
99c317b131
cherry-pick 上游 #42716(运行时部分):encoder SWA bounded replay 下放行 host 层前缀命中——request_window 形态的 SWA 不进树,校验门不应拦截 host-only 节点;rust adapter/test 两处 hunk 与本树结构不符未合(默认 python 树芯用不到)
build-sglang-image / build (push) Failing after 1s
2026-10-10 01:02:33 +08:00
minke.yu
b4dbcebbec
mooncake standalone(dummy) 模式打通 dsv41 unified cache:放宽 standalone 检查对 LogicalHostPool 锚点的误拦(实体 buffer 由旁池运行时 register_buffer);新增 l3dummy compose(单 store 服务 64GB shm 段 + fake-cargo 跳过 rustup 下载)。实测 24x160k 热/冷 +26.4%(vs TCP best_effort -4.4%)
build-sglang-image / build (push) Failing after 1h10m35s
2026-10-10 00:06:32 +08:00
minke.yu
4eabf5e228
G9 P0: loads 排队预估字段 + /query_cache_hit 命中查询端点
...
A. 负载预估数据源(决策②总输入长度口径):
- LoadSnapshot 新增 queued_total_input_tokens / drain_rate_tokens_per_s
- load_inquirer 复用 pending_token_queues 求和;scheduler 60s 滑动窗
记录每请求最终 prefill chunk 完成事件(batch_result_processor 挂钩
set_prefill_finished_time 处,恰好每请求一次)
- /v1/loads 与 PUB socket 同源自动暴露,gateway LoadMonitor 零改动消费
B. /query_cache_hit(决策④放 D):
- io_struct 新增 QueryCacheHitReqInput/Output IPC 类型
- scheduler.handle_query_cache_hit:tree_cache.match_prefix(L1/L2)
+ query_storage_hit_length(共享 L3,逻辑镜像 decode_hicache_mixin)
- tokenizer_control_mixin:fan-out 全部 DP rank 取最优命中
- serving_chat.query_cache_hit:复用 _convert_to_internal_request +
_tokenize_one_request 保证 chat 模板/分词与真实路径一致;含图请求
v1 拒绝(multimodal 400);http_server 注册 /query_cache_hit
2026-10-09 18:48:27 +08:00
minke.yu
7578cdd218
revert: #42035 continuous prefill input polling (CP rank collective mismatch deadlock)
...
build-sglang-image / build (push) Failing after 39m1s
compose 3p1d regression on 1f6ca2cf6 : all 3 P nodes wedged with split-brain
CP ranks (one blocked in _sync_hicache_ready_counts all_reduce, partner in
request broadcast). Root cause: #42035 wraps get_next_disagg_prefill_batch_to_run
(which contains the hicache attn_cp_group all_reduce) in rank-local guards
(len(result_queue)<2, _deferred_input_requests, skip_batch_creation); any
cross-rank divergence in batch-None decisions misaligns collective pairing
permanently. Continuous polling requires attn_cp_size==1 anyway, so cp2
deployments get zero benefit. Reverts 93d610c1fd + test adaptations
6ba34507f1/bde1fcbf51/1f6ca2cf68 + deploy bump 236686454c .
Keeps #42194 (9ef65b51c2 ) and checkpoint_disagg_prefill shim (c4f3e3be33 ).
Evidence: D:\B300\investigations\pd-deadlock-20261009.md
2026-10-09 16:39:06 +08:00
minke.yu
c4f3e3be33
refactor: route disagg prefill checkpoints through checkpoint_disagg_prefill shim
...
Upstream #42035 's tests and call structure expect Scheduler.checkpoint_disagg_prefill
(introduced upstream with #41520 's checkpoint_kv_cache / insert_req cache API).
This branch still uses maybe_cache_unfinished_req (old cache_unfinished_req API),
so add a thin shim with the upstream name and route the three call sites through it.
No behavior change.
