@@ -51,6 +51,7 @@ cc_library_static {
5151 " src/core/AccessWindowAutoPadding.cpp" ,
5252 " src/core/AccessWindowStatic.cpp" ,
5353 " src/core/AccessWindowTranspose.cpp" ,
54+ " src/core/CL/CLCompileContext.cpp" ,
5455 " src/core/CL/CLCoreRuntimeContext.cpp" ,
5556 " src/core/CL/CLHelpers.cpp" ,
5657 " src/core/CL/CLKernelLibrary.cpp" ,
@@ -67,8 +68,12 @@ cc_library_static {
6768 " src/core/CL/OpenCL.cpp" ,
6869 " src/core/CL/gemm/CLGEMMHelpers.cpp" ,
6970 " src/core/CL/gemm/native/CLGEMMNativeKernelConfigurationBifrost.cpp" ,
71+ " src/core/CL/gemm/native/CLGEMMNativeKernelConfigurationMidgard.cpp" ,
72+ " src/core/CL/gemm/native/CLGEMMNativeKernelConfigurationValhall.cpp" ,
7073 " src/core/CL/gemm/reshaped/CLGEMMReshapedKernelConfigurationBifrost.cpp" ,
74+ " src/core/CL/gemm/reshaped/CLGEMMReshapedKernelConfigurationValhall.cpp" ,
7175 " src/core/CL/gemm/reshaped_only_rhs/CLGEMMReshapedOnlyRHSKernelConfigurationBifrost.cpp" ,
76+ " src/core/CL/gemm/reshaped_only_rhs/CLGEMMReshapedOnlyRHSKernelConfigurationValhall.cpp" ,
7277 " src/core/CL/kernels/CLAbsoluteDifferenceKernel.cpp" ,
7378 " src/core/CL/kernels/CLAccumulateKernel.cpp" ,
7479 " src/core/CL/kernels/CLActivationLayerKernel.cpp" ,
@@ -117,17 +122,16 @@ cc_library_static {
117122 " src/core/CL/kernels/CLFlattenLayerKernel.cpp" ,
118123 " src/core/CL/kernels/CLFloorKernel.cpp" ,
119124 " src/core/CL/kernels/CLFuseBatchNormalizationKernel.cpp" ,
120- " src/core/CL/kernels/CLGEMMLowpMatrixMultiplyKernel.cpp" ,
121125 " src/core/CL/kernels/CLGEMMLowpMatrixMultiplyNativeKernel.cpp" ,
122126 " src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedKernel.cpp" ,
123127 " src/core/CL/kernels/CLGEMMLowpMatrixMultiplyReshapedOnlyRHSKernel.cpp" ,
124128 " src/core/CL/kernels/CLGEMMLowpOffsetContributionKernel.cpp" ,
125129 " src/core/CL/kernels/CLGEMMLowpOffsetContributionOutputStageKernel.cpp" ,
130+ " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleByFloatKernel.cpp" ,
131+ " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ScaleKernel.cpp" ,
126132 " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp" ,
127133 " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp" ,
128134 " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp" ,
129- " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleByFloatKernel.cpp" ,
130- " src/core/CL/kernels/CLGEMMLowpQuantizeDownInt32ToUint8ScaleKernel.cpp" ,
131135 " src/core/CL/kernels/CLGEMMLowpReductionKernel.cpp" ,
132136 " src/core/CL/kernels/CLGEMMMatrixAccumulateBiasesKernel.cpp" ,
133137 " src/core/CL/kernels/CLGEMMMatrixMultiplyKernel.cpp" ,
@@ -169,6 +173,7 @@ cc_library_static {
169173 " src/core/CL/kernels/CLPixelWiseMultiplicationKernel.cpp" ,
170174 " src/core/CL/kernels/CLPoolingLayerKernel.cpp" ,
171175 " src/core/CL/kernels/CLPriorBoxLayerKernel.cpp" ,
176+ " src/core/CL/kernels/CLQLSTMLayerNormalizationKernel.cpp" ,
172177 " src/core/CL/kernels/CLQuantizationLayerKernel.cpp" ,
173178 " src/core/CL/kernels/CLROIAlignLayerKernel.cpp" ,
174179 " src/core/CL/kernels/CLROIPoolingLayerKernel.cpp" ,
@@ -209,7 +214,6 @@ cc_library_static {
209214 " src/core/CPP/kernels/CPPBoxWithNonMaximaSuppressionLimitKernel.cpp" ,
210215 " src/core/CPP/kernels/CPPCornerCandidatesKernel.cpp" ,
