ggml-hexagon: Implement true Q8_0 quantization on Hexagon NPU for more accurate mixed-precision matmul operations

docs/backend/hexagon/CMakeUserPresets.json

@@ -22,6 +22,7 @@
             "GGML_LLAMAFILE":   "OFF",
             "GGML_OPENCL":      "ON",
             "GGML_HEXAGON":     "ON",
+            "GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE": "128",
             "LLAMA_CURL":       "OFF"
         }
     },
@@ -36,6 +37,7 @@
             "GGML_LLAMAFILE":   "OFF",
             "GGML_OPENCL":      "ON",
             "GGML_HEXAGON":     "ON",
+            "GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE": "128",
             "LLAMA_CURL":       "OFF"
         }
     },

ggml/CMakeLists.txt

@@ -254,6 +254,7 @@ set   (GGML_OPENCL_TARGET_VERSION "300" CACHE STRING
                                             "gmml: OpenCL API version to target")
 
 option(GGML_HEXAGON                         "ggml: enable Hexagon backend"                    OFF)
+set(GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE 128 CACHE STRING "ggml: quantize group size (32, 64, or 128)")
 
 # toolchain for vulkan-shaders-gen
 set   (GGML_VULKAN_SHADERS_GEN_TOOLCHAIN "" CACHE FILEPATH "ggml: toolchain file for vulkan-shaders-gen")

ggml/src/ggml-hexagon/CMakeLists.txt

@@ -2,6 +2,7 @@ include(${HEXAGON_SDK_ROOT}/build/cmake/hexagon_fun.cmake)
 include(ExternalProject)
 
 option(GGML_HEXAGON_HTP_DEBUG "ggml-hexagon: enable HTP debug output" OFF)
+set(GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE 128 CACHE STRING "ggml-hexagon: quantize group size (32, 64, or 128)")
 
 add_library(htp_iface OBJECT
     ${CMAKE_CURRENT_BINARY_DIR}/htp_iface_stub.c)
@@ -41,7 +42,8 @@ set(HTP_CMAKE_ARGS
     -DCMAKE_INSTALL_LIBDIR=${CMAKE_CURRENT_BINARY_DIR}
     -DHEXAGON_SDK_ROOT=$ENV{HEXAGON_SDK_ROOT}
     -DHEXAGON_TOOLS_ROOT=$ENV{HEXAGON_TOOLS_ROOT}
-    -DHEXAGON_HTP_DEBUG=${GGML_HEXAGON_HTP_DEBUG})
+    -DHEXAGON_HTP_DEBUG=${GGML_HEXAGON_HTP_DEBUG}
+    -DGGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE=${GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE})
 
 ExternalProject_Add(htp-v68
     SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR}/htp BUILD_ALWAYS ON

ggml/src/ggml-hexagon/htp/CMakeLists.txt

@@ -31,7 +31,8 @@ add_library(${HTP_LIB} SHARED
 )
 
 target_compile_definitions(${HTP_LIB} PRIVATE
-    $<IF:$<BOOL:${HEXAGON_HTP_DEBUG}>,HTP_DEBUG=1,NDEBUG=1>)
+    $<IF:$<BOOL:${HEXAGON_HTP_DEBUG}>,HTP_DEBUG=1,NDEBUG=1>
+    FP32_QUANTIZE_GROUP_SIZE=${GGML_HEXAGON_FP32_QUANTIZE_GROUP_SIZE})
 
 build_idl(htp_iface.idl ${HTP_LIB})
 

ggml/src/ggml-hexagon/htp/matmul-ops.c

@@ -92,6 +92,18 @@ static const uint8_t __attribute__((aligned(128))) repl_1x_fp16[128] = {
     0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
 };
 