2026-10-09 11:22:42 +08:00
Liangsheng Yin
9ef65b51c2
[mem_cache] Skip the release-time insert on optimistic prefill requeue; make refresh_fill_ids public ( #42194 )
...
build-sglang-image / build (push) Successful in 25m53s
(cherry picked from commit 07e1a4f082 )
2026-10-09 11:18:17 +08:00
metamergebot
93d610c1fd
[PD] Keep ingesting requests while a prefill forward result is pending ( #42035 )
...
build-sglang-image / build (push) Successful in 26m6s
Co-authored-by: merrymercy <merrymercy@users.noreply.github.com >
Co-authored-by: Jialin <Jialin@users.noreply.github.com >
(cherry picked from commit 6acf2f4736 )
2026-10-09 10:56:14 +08:00
minke.yu
1068926cd6
fix: import is_hip_runtime in dsv4 low_ratio_compress (cherry-pick #41660 lost it)
...
build-sglang-image / build (push) Successful in 32m18s
Upstream's file gets is_hip_runtime from sglang.kernels.jit.utils via an
intermediate refactor we don't have; our cherry-pick applied cleanly but
the resulting file references is_hip_runtime() without importing it,
crashing every prefill (NameError in c2_prefill_norm_rope_store).
Also unify Dockerfile.fast with the v3.1 full path (check_deps + --no-deps
delta installs) so fast builds after a kernel-wheel bump don't silently
revert to the base image's older kernels.
2026-10-08 16:31:34 +08:00
metamergebot
75758cb0a0
[HiCache] Fix cgroup page-cache accounting and the sizing fallback when cgroup discovery fails ( #42420 )
...
Co-authored-by: merrymercy <merrymercy@users.noreply.github.com >
Co-authored-by: xiezhq-hermann <xiezhq-hermann@users.noreply.github.com >
Co-authored-by: nvjullin <nvjullin@users.noreply.github.com >
(cherry picked from commit 4ab720e655 )
2026-10-08 15:41:51 +08:00
metamergebot
60e80ac7d1
[HiCache] Fix write-back SWA insert backups tripping the write-through pending-ack assert ( #42264 )
...
Co-authored-by: merrymercy <merrymercy@users.noreply.github.com >
Co-authored-by: 842974287 <842974287@users.noreply.github.com >
(cherry picked from commit 2bae12b9b3 )
2026-10-08 15:37:43 +08:00
Sasha Sidorov
4ef799aaa6
[PD] Wait for prefill completion before Mooncake early KV transfer ( #41395 )
...
Co-authored-by: Shangming Cai <csmthu@gmail.com >
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com >
(cherry picked from commit c683190b40 )
2026-10-08 15:36:21 +08:00
Aurick Qiao
1be762368b
[Disagg] Validate state strides before Mooncake transfers ( #41607 )
...
Co-authored-by: Aurick Qiao <6137920+aurickq@users.noreply.github.com >
(cherry picked from commit 1a0250c5b6 )
2026-10-08 15:35:05 +08:00
DarkSharpness
f577b61b52
[DSv4.1] Trim prefill model-entry overhead ( #41704 )
...
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
Co-authored-by: BBuf <1182563586@qq.com >
(cherry picked from commit 2f6e0ec45d )
2026-10-08 15:32:48 +08:00
DarkSharpness
4792b61332
[DSv4.1] Fused c1/c2 compress for eager extend, faster c2 decode ( #41660 )
...
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
Co-authored-by: BBuf <1182563586@qq.com >
(cherry picked from commit 5916999afb )
2026-10-08 15:32:48 +08:00
DarkSharpness
62f1ed099e
[DSv4.1] Faster fp4 index-K gather and combine_topk_swa_indices ( #41658 )
...
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
Co-authored-by: BBuf <1182563586@qq.com >
(cherry picked from commit d75d5b33f7 )
2026-10-08 15:32:47 +08:00
AMD-yanfeiwang
bfd9bb2e03
fix(cuda-graph): remove unnecessary CP batch alignment ( #40543 )
...
(cherry picked from commit 3dc7f4b816 )
2026-10-08 15:32:46 +08:00
metamergebot
6deac260bf
[PD] Make the prebuilt last-token H2D copy non-blocking under overlap ( #42113 )
...
