aboutsummaryrefslogtreecommitdiff
path: root/src/core/NEON/kernels/arm_gemm/gemm_quint8.cpp
AgeCommit message (Collapse)Author
2024-05-01New SME2 heuristics.David Mansell
Change-Id: I69aa973e61df950060807a31230a1edd91add498 Signed-off-by: David Mansell <David.Mansell@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11514 Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2024-03-18Fix quant. gemv kernel driver by adding set_quantized_bias()Gunes Bayir
arm_gemm fuses the actual bias addition with the output stage in quantized gemm. The output stage, in its very basic form, is: A_offset * B_offset - sum(A_row_i) * B_offset - sum(B_col_j) * A_offset Matrix B is usually constant (e.g. weight matrix in convolutions). Therefore, except the middle term above, the expression is constant across the same output row because the column sums of matrix B are pre-calculated. The bias is also usually constant. When it is, it makes sense to add the bias vector to the above sum and just perform a single addition on top of the output tensor. For this to happen, the column sum computation of B tensor must account for the bias. This is ensured by set_quantized_bias() method in the interface. This function passes the bias pointer and strides to arm_gemm. Gemv_pretransposed does not implement set_quantized_bias() and uses the parent function, which does nothing. Therefore, the bias is not added to the output. This causes tests to fail. Resolves: COMPMID-6928 Change-Id: Iba24fabc65fdc47edb12db6abff2fb47784c0743 Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11310 Benchmark: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
2024-02-06arm_gemm: SME: Remove artificial single-thread constraint on quantized int8 ↵David Mansell
kernels. Change-Id: I81b71ecc0d2e776d132091e074798a79b3141bec Signed-off-by: David Mansell <David.Mansell@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/11085 Reviewed-by: Jakub Sujak <jakub.sujak@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2023-06-05Update CPU kernel implementations and guard directivesMichael Tyler
Resolves COMPMID-6023 Change-Id: I868975d14c4f98af6716726feda22405a6a4c891 Signed-off-by: Michael Tyler <michael.tyler@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9686 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
2023-03-07GEMM: SME: Allow threading for quantized GEMMs.David Mansell
The SME kernels for quantized int8/uint8 GEMMs erroneously required that maxthreads==1 before they could be selected. This resulted in them not being available on multi-thread runs. Remove that restriction. Resolves COMPMID-5962 Change-Id: Ia7933d0c66020b5e2981604ca97ff7ead95ec14e Signed-off-by: David Mansell <David.Mansell@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9274 Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2023-02-08Update CPU kernels to remove x19 and w19Michael Tyler
Resolves: COMPMID-5805 Change-Id: Idf720bbb136474810086f5089c5ed23b3f79835a Signed-off-by: Michael Tyler <michael.tyler@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/9081 Benchmark: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com>
2023-01-18Revert "Update CPU kernels to remove x19"Michael Tyler
This reverts commit 3c59f01c209d2732a15d97d65565ead964787a8b. Resolves: COMPMID-5817 Change-Id: Ie2443a21854a95db1e3d0cafa2121c0187a5e237 Signed-off-by: Michael Tyler <michael.tyler@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8974 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
2023-01-16Update CPU kernels to remove x19Michael Tyler
Resolves: COMPMID-5805 Signed-off-by: Michael Tyler <michael.tyler@arm.com> Change-Id: I250f64531e209625e4ff176dd5a552c1c34bc484 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8909 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Reviewed-by: Viet-Hoa Do <viet-hoa.do@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
2022-11-28Integrate SME2 kernelsViet-Hoa Do
* Add SME/SME2 detection. * Integrate SME2 implementation for: - Normal convolution - Winograd - Depthwise convolution - Pooling Resolves: COMPMID-5700 Signed-off-by: Viet-Hoa Do <viet-hoa.do@arm.com> Change-Id: I2f1ca1d05f8cfeee9309ed1c0a36096a4a6aad5c Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/8692 Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-07-19[ONCPUML-951] Variable weight support for Convolution.Francesco Petrogalli