+// vdelta control to replicate first fp16 value across all elements
+static const uint8_t __attribute__((aligned(128))) repl_2x_fp16[128] = {
+    0x00, 0x00, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x10, 0x10, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x20, 0x20, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x10, 0x10, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x00, 0x00, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x10, 0x10, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x20, 0x20, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+    0x10, 0x10, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02, 0x08, 0x08, 0x02, 0x02, 0x04, 0x04, 0x02, 0x02,
+};
+
 // vdelta control to expand first 32 e8m0 values into 32 uint32 elements
 static const uint8_t __attribute__((aligned(128))) expand_x32_e8m0[128] = {
     0x00, 0x00, 0x00, 0x00, 0x01, 0x04, 0x00, 0x00, 0x02, 0x00, 0x08, 0x08, 0x01, 0x02, 0x00, 0x04, 0x04, 0x00, 0x00,
@@ -1585,6 +1597,118 @@ static void matmul_f16_f32(struct htp_tensor * restrict src0,
 
 // *** dynamic quant
 
+static inline void quantize_block_fp32_q8x1(float * restrict x, uint8_t * restrict y_q, uint8_t * restrict y_d) {
+    assert((unsigned long) x % 128 == 0);
+    assert((unsigned long) y_q % 128 == 0);
+
+    HVX_Vector * vx = (HVX_Vector *) x;
+    HVX_Vector zero   = Q6_V_vsplat_R(0);
+
+    // Use reduce max fp32 to find max(abs(e)) first
+    HVX_Vector vmax0_sf = hvx_vec_reduce_max_fp32(hvx_vec_abs_fp32(vx[0]));
+    HVX_Vector vmax1_sf = hvx_vec_reduce_max_fp32(hvx_vec_abs_fp32(vx[1]));
+    HVX_Vector vmax2_sf = hvx_vec_reduce_max_fp32(hvx_vec_abs_fp32(vx[2]));
+    HVX_Vector vmax3_sf = hvx_vec_reduce_max_fp32(hvx_vec_abs_fp32(vx[3]));
+    // Load and convert into QF32
+    HVX_Vector vx0_qf = Q6_Vqf32_vsub_VsfVsf(vx[0], zero);  // 32 elements
+    HVX_Vector vx1_qf = Q6_Vqf32_vsub_VsfVsf(vx[1], zero);  // 32 elements
+    HVX_Vector vx2_qf = Q6_Vqf32_vsub_VsfVsf(vx[2], zero);  // 32 elements
+    HVX_Vector vx3_qf = Q6_Vqf32_vsub_VsfVsf(vx[3], zero);  // 32 elements
+
+    // Convert to QF32
+    HVX_Vector vmax0_qf = Q6_Vqf32_vsub_VsfVsf(vmax0_sf, zero);
+    HVX_Vector vmax1_qf = Q6_Vqf32_vsub_VsfVsf(vmax1_sf, zero);
+    HVX_Vector vmax2_qf = Q6_Vqf32_vsub_VsfVsf(vmax2_sf, zero);
+    HVX_Vector vmax3_qf = Q6_Vqf32_vsub_VsfVsf(vmax3_sf, zero);
+
+    // Combine and convert to fp16
+    HVX_Vector vmax01_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vmax1_qf, vmax0_qf)));
+    HVX_Vector vmax23_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vmax3_qf, vmax2_qf)));
+
+    // Convert into fp16
+    HVX_Vector vx01_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vx1_qf, vx0_qf)));
+    HVX_Vector vx23_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vx3_qf, vx2_qf)));
+
+    // Replicate first fp16 scale across all lanes
+    HVX_Vector ctrl = *(const HVX_Vector *) repl_2x_fp16;
+    vmax01_hf         = Q6_V_vdelta_VV(vmax01_hf, ctrl);
+    vmax23_hf         = Q6_V_vdelta_VV(vmax23_hf, ctrl);
+
+    HVX_Vector vd01_qf16 = Q6_Vqf16_vmpy_VhfVhf(vmax01_hf, Q6_Vh_vsplat_R(0x2008));  // 1.0 / 127.0
+    HVX_Vector vd23_qf16 = Q6_Vqf16_vmpy_VhfVhf(vmax23_hf, Q6_Vh_vsplat_R(0x2008));  // 1.0 / 127.0
+    HVX_Vector vd01_hf   = Q6_Vhf_equals_Vqf16(vd01_qf16);
+    HVX_Vector vd23_hf   = Q6_Vhf_equals_Vqf16(vd23_qf16);
+
+    hvx_vec_store_u(y_d + 0, 2, vd01_hf);
+    HVX_Vector rotated_vd_hf = Q6_V_vror_VR(vd01_hf, 64);