211216 " src/core/CPP/kernels/CPPDetectionWindowNonMaximaSuppressionKernel.cpp" ,
212- " src/core/CPP/kernels/CPPFlipWeightsKernel.cpp" ,
213217 " src/core/CPP/kernels/CPPNonMaximumSuppressionKernel.cpp" ,
214218 " src/core/CPP/kernels/CPPPermuteKernel.cpp" ,
215219 " src/core/CPP/kernels/CPPSortEuclideanDistanceKernel.cpp" ,
@@ -278,10 +282,10 @@ cc_library_static {
278282 " src/core/NEON/kernels/NEGEMMLowpMatrixMultiplyKernel.cpp" ,
279283 " src/core/NEON/kernels/NEGEMMLowpOffsetContributionKernel.cpp" ,
280284 " src/core/NEON/kernels/NEGEMMLowpOffsetContributionOutputStageKernel.cpp" ,
285+ " src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ScaleKernel.cpp" ,
281286 " src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt16ScaleByFixedPointKernel.cpp" ,
282287 " src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToInt8ScaleByFixedPointKernel.cpp" ,
283288 " src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleByFixedPointKernel.cpp" ,
284- " src/core/NEON/kernels/NEGEMMLowpQuantizeDownInt32ToUint8ScaleKernel.cpp" ,
285289 " src/core/NEON/kernels/NEGEMMLowpReductionKernel.cpp" ,
286290 " src/core/NEON/kernels/NEGEMMMatrixAccumulateBiasesKernel.cpp" ,
287291 " src/core/NEON/kernels/NEGEMMMatrixAdditionKernel.cpp" ,
@@ -319,6 +323,7 @@ cc_library_static {
319323 " src/core/NEON/kernels/NEPixelWiseMultiplicationKernel.cpp" ,
320324 " src/core/NEON/kernels/NEPoolingLayerKernel.cpp" ,
321325 " src/core/NEON/kernels/NEPriorBoxLayerKernel.cpp" ,
326+ " src/core/NEON/kernels/NEQLSTMLayerNormalizationKernel.cpp" ,
322327 " src/core/NEON/kernels/NEQuantizationLayerKernel.cpp" ,
323328 " src/core/NEON/kernels/NEROIAlignLayerKernel.cpp" ,
324329 " src/core/NEON/kernels/NEROIPoolingLayerKernel.cpp" ,
@@ -349,6 +354,7 @@ cc_library_static {
349354 " src/core/NEON/kernels/NEWidthConcatenateLayerKernel.cpp" ,
350355 " src/core/NEON/kernels/NEWinogradConvolutionLayerKernel.cpp" ,
351356 " src/core/NEON/kernels/NEYOLOLayerKernel.cpp" ,
357+ " src/core/NEON/kernels/arm_gemm/gemm_bf16.cpp" ,
352358 " src/core/NEON/kernels/arm_gemm/gemm_fp16.cpp" ,
353359 " src/core/NEON/kernels/arm_gemm/gemm_fp32.cpp" ,
354360 " src/core/NEON/kernels/arm_gemm/gemm_int16.cpp" ,
@@ -381,26 +387,32 @@ cc_library_static {
381387 " src/core/NEON/kernels/convolution/winograd/padding.cpp" ,
382388 " src/core/NEON/kernels/convolution/winograd/winograd.cpp" ,
383389 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/input_1x8_fp32_fp32_integers.cpp" ,
390+ " src/core/NEON/kernels/convolution/winograd/winograd_transforms/input_4x4_fp16_fp16_integers.cpp" ,
384391 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/input_4x4_fp32_fp32_integers.cpp" ,
392+ " src/core/NEON/kernels/convolution/winograd/winograd_transforms/input_6x6_fp16_fp16_integers.cpp" ,
385393 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/input_6x6_fp32_fp32_integers.cpp" ,
386394 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_2_7_fp32_fp32_integers.cpp" ,
387395 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_2x2_3x3_fp32_fp32_integers.cpp" ,
388396 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_2x2_5x5_fp32_fp32_integers.cpp" ,
389397 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_4_5_fp32_fp32_integers.cpp" ,