Co-authored-by: merrymercy <merrymercy@users.noreply.github.com >
Co-authored-by: bbus <bbus@users.noreply.github.com >
(cherry picked from commit 7dd93b01a5 )
2026-10-08 15:32:46 +08:00
Yuwei An
842ad7c4d4
[Fix][DSV4.1] SWA page size with bounded replay ( #42128 )
...
(cherry picked from commit ef867fa40d )
2026-10-08 15:32:45 +08:00
Mohammad Miadh Angkad
b605d2d370
[DSA] Stop forcing the per-step CPU seq_lens sync for the k-pool indexer ( #41987 )
...
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai >
(cherry picked from commit 42f2e4fcb7 )
2026-10-08 15:32:44 +08:00
minke.yu
b91137ab98
scheduler: CP-symmetric idle check for health-check admission
...
build-sglang-image / build (push) Successful in 27m52s
Health-check admit/skip used is_fully_idle(), which includes rank-local
hicache drain queues; ranks diverge right after activity, so one rank
dispatched the health-check generate while others piggyback-skipped,
deadlocking CP (hicache drain all_reduce vs CP request broadcast).
Seen on cp2/cp4 + hicache L3 after router health checks.
Recovered from b300 /data/ymk/sglang working copy (uncommitted WIP).
2026-09-24 12:01:57 +08:00
minke.yu
db7d2cb7db
fix: vision check uses get_parallel().pp_group (get_pp_group undefined in this tree)
build-sglang-image / build (push) Successful in 27m58s
2026-09-23 18:05:57 +08:00
Xinyuan Tong
6833498646
model: prune comments and redundant tests in dsv41 vision CP
2026-09-23 14:36:01 +08:00
Xinyuan Tong
b48e2cb1eb
model: TP-wide single-owner image encoding for DeepSeek V4.1
...
ViT and Aligner are replicated per TP rank, encoding each image eight
times with TP8 on both CP1 and CP8. Elect one owner per image and use
ordered full-span broadcasts with a six-phase agreement protocol.
Both CP1 and CP8 benefit while local cache hits preserve collective order.
2026-09-23 14:36:01 +08:00
Xinyuan Tong
bfeb7cd9b2
model: support DeepSeek V4.1 vision with interleave prefill CP
...
The CP runner bypassed the vision merge and used bare text embeddings.
Merge image features before sharding so request-global offsets stay valid.
Canonicalize model IDs separately to preserve scheduler hash IDs.
Keep unsupported combinations guarded and isolate embedding overrides
from multimodal prefills without starving queued FCFS requests.
2026-09-23 14:36:00 +08:00
Yuwei An
ddf5207630
[Fix] Handle chunked paged MQA metadata in DSV4.1 eager forwards ( #40637 )
build-sglang-image / build (push) Successful in 32m15s
2026-09-23 13:35:55 +08:00
minke.yu
b081dd3d23
Merge branch 'main' into dsv41-pd
2026-09-22 14:56:35 +08:00
Kevin Mi
8ac19cc19f
[AMD][Kimi-K3] Fix deferred KDA gate projection and update DCP cookbook ( #39066 )
PR Test (Arm64) / check-changes (push) Successful in 10s
PR Test (NPU) / set-image-config (push) Successful in 1s
PR Test (NPU) / Recommend tests from coverage (push) Skipped
PR Test (NPU) / check-changes (push) Successful in 12s
PR Test (sgl-router) / gate (push) Successful in 8s
PR Test (Xeon) / check-changes (push) Successful in 9s
PR Test (XPU) / check-changes (push) Successful in 16s
pr-test-arm64.yml / pr-gate (push) Successful in 3s
PR Test (Arm64) / pr-gate (push) Successful in 3s
pr-test-npu.yml / pr-gate (push) Successful in 2s
