ggml : add mmla kernels for quantized GEMM (ggml-org#4966)

* ggml: aarch64: implement smmla kernel for q8_0_q8_0 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q8_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: aarch64: implement smmla kernel for q4_0_q8_0 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: aarch64: implement smmla kernel for q4_1_q8_1 quantized gemm

armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_1_q8_1 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"

On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.

* ggml: update unit tests for the new vec_dot interface

* llama.cpp: add MATMUL_INT8 capability to system_info

Author: snadampal

common/common.cpp

@@ -1550,6 +1550,7 @@ void dump_non_result_info_yaml(FILE * stream, const gpt_params & params, const l
     fprintf(stream, "cpu_has_blas: %s\n",        ggml_cpu_has_blas()        ? "true" : "false");
     fprintf(stream, "cpu_has_sse3: %s\n",        ggml_cpu_has_sse3()        ? "true" : "false");
     fprintf(stream, "cpu_has_vsx: %s\n",         ggml_cpu_has_vsx()         ? "true" : "false");
+    fprintf(stream, "cpu_has_matmul_int8: %s\n", ggml_cpu_has_matmul_int8() ? "true" : "false");
 
 #ifdef NDEBUG
     fprintf(stream, "debug: false\n");

ggml-quants.c

@@ -245,20 +245,20 @@ void dequantize_row_iq2_xs (const block_iq2_xs  * GGML_RESTRICT x, float * GGML_
 void dequantize_row_iq3_xxs(const block_iq3_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int k);
 
 // Dot product
-void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q5_0_q8_0(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q5_1_q8_1(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-
-void ggml_vec_dot_q2_K_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q3_K_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q4_K_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q5_K_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_q6_K_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_iq2_xxs_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_iq2_xs_q8_K (int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
-void ggml_vec_dot_iq3_xxs_q8_K(int n, float * GGML_RESTRICT s, const void * GGML_RESTRICT vx, const void * GGML_RESTRICT vy);
+void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q5_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q5_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+
+void ggml_vec_dot_q2_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q3_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q4_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q5_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_q6_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_iq2_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_iq2_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
+void ggml_vec_dot_iq3_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc);
 
 //
 // Quantization utilizing an importance matrix (a.k.a. "Activation aWare Quantization")