Updated the int4 per-group dequant. Meet function bug.

example/65_gemm_multiply_multiply/CMakeLists.txt

 add_example_executable(example_gemm_multiply_multiply_xdl_fp8 gemm_multiply_multiply_xdl_fp8.cpp)
 add_example_executable(example_gemm_multiply_multiply_xdl_fp8_ab_scale gemm_multiply_multiply_xdl_fp8_ab_scale.cpp)
 add_example_executable(example_gemm_fp16int8_b_scale gemm_fp16int8_b_scale.cpp)
+add_example_executable(example_gemm_fp16int4_b_scale gemm_fp16int4_b_scale.cpp)
 add_example_executable(example_gemm_add_add_xdl_fp16 gemm_add_add_xdl_fp16.cpp)

example/65_gemm_multiply_multiply/gemm_fp16int8_b_scale.cpp

@@ -61,7 +61,7 @@ using CDEElementOp = PassThrough;
 static constexpr auto GemmSpec = ck::tensor_operation::device::GemmSpecialization::Default;
 
 // static constexpr ck::index_t Scale_Block_M = 128;
-static constexpr ck::index_t Scale_Block_N = 128;
+static constexpr ck::index_t Scale_Block_N = 1;
 static constexpr ck::index_t Scale_Block_K = 128;
 
 using DeviceOpInstance = ck::tensor_operation::device::DeviceGemmMultiD_BScale_Xdl_CShuffle_V3
@@ -217,7 +217,8 @@ int main(int argc, char* argv[])
         a0_m_k.GenerateTensorValue(GeneratorTensor_1<A0DataType>{});
         quant_b0_k_n.GenerateTensorValue(GeneratorTensor_1<QuantDataType>{});
         // a1_m_k.GenerateTensorValue(GeneratorTensor_1<A1DataType>{});
-        b1_k_n.GenerateTensorValue(GeneratorTensor_1<B1DataType>{});
+        // b1_k_n.GenerateTensorValue(GeneratorTensor_1<B1DataType>{});
+        b1_k_n.GenerateTensorValue(GeneratorTensor_3<B1DataType>{0, 1.0});
         break;
     case 3:
         a0_m_k.GenerateTensorValue(GeneratorTensor_2<A0DataType>{-2, 2});

include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_b_scale_selector.hpp

@@ -39,7 +39,6 @@ template <BlockGemmPipelineVersion BlkGemmPipelineVer,
           index_t KPack>
 constexpr auto BlockGemmBScalePipeline_Selector()
 {
-    printf("I'm Here\n");
     return BlockwiseGemmXdlops_pipeline_v3_b_scale<BlkGemmPipeSche,
                                                    BlockSize,
                                                    ADataType,

include/ck/tensor_operation/gpu/block/blockwise_gemm_pipeline_xdlops_v3_b_scale.hpp

@@ -440,7 +440,7 @@ struct BlockwiseGemmXdlops_pipeline_v3_b_scale<BlockGemmPipelineScheduler::Intra
                             c_thread_buf(Number<c_offset>{}) +=
                                 c_thread_buf_per_scale[Number<t>{}] *
                                 // type_convert<AccDataType>(a_scale_thread_buf[I0]) *
-                                type_convert<AccDataType>(b_scale_thread_buf[I0]);
+                                type_convert<AccDataType>(b_scale_thread_buf[n0]);
                         });
                     });
                 });

include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_xdl_cshuffle_v3_b_scale.hpp

@@ -360,7 +360,8 @@ struct DeviceGemmMultiD_BScale_Xdl_CShuffle_V3
             return false;
         }
 
-        if(ScaleBlockN % NPerBlock != 0 || ScaleBlockK != KPerBlock)
+        // if(ScaleBlockN % NPerBlock != 0 || ScaleBlockK != KPerBlock)
+        if(ScaleBlockK != KPerBlock)
         {
             printf("Return 1\n");
             return false;

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_cshuffle_v3_multi_d_b_scale.hpp

@@ -1359,15 +1359,15 @@ struct GridwiseGemmMultiD_BScale_xdl_cshuffle_v3
             (a_grid_desc_ak0_m_ak1.GetLength(I0) * a_grid_desc_ak0_m_ak1.GetLength(I2)) /
             KPerBlock);
 
-        const index_t ScaleSliceSizeM = 1;
-        const index_t ScaleSliceSizeN = 1;
+        //const index_t ScaleSliceSizeM = 1;
+        const index_t ScaleSliceSizeN = NXdlPerWave;
         const index_t ScaleSliceSizeK = 1;
 
         // constexpr auto a_scale_thread_desc = make_naive_tensor_descriptor_packed(
         //     make_tuple(Number<ScaleSliceSizeM>{}, Number<ScaleSliceSizeK>{}));
 
         constexpr auto b_scale_thread_desc = make_naive_tensor_descriptor_packed(
-            make_tuple(Number<ScaleSliceSizeM>{}, Number<ScaleSliceSizeK>{}));
+            make_tuple(Number<ScaleSliceSizeN>{}, Number<ScaleSliceSizeK>{}));
 
         // auto a_scale_thread_copy =
         //     ThreadwiseTensorSliceTransfer_v2<AScaleType,

include/ck/utility/data_type.hpp

@@ -12,6 +12,7 @@ using half_t  = _Float16;
 using int4_t  = _BitInt(4);
 using f8_t    = _BitInt(8);
 using bf8_t   = unsigned _BitInt(8);
+using pk_i4_t = unsigned char;
 
 inline constexpr auto next_pow2(uint32_t x)
 {