+    hvx_vec_store_u(y_d + 2, 2, rotated_vd_hf);
+
+    hvx_vec_store_u(y_d + 4, 2, vd23_hf);
+    rotated_vd_hf = Q6_V_vror_VR(vd23_hf, 64);
+    hvx_vec_store_u(y_d + 6, 2, rotated_vd_hf);
+
+    // Divide input by the scale
+    HVX_Vector vd01_inv_hf = hvx_vec_inverse_fp16(vd01_hf);
+    HVX_Vector vd23_inv_hf = hvx_vec_inverse_fp16(vd23_hf);
+    vx01_hf              = Q6_Vhf_equals_Vqf16(Q6_Vqf16_vmpy_VhfVhf(vx01_hf, vd01_inv_hf));
+    vx23_hf              = Q6_Vhf_equals_Vqf16(Q6_Vqf16_vmpy_VhfVhf(vx23_hf, vd23_inv_hf));
+
+    // Convert to int8
+    HVX_Vector vx01_i16 = hvx_vec_i16_from_hf_rnd_sat(vx01_hf);
+    HVX_Vector vx23_i16 = hvx_vec_i16_from_hf_rnd_sat(vx23_hf);
+    HVX_Vector vx_i8    = Q6_Vb_vpack_VhVh_sat(vx23_i16, vx01_i16);
+
+    *(HVX_Vector *) y_q = vx_i8;
+}
+
+static inline void quantize_block_fp32_q8x2(float * restrict x, uint8_t * restrict y_q, uint8_t * restrict y_d) {
+    assert((unsigned long) x % 128 == 0);
+    assert((unsigned long) y_q % 128 == 0);
+
+    HVX_Vector * vx = (HVX_Vector *) x;
+
+    // Load and convert into QF32
+    HVX_Vector zero   = Q6_V_vsplat_R(0);
+    HVX_Vector vx0_qf = Q6_Vqf32_vsub_VsfVsf(vx[0], zero);  // 32 elements
+    HVX_Vector vx1_qf = Q6_Vqf32_vsub_VsfVsf(vx[1], zero);  // 32 elements
+    HVX_Vector vx2_qf = Q6_Vqf32_vsub_VsfVsf(vx[2], zero);  // 32 elements
+    HVX_Vector vx3_qf = Q6_Vqf32_vsub_VsfVsf(vx[3], zero);  // 32 elements
+
+    // Convert into fp16
+    HVX_Vector vx01_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vx1_qf, vx0_qf)));
+    HVX_Vector vx23_hf = Q6_Vh_vdeal_Vh(Q6_Vhf_equals_Wqf32(Q6_W_vcombine_VV(vx3_qf, vx2_qf)));
+
+    // Compute max and scale
+    HVX_Vector vmax01_hf = hvx_vec_reduce_max_fp16(hvx_vec_abs_fp16(vx01_hf));
+    HVX_Vector vmax23_hf = hvx_vec_reduce_max_fp16(hvx_vec_abs_fp16(vx23_hf));
+
+    // Replicate first fp16 scale across all lanes
+    HVX_Vector ctrl = *(const HVX_Vector *) repl_1x_fp16;
+    vmax01_hf         = Q6_V_vdelta_VV(vmax01_hf, ctrl);
+    vmax23_hf         = Q6_V_vdelta_VV(vmax23_hf, ctrl);
+
+    HVX_Vector vd01_qf16 = Q6_Vqf16_vmpy_VhfVhf(vmax01_hf, Q6_Vh_vsplat_R(0x2008));  // 1.0 / 127.0
+    HVX_Vector vd23_qf16 = Q6_Vqf16_vmpy_VhfVhf(vmax23_hf, Q6_Vh_vsplat_R(0x2008));  // 1.0 / 127.0
+    HVX_Vector vd01_hf   = Q6_Vhf_equals_Vqf16(vd01_qf16);
+    HVX_Vector vd23_hf   = Q6_Vhf_equals_Vqf16(vd23_qf16);
+
+    hvx_vec_store_u(y_d + 0, 4, vd01_hf);
+    hvx_vec_store_u(y_d + 4, 4, vd23_hf);
+
+    // Divide input by the scale
+    HVX_Vector vd01_inv_hf = hvx_vec_inverse_fp16(vd01_hf);
+    HVX_Vector vd23_inv_hf = hvx_vec_inverse_fp16(vd23_hf);
+    vx01_hf              = Q6_Vhf_equals_Vqf16(Q6_Vqf16_vmpy_VhfVhf(vx01_hf, vd01_inv_hf));
+    vx23_hf              = Q6_Vhf_equals_Vqf16(Q6_Vqf16_vmpy_VhfVhf(vx23_hf, vd23_inv_hf));
+
+    // Convert to int8
+    HVX_Vector vx01_i16 = hvx_vec_i16_from_hf_rnd_sat(vx01_hf);
+    HVX_Vector vx23_i16 = hvx_vec_i16_from_hf_rnd_sat(vx23_hf);
+    HVX_Vector vx_i8    = Q6_Vb_vpack_VhVh_sat(vx23_i16, vx01_i16);
+
+    *(HVX_Vector *) y_q = vx_i8;
+}
+
 static inline void quantize_block_fp32_q8x4(float * restrict x, uint8_t * restrict y_q, uint8_t * restrict y_d) {
     assert((unsigned long) x % 128 == 0);
     assert((unsigned long) y_q % 128 == 0);
@@ -1646,10 +1770,24 @@ static void quantize_row_fp32_q8x4x2(float * restrict x, uint8_t * restrict y, u
     uint8_t * restrict t_d = (uint8_t *) x;
 
