[Kernel] Expand MoE weight loading + Add Fused Marlin MoE Kernel (vll…

…m-project#7527) Co-authored-by: ElizaWszola <eliza@neuralmagic.com>
AnswerDotAI · Aug 21, 2024 · 8678a69 · 8678a69
1 parent 5844017
commit 8678a69
Show file tree

Hide file tree

Showing 15 changed files with 2,374 additions and 84 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -286,7 +286,8 @@ define_gpu_extension_target(
 
 set(VLLM_MOE_EXT_SRC
   "csrc/moe/torch_bindings.cpp"
-  "csrc/moe/topk_softmax_kernels.cu")
+  "csrc/moe/topk_softmax_kernels.cu"
+  "csrc/moe/marlin_moe_ops.cu")
 
 define_gpu_extension_target(
   _moe_C