Jiejing 8e488bbd55 serve: prefill/decode split prototype over two hosts (pd_server.py) ... A prompt-only engine tray (PE) and a decode-only engine tray (CE) behind the OpenAI-compatible frontend from serve/openai_server.py: - serve/pd_server.py: --role pe / --role ce. The CE frontend renders the prompt, picks a CE rank (affinity or least-loaded routing) and POSTs /prefill to the PE. The PE prefills with the host prefix cache on, exports the prompt's KV blocks and the full SWA state in the pe_do_save_kvcache hook, streams the first sampled token back in the HTTP response and pushes the KV bytes to the CE rank over TCP. The CE rank admits the sequence into decode via Scheduler.add_prompted_sequence and reclaims PE sequences through Scheduler.free_finished_seqs. - deepinfer/hicache/pd_transfer.py: export/admit of KV pages and state through the hicache page-copy kernels. The CE keeps the pinned receive buffers alive until the admit copy kernels finish (the kernels read them by raw pointer; the caching host allocator otherwise reused the memory while the copy was queued -> cudaErrorLaunchFailure). - serve/pd_run.sh start/stop/logs over ssh; serve/check_pd.py compares outputs against a single-engine server token by token; serve/nccl_xtray_test.py reports the cross-host NCCL transport. Measured on 4 GB300 (PE) + 4 GB300 (CE) against one 4-GPU tray with the AgentX replay, outputs identical: ITL p90 stays within 8.6-15 ms from conc 8 to 256 (single tray 12-98 ms), P90 interactivity 66-117 tok/s/user where the single tray falls to 10; per total chip PD is behind up to conc 64 and ahead at 256 (doc/benchmark_gb300.md). Depends on the host prefix cache (#144), the OpenAI server (#147) and the scheduler bindings (add_prompted_sequence / free_finished_seqs), which are included as the first three commits. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFPG29daR1FhchJBTjiEKr		2026-10-21 22:26:29 +00:00
..
_C	engine: expose Scheduler.add_prompted_sequence and free_finished_seqs	2026-10-21 22:26:29 +00:00
distributed	initial commit	2026-10-20 12:09:42 +08:00
engine	initial commit	2026-10-20 12:09:42 +08:00
entrypoints	initial commit	2026-10-20 12:09:42 +08:00
hicache	serve: prefill/decode split prototype over two hosts (pd_server.py)	2026-10-21 22:26:29 +00:00
hugging_face	initial commit	2026-10-20 12:09:42 +08:00
model_executor	initial commit	2026-10-20 12:09:42 +08:00
npu_ops	initial commit	2026-10-20 12:09:42 +08:00
structure_output	initial commit	2026-10-20 12:09:42 +08:00
ve	initial commit	2026-10-20 12:09:42 +08:00
worker	initial commit	2026-10-20 12:09:42 +08:00
__init__.py	initial commit	2026-10-20 12:09:42 +08:00
attrdict.py	initial commit	2026-10-20 12:09:42 +08:00
config.py	initial commit	2026-10-20 12:09:42 +08:00
environ.py	initial commit	2026-10-20 12:09:42 +08:00
load_class.py	initial commit	2026-10-20 12:09:42 +08:00
logger.py	initial commit	2026-10-20 12:09:42 +08:00
model_config_loader.py	initial commit	2026-10-20 12:09:42 +08:00
multimodal_tokens.py	initial commit	2026-10-20 12:09:42 +08:00
platform.py	initial commit	2026-10-20 12:09:42 +08:00
py.typed	initial commit	2026-10-20 12:09:42 +08:00
sampling_params.py	initial commit	2026-10-20 12:09:42 +08:00
sequence.py	initial commit	2026-10-20 12:09:42 +08:00
shared_memory.py	initial commit	2026-10-20 12:09:42 +08:00
testing.py	initial commit	2026-10-20 12:09:42 +08:00
timer.py	initial commit	2026-10-20 12:09:42 +08:00
utils.py	initial commit	2026-10-20 12:09:42 +08:00