[Kernel] Marlin Expansion: Support AutoGPTQ Models with Marlin (vllm-…

…project#3922) Co-authored-by: alexm <[email protected]> Co-authored-by: mgoin <[email protected]>
alexm-redhat · Apr 29, 2024 · 73c8d67 · 73c8d67
1 parent df29793
commit 73c8d67
Show file tree

Hide file tree

Showing 14 changed files with 2,626 additions and 104 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -177,6 +177,8 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
     "csrc/quantization/aqlm/gemm_kernels.cu"
     "csrc/quantization/awq/gemm_kernels.cu"
     "csrc/quantization/marlin/marlin_cuda_kernel.cu"
+    "csrc/quantization/gptq_marlin/gptq_marlin.cu"
+    "csrc/quantization/gptq_marlin/gptq_marlin_repack.cu"
     "csrc/custom_all_reduce.cu")
 endif()
 

diff --git a/csrc/ops.h b/csrc/ops.h
@@ -124,6 +124,24 @@ torch::Tensor marlin_gemm(
     int64_t size_m, 
     int64_t size_n, 
     int64_t size_k);
+
+torch::Tensor gptq_marlin_gemm(
+  torch::Tensor &a,
+  torch::Tensor &b_q_weight,
+  torch::Tensor &b_scales,
+  torch::Tensor &g_idx,
+  torch::Tensor &perm,
+  torch::Tensor &workspace,
+  int64_t size_m,
+  int64_t size_n,
+  int64_t size_k,
+  bool is_k_full);
+
+torch::Tensor gptq_marlin_repack(
+  torch::Tensor &b_q_weight,
+  torch::Tensor &perm,
+  int64_t size_k,
+  int64_t size_n);
 #endif
 
 void squeezellm_gemm(

diff --git a/csrc/pybind.cpp b/csrc/pybind.cpp
@@ -67,6 +67,8 @@ PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {
   ops.def("aqlm_dequant", &aqlm_dequant, "Decompression method for AQLM");
   ops.def("awq_gemm", &awq_gemm, "Quantized GEMM for AWQ");
   ops.def("marlin_gemm", &marlin_gemm, "Marlin Optimized Quantized GEMM for GPTQ");
+  ops.def("gptq_marlin_gemm", &gptq_marlin_gemm, "gptq_marlin Optimized Quantized GEMM for GPTQ");
+  ops.def("gptq_marlin_repack", &gptq_marlin_repack, "gptq_marlin repack from GPTQ");
   ops.def("awq_dequantize", &awq_dequantize, "Dequantization for AWQ");
 #endif