[Kernel/Quant] Remove AQLM (#22943)

Signed-off-by: mgoin <[email protected]>
Co-authored-by: Wentao Ye <[email protected]>

Author: mgoin

.buildkite/scripts/hardware_ci/run-amd-test.sh

@@ -121,7 +121,6 @@ fi
 if [[ $commands == *" kernels/quantization"* ]]; then
   commands="${commands} \
   --ignore=kernels/quantization/test_int8_quant.py \
-  --ignore=kernels/quantization/test_aqlm.py \
   --ignore=kernels/quantization/test_machete_mm.py \
   --ignore=kernels/quantization/test_block_fp8.py \
   --ignore=kernels/quantization/test_block_int8.py \

CMakeLists.txt

@@ -286,7 +286,6 @@ if(VLLM_GPU_LANG STREQUAL "CUDA")
   FetchContent_MakeAvailable(cutlass)
 
   list(APPEND VLLM_EXT_SRC
-    "csrc/quantization/aqlm/gemm_kernels.cu"
     "csrc/quantization/awq/gemm_kernels.cu"
     "csrc/permute_cols.cu"
     "csrc/quantization/cutlass_w8a8/scaled_mm_entry.cu"

benchmarks/kernels/benchmark_aqlm.py

@@ -154,15 +154,6 @@ void cutlass_mla_decode(torch::Tensor const& out, torch::Tensor const& q_nope,
 torch::Tensor get_cuda_view_from_cpu_tensor(torch::Tensor& cpu_tensor);
 
 #ifndef USE_ROCM
-torch::Tensor aqlm_gemm(const torch::Tensor& input, const torch::Tensor& codes,
-                        const torch::Tensor& codebooks,
-                        const torch::Tensor& scales,
-                        const std::vector<int64_t>& codebook_partition_sizes,
-                        const std::optional<torch::Tensor>& bias);
-
-torch::Tensor aqlm_dequant(
-    const torch::Tensor& codes, const torch::Tensor& codebooks,
-    const std::vector<int64_t>& codebook_partition_sizes);
 
 torch::Tensor awq_gemm(torch::Tensor _in_feats, torch::Tensor _kernel,
                        torch::Tensor _scaling_factors, torch::Tensor _zeros,