GitLab

[qwen3-480b] MoE(TN) configs for nmz TP=8
[opt] 优化deepep相关代码
[fix] 修复deepseek moe模型的awq量化推理bug和精度问题, 修复awq模型的VLLM_USE_LIGHTOP_MOE_SUM_MUL_ADD设置位置, update_state,优化性能，去除冗余操作
pcie 解决custom cudagraph模式需要拷贝的问题，需要配合dtk进行使用
[feat] Switch default w8a8 gemm impl to blaslt. Support w8a8-fp8 GEMM backend.MoE 路由抓取：新增 router_capture 工具链与 envs 统一配置
[envs] set VLLM_CUSTOM_CACHE=1、VLLM_USE_FUSED_RMS_ROPE=1、VLLM_USE_FUSED_FILL_RMS_CAT=1、VLLM_USE_FLASH_ATTN_FP8=1、VLLM_USE_FLASH_MLA_FP8=1、update VLLM_USE_TOPK_RENORM

vllm：export VLLM_CUSTOM_CACHE=1
dtk：export HIP_KERNEL_EVENT_SYSTENFENCE=1

Signed-off-by: Tianer Zhou <ezhoureal@gmail.com>

Signed-off-by: ilmarkov <imarkov@redhat.com>
Co-authored-by: ilmarkov <imarkov@redhat.com>

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Signed-off-by: Hanzhi Zhou <hanzhi713@gmail.com>