ViperEkura/AstrAI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
835 Commits 1 Branch 14 Tags
587b0ee046d1f0ea0ef7f09fc2c2f28dcb6fbe57
Commit Graph
6 Commits
Author SHA1 Message Date
ViperEkura 92e3cdf044 fix: allocate inference workspace buffers outside inference mode 
- workspace buffers are transport storage mutated in-place every step by the scheduler loop thread, which holds no ambient inference-mode context because torch.inference_mode is thread-local
- callers building the engine inside torch.inference_mode (scripts/tools/generate.py) produced inference tensors that reject off-thread in-place updates, crashing the first decode fill and aborting tasks after a single token
- force inference mode off around all workspace allocation so every buffer is a plain tensor regardless of caller context
 2026-09-02 20:40:49 +08:00
ViperEkura 1cf7d6c76b perf: fill steady-state decode input ids via d2d copy 
- add InferenceWorkspace.fill_input_ids_from_device copying device tokens straight into the fixed-address input_ids buffer
- cache each decode step's sampled tokens on-device in DecodeSteadyState.last_tokens; when the task signature is unchanged the next step reuses them, replacing the tolist -> python list -> elementwise host fill -> pageable h2d round-trip
- _sample_logits returns (host payload, device tokens); prefill discards the device tensor
- signature change (task join/leave/first decode) still takes the host path; both dispatch paths covered by tests

Benchmark: NVIDIA L20, BF16, 1B model + 0.11B test model (4 layers, hidden 512), contiguous KV cache, CUDA Graph, greedy, prompt 512, generation 256, engine decode via scripts/tools/benchmark.py (alternating A/B, 2-4 paired runs)
- 0.11B batch 32: 21429 -> 24415 tok/s mean (1.14x, +13.9%), 4/4 paired runs faster
- 1B batch 32: 4242 -> 4388 tok/s (1.034x, +3.4%), 7.54 -> 7.29 ms/step
- batch 1: no measurable change (<0.5%)
 2026-08-31 14:24:51 +08:00
ViperEkura f252af495c refactor: remove dead code and deduplicate scheduler setup 2026-08-19 14:58:37 +08:00
ViperEkura 1bcd8f53ab perf: precompute ragged Q tile scheduling 2026-08-16 23:32:46 +08:00
ViperEkura 3fb4b8ab13 perf: use int32 paged KV indices 
- store page-table, request-row, and cache-location indices as int32
- preserve CUDA graph replay with bit-exact logits and KV cache coverage
- improve B=1 decode latency by 1-6% across 1K-32K contexts on L20
 2026-08-15 13:17:06 +08:00
ViperEkura 0c1b7664c1 refactor: split infer core into subpackages by concern 
- Eliminate core/ directory into cache/, runtime/, network/ subpackages plus flat modules
- Split cache.py (647 lines) into cache/{buffer,strategy,pool}.py by layer
- Add explicit ContiguousStrategy, make AllocationStrategy a real ABC
- Move TaskCacheState to cache/strategy.py, drop string forward references
- Rename api/ to network/, server.py to app.py
- Move sample.py into runtime/ alongside executor and graph
- Simplify TaskCacheManager.__init__ to single pool param
- Expose pool.strategy and pool.req_pool as public properties
- Fix KVCache import in attention_backend.py (TYPE_CHECKING guard)
- Fix steady-state decode reading uninitialized position_ids on first step
 2026-08-08 23:43:05 +08:00