API changes for NEGEMMConvolutionLayer and CpuGemmConv2d Built with: scons neon=1 opencl=0 os=linux arch=armv8.2-a multi_isa=1 \ build=native -j32 Werror=false validation_tests=1 build_dir=opt \ standalone=1 asserts=1 experimental_fixed_format_kernels=1 . Tested with: ./build/opt/tests/arm_compute_validation Hardware where the test executable was run: Neoverse N1 Test coverage: * NEGEMMConvolutionLayer, CpuGemmConv2d * NHWC (the only one supported by the fixed-format kernels) * F16, F32 * Shapes: RunSmall Change-Id: I4fd3e495a7cbf61210ea02d37440ba9652934e99 Signed-off-by: Francesco Petrogalli <francesco.petrogalli@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7632 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gunes Bayir <gunes.bayir@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Benchmark: Arm Jenkins <bsgcomp@arm.com>
2022-04-06[arm_gemm] Use static validate to find arm_gemm kernels.Francesco.Petrogalli@arm.com
The static method `CpuGemmAssemblyDispatch::validate` should look into the list of the available kernels to make sure the one requested by the user was found. Formatting changes in the files touched by the patch have been automatically inserted by the formatting script. Resolves: ONCPUML-840 Change-Id: Icd650a30e142284a942c64f8a2b72441ee7b3f4e Signed-off-by: Francesco.Petrogalli@arm.com <francesco.petrogalli@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7375 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Giorgio Arena <giorgio.arena@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2022-03-16Remove deprecated interface from arm_compute.Francesco.Petrogalli@arm.com
The function `get_gemm_method` in arm_compute is deprecated in favor of the method `arm_gemm::GemmCommon<TypeInput, TypeOutput>::get_config()`. Signed-off-by: francesco.petrogalli@arm.com Change-Id: Idd5d879180c3995d5a07a727aa9216b8f94f01ba Signed-off-by: Francesco.Petrogalli@arm.com <francesco.petrogalli@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/7304 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sheri Zhang <sheri.zhang@arm.com>
2021-09-15Update cpu heuristics for GEMMGunes Bayir
If available, it prefers MMLA over DOT for smallK kernels for int8 types for faster execution. Signed-off-by: Gunes Bayir <gunes.bayir@arm.com> Change-Id: I190142dc66555476b4d8266d22e54ce8b45fffdb Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6251 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-07-22Update GEMM assembly kernelsGeorgios Pinitas
- Introduce Fp32 kernels with internal calculations in Bfloat16 when fast_mode is enabled - Improve kernel selection heuristics Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: I68a9e7e862b6fd2721b46e0d7cc791091c4ab279 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5965 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-06-23Create core library using high priority operatorsMichalis Spyrou
A smaller core library is created using a subset of the operators. Changed the structure of filelist.json in order to include more information about the kernels and make the selection easier. Resolves: COMPMID-4514 Change-Id: I079ca7d8e64346174eebdd13b834e1dd4dc36ca2 Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5786 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-03-17Updated cpu detectionPablo Marquez Tello
* Added the case in the cpu detection code for Klein cores * Added has_sve() and set_sve() methods in CpuInfo * Detection code checks for presence of SVE via HWCAP_SVE * Updated the heuristic in sve kernels to check for the absence of Klein * Resolves: COMPMID-4085 Change-Id: I0b8c72ff19dc5a3a81628d121a1afa836e724b4f Signed-off-by: Pablo Marquez Tello <pablo.tello@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/5257 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2021-01-18Cycle estimate-based kernel selection for dot product quantized s8/u8 kernelsGeorgios Pinitas
Resolves: COMPMID-3990 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: If840c79209940535450f4ea1cbf6b0ec646a168e Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4866 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-12-23Fix baremetal arm_compute_validation build errorsSiCongLi