398+ " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_4x4_3x3_fp16_fp16_integers.cpp" ,
390399 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_4x4_3x3_fp32_fp32_integers.cpp" ,
391400 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/output_6_3_fp32_fp32_integers.cpp" ,
392401 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_2_7_fp32_fp32_integers.cpp" ,
393402 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_2x2_3x3_fp32_fp32_integers.cpp" ,
394403 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_2x2_5x5_fp32_fp32_integers.cpp" ,
395404 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_4_5_fp32_fp32_integers.cpp" ,
405+ " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_4x4_3x3_fp16_fp16_integers.cpp" ,
396406 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_4x4_3x3_fp32_fp32_integers.cpp" ,
397407 " src/core/NEON/kernels/convolution/winograd/winograd_transforms/weights_6_3_fp32_fp32_integers.cpp" ,
398408 " src/core/PyramidInfo.cpp" ,
399409 " src/core/Rounding.cpp" ,
410+ " src/core/Size2D.cpp" ,
400411 " src/core/SubTensorInfo.cpp" ,
401412 " src/core/TensorInfo.cpp" ,
402413 " src/core/Utils.cpp" ,
403414 " src/core/Validate.cpp" ,
415+ " src/core/Version.cpp" ,
404416 " src/core/utils/helpers/fft.cpp" ,
405417 " src/core/utils/helpers/tensor_transform.cpp" ,
406418 " src/core/utils/io/FileHandler.cpp" ,
@@ -431,6 +443,7 @@ cc_library_static {
431443 " src/runtime/CL/CLTensorAllocator.cpp" ,
432444 " src/runtime/CL/CLTuner.cpp" ,
433445 " src/runtime/CL/ICLSimpleFunction.cpp" ,
446+ " src/runtime/CL/Utils.cpp" ,
434447 " src/runtime/CL/functions/CLAbsoluteDifference.cpp" ,
435448 " src/runtime/CL/functions/CLAccumulate.cpp" ,
436449 " src/runtime/CL/functions/CLActivationLayer.cpp" ,
@@ -522,6 +535,7 @@ cc_library_static {
522535 " src/runtime/CL/functions/CLPixelWiseMultiplication.cpp" ,
523536 " src/runtime/CL/functions/CLPoolingLayer.cpp" ,
524537 " src/runtime/CL/functions/CLPriorBoxLayer.cpp" ,
538+ " src/runtime/CL/functions/CLQLSTMLayer.cpp" ,
525539 " src/runtime/CL/functions/CLQuantizationLayer.cpp" ,
526540 " src/runtime/CL/functions/CLRNNLayer.cpp" ,
527541 " src/runtime/CL/functions/CLROIAlignLayer.cpp" ,
@@ -557,6 +571,9 @@ cc_library_static {
557571 " src/runtime/CL/functions/CLWinogradConvolutionLayer.cpp" ,
558572 " src/runtime/CL/functions/CLWinogradInputTransform.cpp" ,
559573 " src/runtime/CL/functions/CLYOLOLayer.cpp" ,
574+ " src/runtime/CL/gemm/CLGEMMKernelSelectionBifrost.cpp" ,
575+ " src/runtime/CL/gemm/CLGEMMKernelSelectionMidgard.cpp" ,
576+ " src/runtime/CL/gemm/CLGEMMKernelSelectionValhall.cpp" ,
560577 " src/runtime/CL/tuners/BifrostTuner.cpp" ,
561578 " src/runtime/CL/tuners/CLLWSList.cpp" ,
562579 " src/runtime/CL/tuners/MidgardTuner.cpp" ,
@@ -683,6 +700,7 @@ cc_library_static {
683700 " src/runtime/NEON/functions/NEPixelWiseMultiplication.cpp" ,
684701 " src/runtime/NEON/functions/NEPoolingLayer.cpp" ,
685702 " src/runtime/NEON/functions/NEPriorBoxLayer.cpp" ,
703+ " src/runtime/NEON/functions/NEQLSTMLayer.cpp" ,
686704 " src/runtime/NEON/functions/NEQuantizationLayer.cpp" ,
687705 " src/runtime/NEON/functions/NERNNLayer.cpp" ,
688706 " src/runtime/NEON/functions/NEROIAlignLayer.cpp" ,
@@ -764,6 +782,8 @@ cc_library_static {
764782 " src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_s8s32_dot_16x4/generic.cpp" ,