PR Test (NPU) / pr-gate (push) Successful in 2s
PR Test (sgl-router) / tier-1 — lint (push) Failing after 33s
PR Test (sgl-router) / tier-2 — build + test (push) Skipped
PR Test (sgl-router) / tier-3 — docker (placeholder) (push) Skipped
PR Test (sgl-router) / tier-3 — k8s integration (push) Skipped
PR Test (sgl-router) / tier-3 — e2e (push) Skipped
pr-test-xpu.yml / pr-gate (push) Successful in 2s
PR Test (XPU) / pr-gate (push) Successful in 2s
pr-test-xeon.yml / pr-gate (push) Successful in 3s
PR Test (Xeon) / pr-gate (push) Successful in 3s
PR Test (sgl-router) / finish (push) Successful in 1s
Lint / lint (push) Failing after 2m51s
PR Test (NPU) / base-a-test-1-npu-a2 (push) Canceled after 0s
PR Test (NPU) / base-b-test-1-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-b-test-2-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-b-test-4-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-b-test-8-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-b-test-16-npu-a3 (push) Canceled after 0s
PR Test (NPU) / multimodal-gen-test-1-npu-a3 (0) (push) Canceled after 0s
PR Test (NPU) / multimodal-gen-test-1-npu-a3 (1) (push) Canceled after 0s
PR Test (NPU) / multimodal-gen-test-4-npu-a3 (0) (push) Canceled after 0s
PR Test (NPU) / multimodal-gen-test-4-npu-a3 (1) (push) Canceled after 0s
PR Test (NPU) / base-c-test-acc-2-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-c-test-acc-16-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-c-test-perf-2-npu-a3 (push) Canceled after 0s
PR Test (NPU) / base-c-test-perf-16-npu-a3 (push) Canceled after 0s
PR Test (NPU) / Analyze failure report (push) Canceled after 0s
PR Test (NPU) / setup-covstub (push) Canceled after 0s
PR Test (NPU) / pr-test-npu-finish (push) Canceled after 0s
pr-test-npu.yml / run (${{ fromJson(inputs.partitions).arr }}) (push) Canceled after 0s
PR Test (XPU) / finish (push) Canceled after 0s
PR Test (Arm64) / build-test (push) Canceled after 0s
PR Test (XPU) / stage-a-test-1-gpu-xpu (push) Canceled after 0s
PR Test (XPU) / multimodal-gen-test-1-gpu-xpu (push) Canceled after 0s
PR Test (Xeon) / build-test (gnr, gnr, xeon-gnr, stage-a-tp-test-cpu-intel) (push) Canceled after 0s
PR Test (Xeon) / build-test (spr1, 0, 3, spr, xeon-spr, stage-a-test-cpu-intel,stage-b-test-cpu-intel) (push) Canceled after 0s
PR Test (Xeon) / build-test (spr2, 1, 3, spr, xeon-spr, stage-a-test-cpu-intel,stage-b-test-cpu-intel) (push) Canceled after 0s
PR Test (Xeon) / build-test (spr3, 2, 3, spr, xeon-spr, stage-a-test-cpu-intel,stage-b-test-cpu-intel) (push) Canceled after 0s
2026-09-22 06:35:12 +00:00
Mohammad Miadh Angkad and Mohammad Angkad
4c81cd1b09
[KDA] Fix missing beta sigmoid in PTX prefill ( #40685 )
...
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai >
2026-09-21 23:24:38 -07:00
Khoa Pham
bc22e1de9e
[DSpark] Fix draft CUDA graph stream explosion ( #40658 )
2026-09-21 23:05:45 -07:00
Guangda Liu and Guangda Liu
04c0913434
[HiSparse] Add MHA hisparse support for MiniMax M3 ( #31446 )
...
Co-authored-by: Guangda Liu <bingps@users.noreply.github.com >
2026-09-22 13:28:03 +08:00
095e45100b
[AMD] [GLM-5.3-Flash Day 0] Route mHC through AITER on gfx950 ( #38545 )
...
