alpha
Login
Join now
clee.sh
/
qxmx
Star
Starred
Star
Something went wrong. Try again.
Fork
1
Eye filled
Watch
Watching
Watch
A single-model (Bonsai-27B) inference engine for Intel Arc GPUs
Star
Starred
Star
Something went wrong. Try again.
Fork
1
Eye filled
Watch
Watching
Watch
Overview
Overview
Issues
Issues
Pulls
Pulls
Pipelines
Pipelines
qxmx
/
tests
20 files
at
main
Chris Lee
·
coarsen prefix-cache snapshot stride + real device banner in CLI tools
·
2mo ago
bf21675d
attn_forward_batch_test.cpp
M1.7b.7: 3-kernel FA split (DPAS QK + softmax + DPAS PV) -- 439 tok/s
2mo ago
chat_template_test.cpp
M3.2-M3.6: OpenAI endpoint, ChatML template, tool calling, lifecycle
2mo ago
deltanet_naive_test.cpp
M1.7b.12: naive register-recurrence DeltaNet -- 591 tok/s (+18% over WY/UT)
2mo ago
deltanet_prefill_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
deltanet_wyut_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
elementwise_batch_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
fa_fused_bench.cpp
M1.7c: fused FlashAttention via sycl::joint_matrix -- 741 tok/s (+20%)
2mo ago
fa_kernel_test.cpp
M4: FlashDecoding split-KV decode attention (7.4x attn, 2.8x decode)
2mo ago
ffn_batch_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
ffn_timing_test.cpp
FFN GEMM tile tune + naive DeltaNet default -- 613 tok/s (+22%)
2mo ago
gemm_engine_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
gemm_tf32_test.cpp
Reorganize source tree: src/ + tools/ + tests/
2mo ago
gemm_tune.cpp
FFN GEMM tile tune + naive DeltaNet default -- 613 tok/s (+22%)
2mo ago
prefix_cache_engine_test.cpp
coarsen prefix-cache snapshot stride + real device banner in CLI tools
2mo ago
prefix_cache_test.cpp
M2.6.2+2.6.3: engine restore path + inline SSM snapshotting
2mo ago
scheduler_test.cpp
M2.5: multi-slot token-interleaved scheduler (closes M2)
2mo ago
ssm_batch_test.cpp
FFN GEMM tile tune + naive DeltaNet default -- 613 tok/s (+22%)
2mo ago
ssm_timing_test.cpp
FFN GEMM tile tune + naive DeltaNet default -- 613 tok/s (+22%)
2mo ago
tool_call_test.cpp
M3.2-M3.6: OpenAI endpoint, ChatML template, tool calling, lifecycle
2mo ago
tpl_crosscheck.py
M3.2-M3.6: OpenAI endpoint, ChatML template, tool calling, lifecycle
2mo ago