765783 " src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_u8u32_dot_16x4/a55.cpp" ,
766784 " src/core/NEON/kernels/arm_gemm/kernels/a64_hybrid_u8u32_dot_16x4/generic.cpp" ,
785+ " src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_bf16fp32_dot_12x8/generic.cpp" ,
786+ " src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_bf16fp32_mmla_12x8/generic.cpp" ,
767787 " src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_s8s32_mmla_12x8/generic.cpp" ,
768788 " src/core/NEON/kernels/arm_gemm/kernels/a64_interleaved_u8u32_mmla_12x8/generic.cpp" ,
769789 " src/core/NEON/kernels/arm_gemm/kernels/a64_native_fp32_mla_16x4/generic.cpp" ,
@@ -783,20 +803,28 @@ cc_library_static {
783803 " src/core/NEON/kernels/arm_gemm/kernels/a64_smallK_hybrid_u8u32_dot_4x6/generic.cpp" ,
784804 " src/core/NEON/kernels/arm_gemm/kernels/a64_smallK_hybrid_u8u32_dot_4x8/a55.cpp" ,
785805 " src/core/NEON/kernels/arm_gemm/kernels/a64_smallK_hybrid_u8u32_dot_4x8/generic.cpp" ,
806+ " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_bf16fp32_dot_4VLx4/generic.cpp" ,
807+ " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_bf16fp32_mmla_4VLx4/generic.cpp" ,
808+ " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_bf16fp32_mmla_6VLx2/generic.cpp" ,
809+ " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_bf16fp32_mmla_8VLx2/generic.cpp" ,
786810 " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_fp16_mla_4VLx4/generic.cpp" ,
787811 " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_fp32_mla_4VLx4/generic.cpp" ,
788812 " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_s8s32_dot_4VLx4/generic.cpp" ,
789813 " src/core/NEON/kernels/arm_gemm/kernels/sve_hybrid_u8u32_dot_4VLx4/generic.cpp" ,
814+ " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_bf16fp32_dot_3VLx8/generic.cpp" ,
815+ " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_bf16fp32_mmla_3VLx8/generic.cpp" ,
790816 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_fp16_mla_3VLx8/generic.cpp" ,
791817 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_fp32_mla_3VLx8/generic.cpp" ,
792818 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_s8s32_dot_3VLx8/generic.cpp" ,
793819 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_s8s32_mmla_3VLx8/generic.cpp" ,
794820 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_u8u32_dot_3VLx8/generic.cpp" ,
795821 " src/core/NEON/kernels/arm_gemm/kernels/sve_interleaved_u8u32_mmla_3VLx8/generic.cpp" ,
822+ " src/core/NEON/kernels/arm_gemm/kernels/sve_native_bf16fp32_dot_4VLx4/generic.cpp" ,
796823 " src/core/NEON/kernels/arm_gemm/kernels/sve_native_fp16_mla_4VLx4/generic.cpp" ,
797824 " src/core/NEON/kernels/arm_gemm/kernels/sve_native_fp32_mla_4VLx4/generic.cpp" ,
798825 " src/core/NEON/kernels/arm_gemm/kernels/sve_native_s8s32_dot_4VLx4/generic.cpp" ,
799826 " src/core/NEON/kernels/arm_gemm/kernels/sve_native_u8u32_dot_4VLx4/generic.cpp" ,
827+ " src/core/NEON/kernels/arm_gemm/kernels/sve_smallK_hybrid_fp32_mla_1VLx8/generic.cpp" ,
800828 " src/core/NEON/kernels/arm_gemm/kernels/sve_smallK_hybrid_s8s32_dot_1VLx8/generic.cpp" ,
801829 " src/core/NEON/kernels/arm_gemm/kernels/sve_smallK_hybrid_u8u32_dot_1VLx8/generic.cpp" ,
802830
0 commit comments