yangzhuxinyzx fc3901beb4 [Core] Share SM70 AWQ and FP8 MoE stages and capture engine policy ...

Signed-off-by: yangzhuxinyzx <153831768+yangzhuxinyzx@users.noreply.github.com>

2026-10-09 10:54:32 +08:00

..

data

[Test] Report prefill stalls within the measured online phase

2026-10-04 01:13:53 +08:00

fixtures

[Core] Share SM70 AWQ and FP8 MoE stages and capture engine policy

2026-10-09 10:54:32 +08:00

__init__.py

…

fp_quant.py

…

reference_mxfp4.py

…

test_auto_gptq.py

…

test_auto_round.py

…

test_awq_qpn_sm70.py

…

test_awq_sm70_fallback.py

…

test_awq_sm70_persistent_cap.py

…

test_awq_sm70_sorted_output_contract.py

…

test_blackwell_moe.py

…

test_block_fp8_qpn8_capabilities.py

[Kernel] Retain TurboMind for unsupported block QPN8 prefill rows

2026-10-03 08:28:15 +08:00

test_compressed_tensors.py

…

test_configs.py

…

test_cpu_offload.py

…

test_cpu_wna16.py

…

test_cutlass_w4a16.py

…

test_experts_int8.py

…

test_fp8_marlin_bmm_dequant.py

[Kernel] Retain TurboMind for unsupported block QPN8 prefill rows

2026-10-03 08:28:15 +08:00

test_fp8.py

…

test_gfx950_moe.py

…

test_gptq_dynamic.py

…

test_gptq_v2.py

…

test_lm_head.py

…

test_mixed_precision.py

…

test_modelopt.py

…

test_nvfp4_mxfp4_gating.py

…

test_online.py

…

test_per_token_kv_cache.py

…

test_quantization_config_args.py

…

test_quark_w4a16.py

…

test_quark.py

…

test_register_quantization_config.py

…

test_sm70_awq_active_grouped_decode.py

…

test_sm70_awq_compact_metadata.py

…

test_sm70_awq_indexed_prefill.py

…

test_sm70_awq_kernel_selection.py

[Test] Make AWQ selector probes independent of the host platform

2026-10-03 02:01:54 +08:00

test_sm70_awq_prefill_exact_dense.py

[Core] Select SM70 AWQ through the linear kernel framework

2026-10-03 01:02:28 +08:00

test_sm70_batched_gemm_layouts.py

…

test_sm70_channel_fp8_layout_guard.py

…

test_sm70_compressed_tensors_fastpaths.py

[Test] Bring stale core and SM70 tests in line with current main

2026-10-02 22:54:32 +08:00

test_sm70_ct_fp8_cache_workspace.py

…

test_sm70_fp8_kernel_selection.py

[Test] Refresh NVFP4 compile identity for current kernel defaults

2026-10-09 08:40:45 +08:00

test_sm70_fp8_prefill_exact_dense.py

[Core] Move SM70 serialized FP8 dispatch into the linear kernel lifecycle

2026-10-03 03:36:20 +08:00

test_sm70_fp8_workspace_aot_reload.py

[Core] Move SM70 serialized FP8 dispatch into the linear kernel lifecycle

2026-10-03 03:36:20 +08:00

test_sm70_fp32_lm_head_admission.py

…

test_sm70_glm53_modelopt_nvfp4.py

…

test_sm70_local_workspace_capacity.py

…

test_sm70_modelopt_mixed_nvfp4.py

[Test] Repair twelve SM70 test cases broken by recent main changes

2026-10-09 08:39:54 +08:00

test_sm70_modelopt_nvfp4_kernel_oracle.py

…

test_sm70_modelopt_nvfp4.py

…

test_sm70_moe_backend_override.py

…

test_sm70_moe_buffers.py

[Core] Share SM70 AWQ and FP8 MoE stages and capture engine policy

2026-10-09 10:54:32 +08:00

test_sm70_moe_load_cache_release.py

[Core] Share SM70 AWQ and FP8 MoE stages and capture engine policy

2026-10-09 10:54:32 +08:00

test_sm70_moe_mainflow.py

[Core] Share SM70 AWQ and FP8 MoE stages and capture engine policy

2026-10-09 10:54:32 +08:00

test_sm70_mxfp4_moe.py

[Test] Keep MXFP4 repack regressions independent of CUDA devices

2026-10-02 22:54:32 +08:00

test_sm70_nvfp4_grouped_decode_dispatch.py

[Kernel] Enable qualified Flash-Next batch defaults and MTP admission

2026-10-03 05:32:36 +08:00

test_sm70_nvfp4_integration.py

[Test] Gate native batch HC graph transitions and matched concurrency runs

2026-09-27 11:00:58 +08:00

test_sm70_nvfp4_kernel_selection.py

[Kernel] Integrate SM70 CT NVFP4 into linear kernel selection

2026-10-02 21:14:46 +08:00

test_sm70_nvfp4_qpn2.py

[Kernel] Enable measured dense NVFP4 small-prefill dispatch

2026-10-04 00:17:07 +08:00

test_sm70_online_qpn8.py

…

test_sm70_platform_gate_per_device.py

…

test_sm70_quantized_model_compat.py

…

test_sm70_turbomind_adapter.py

[Test][SM70] Mock converter availability in CPU preparation tests

2026-10-02 22:54:32 +08:00

test_sm70_turbomind_turing_gates.py

[Feature][SM75] Run ModelOpt NVFP4 and FP8 linears on Turing through the SM70 QPN kernels

2026-10-03 06:39:47 +08:00

test_sm70_warmup.py

[Core] Move SM70 serialized FP8 dispatch into the linear kernel lifecycle

2026-10-03 03:36:20 +08:00

test_sm72_volta_admission.py

[Build] Support SM72 targets and robust device diagnostics

2026-10-03 06:27:04 +08:00

test_torchao.py

…

test_trtllm_nvfp4_hidden_dim_padding.py

…

test_turboquant.py

…

test_turing_qpn_kernel_selection.py

[Bugfix] Preserve ModelOpt FP8 layout and fallback metadata

2026-10-03 06:46:36 +08:00

utils.py

…