* Add -C flag to instruct preprocessor not to strip comments. This is to prevent marker comments like '// fall through' that suppresses certain warnings from being removed. * Fix unused variable warnings. * Add M_PI definition that's missing from certain toolchain standard libraries. Resolves COMPMID-4054 Change-Id: I1d641db668685d4b678f3d0efed84bfe9e630b4b Signed-off-by: SiCongLi <sicong.li@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4692 Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2020-12-03Update GEMV heuristics for quantized types for A53Georgios Pinitas
Switch assembly kernels to dispatch a 4x4 blocked GEMM kernel for A53 when M <= 4 instead of the 8x12 u16 based one. Resolves: COMPMID-3983 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: Ic46a1b51a7c075e46dcb5cd578c75260ded0540c Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4640 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-11-12COMPMID-3776: Indirect GEMMGeorgios Pinitas
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: I51a1b0f098bc3a8c408c50c92221e4df3061e12c Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/4343 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Sang-Hoon Park <sang-hoon.park@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-07-24[ONCPUML-120]: Tweak of the launch heuristics for hybrid_u8u32_dot_16x4 kernelAleksandr Nikolaev
Hybrid kernel turns to be faster for qasymm8 than quantized_wrapper with interleaved. Signed-off-by: Aleksandr Nikolaev <aleksandr.nikolaev@arm.com> Change-Id: I200646aee6cdcabfe125b746c7d87bfa7d06e0fc Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3585 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-07-16[ONCPUML-97]: Implement "int8" support for 2D decomposition at high core countsAleksandr Nikolaev
Interleaved2d functionality was extended to uint8 and int8 kernels. Change-Id: If78facbce56e9ec7b2f4c23436af0bd5db7f7b69 Signed-off-by: Aleksandr Nikolaev <aleksandr.nikolaev@arm.com> Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3467 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-07-07COMPMID-3324: Remove pretransposed support from NEON backendGeorgios Pinitas
Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Change-Id: I394c6c539969940e0119cbc14174909d47e65de6 Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/3519 Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Michele Di Giorgio <michele.digiorgio@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2020-01-23COMPMID-2923 Integrate arm_gemm per channel quantizationMichalis Spyrou
Signed-off-by: Michalis Spyrou <michalis.spyrou@arm.com> Change-Id: I8667e75843fdd6ac75bd8272a86a348b830da28d Reviewed-on: https://review.mlplatform.org/c/2548 Tested-by: Arm Jenkins <bsgcomp@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
2019-10-23COMPMID-2577: Fuse bias addition and activation in gemm assembly kernelsGeorgios Pinitas
Change-Id: I7f52112d2d05b1ea3d3f3d4b19b8eafab05d6c44 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-on: https://review.mlplatform.org/c/2141 Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com> Reviewed-by: Pablo Marquez <pablo.tello@arm.com>
2019-07-29COMPMID-2336: Guard quint8 gemm files for aarch64 only.Georgios Pinitas
gemm_quint8 is only supported for 64-bit thus guarding to avoid any build related issues. Change-Id: Id8784dbacc467780318bd340f895a5abbd383182 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-on: https://review.mlplatform.org/c/1638 Reviewed-by: Pablo Marquez <pablo.tello@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
2019-07-26COMPMID-2178: Update GEMM assembly code.Georgios Pinitas
Perform offset reduction and requantization within the assembly wrapper. Change-Id: I5d5b3e1f6f9ef4c71805362c57f88ff199c027a3 Signed-off-by: Georgios Pinitas <georgios.pinitas@arm.com> Reviewed-on: https://review.mlplatform.org/c/1541 Comments-Addressed: Pablo Marquez <pablo.tello@arm.com> Reviewed-by: Gian Marco Iodice <gianmarco.iodice@arm.com> Comments-Addressed: Arm Jenkins <bsgcomp@arm.com> Tested-by: Arm Jenkins <bsgcomp@arm.com>
generated by cgit v1.2.1 (git 2.23.0) at 2024-05-17 05:49:39 +0000