Co-authored-by: Raiden-Makoto <Raiden-Makoto@users.noreply.github.com >
Co-authored-by: Thomas Wang <thomawan@amd.com >
Co-authored-by: Kevin Mi <45493463+kevin-mii@users.noreply.github.com >
Co-authored-by: Kevin Mi <mikevin920@yahoo.com >
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com >
2026-09-21 22:24:29 -07:00
Piotr Mazurek
b01961e295
[LFM2-VL] Add DSpark speculative decoding ( #40651 )
2026-09-21 21:49:51 -07:00
YAMY
9b59fc5db5
[ModelOpt][PP] Keep BF16 shared experts out of the NVFP4 fusion so TP1 pipeline stages can load ( #40628 )
2026-09-21 21:45:58 -07:00
a9f02b0fa4
[NPU] Fix xgrammar apply_vocab_mask device dispatch to use torch.ops.npu ( #36120 )
...
Co-authored-by: Even Zhou <even.y.zhou@outlook.com >
Co-authored-by: sglang-npu-bot <sglangnpu@163.com >
2026-09-21 21:34:15 -07:00
877a293d6d
[Benchmark] Optionally clear HiCache storage between cases ( #40659 )
...
Co-authored-by: metamergebot <324680979+metamergebot@users.noreply.github.com >
Co-authored-by: Pengchao Wang <wpc@fb.com >
2026-09-21 21:31:21 -07:00
e1daf68304
[AMD] [GLM-5.3-Flash Day 0] Honor fused and per-expert names in quark exclude ( #39317 )
...
Co-authored-by: Yikai Zhang <ykzhang12@gmail.com >
Co-authored-by: Thomas Wang <thomawan@amd.com >
Co-authored-by: Kevin Mi <45493463+kevin-mii@users.noreply.github.com >
Co-authored-by: Kevin Mi <mikevin920@yahoo.com >
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com >
2026-09-21 21:26:42 -07:00
Khoa Pham and Qiaolin Yu
018b73c7a0
[PD] Pack draft KV head slices for DCP transfers ( #40500 )
...
Co-authored-by: Qiaolin Yu <liin1211@outlook.com >
2026-09-21 21:11:27 -07:00
b44e248682
[AMD] [GLM-5.3-Flash Day 0] Enable FP8 and Quark MXFP4 MoE on gfx950 ( #38546 )
...
Co-authored-by: Raiden-Makoto <Raiden-Makoto@users.noreply.github.com >
Co-authored-by: Thomas Wang <thomawan@amd.com >
Co-authored-by: andyluo7 <andy.luo@amd.com >
Co-authored-by: Kevin Mi <mikevin920@yahoo.com >
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com >
2026-09-21 21:07:37 -07:00
jthomson04
15ba54bd5d
perf(engine): avoid timed waits for Engine responses ( #39486 )
...
Signed-off-by: jthomson04 <jwillthomson19@gmail.com >
2026-09-21 20:42:03 -07:00
Jan Bernlöhr and Po-Han Huang
56fee88e23
fix(moe): support Llama4 NVFP4 router input weights on SM120 ( #35504 )
...
Co-authored-by: Po-Han Huang (NVIDIA) <53919306+nvpohanh@users.noreply.github.com >
2026-09-21 20:18:11 -07:00
90cf471723
[AMD] [GLM-5.3-Flash Day 0] Support non-2048 top-k widths in the DSA page-table transform ( #39340 )
...
Co-authored-by: Thomas Wang <thomawan@amd.com >
Co-authored-by: Kevin Mi <45493463+kevin-mii@users.noreply.github.com >
Co-authored-by: Kevin Mi <mikevin920@yahoo.com >
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-09-21 19:50:17 -07:00
5f9c6b9eb0
[diffusion] fix: separate a use-scoped layerwise release from release_all ( #40590 )
...
Co-authored-by: Mick Qian <mickqian@users.noreply.github.com >
Co-authored-by: Claude Opus 5 <noreply@anthropic.com >
2026-09-22 10:38:44 +08:00
Mick and Mick Qian
a1b2b976fe
[diffusion] CI: restore public Qwen-Image 2.1 TP2 E2E coverage ( #40507 )
...
Co-authored-by: Mick Qian <mickqian@users.noreply.github.com >
2026-09-22 10:38:00 +08:00