     for (uint32_t i = 0; i < nb; i++) {
+#if FP32_QUANTIZE_GROUP_SIZE == 32
+        quantize_block_fp32_q8x1(x + (i * 2 + 0) * qk / 2, y_q + (i * 2 + 0) * qblk_size / 2,
+                                 t_d + (i * 2 + 0) * dblk_size / 2);
+        quantize_block_fp32_q8x1(x + (i * 2 + 1) * qk / 2, y_q + (i * 2 + 1) * qblk_size / 2,
+                                 t_d + (i * 2 + 1) * dblk_size / 2);
+#elif FP32_QUANTIZE_GROUP_SIZE == 64
+        quantize_block_fp32_q8x2(x + (i * 2 + 0) * qk / 2, y_q + (i * 2 + 0) * qblk_size / 2,
+                                 t_d + (i * 2 + 0) * dblk_size / 2);
+        quantize_block_fp32_q8x2(x + (i * 2 + 1) * qk / 2, y_q + (i * 2 + 1) * qblk_size / 2,
+                                 t_d + (i * 2 + 1) * dblk_size / 2);
+#elif FP32_QUANTIZE_GROUP_SIZE == 128
         quantize_block_fp32_q8x4(x + (i * 2 + 0) * qk / 2, y_q + (i * 2 + 0) * qblk_size / 2,
                                  t_d + (i * 2 + 0) * dblk_size / 2);
         quantize_block_fp32_q8x4(x + (i * 2 + 1) * qk / 2, y_q + (i * 2 + 1) * qblk_size / 2,
                                  t_d + (i * 2 + 1) * dblk_size / 2);
+#else
+#error "FP32_QUANTIZE_GROUP_SIZE must be 32, 64, or 128"
+#endif
     }
 
     // now copy the scales into final location
@@ -1662,6 +1800,7 @@ static void quantize_fp32_q8x4x2(const struct htp_tensor * src,
                                  uint32_t          nth,
                                  uint32_t          ith,
                                  uint32_t          nrows_per_thread) {
+
     uint64_t t1 = HAP_perf_get_qtimer_count();
 
     const uint32_t ne0 = src->ne[0];