- Feb 18, 2025
-
-
Signed-off-by:
Jens Elofsson <jens.elofsson@arm.com> Reviewed-by:
Felix Johnny Thomasmathibalan <felixjohnny.thomasmathibalan@arm.com> Approved-by:
-
- Feb 17, 2025
-
-
- Add new assembly ukernel optimized with FEAT_DOTPROD for matrix multiplication with 1x8 block size. - Update build script. - Add to unit test. Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Reviewed-by: -
Add support for GEMV like kernel for producing QAI8 from QAI8 LHS and QSI8CXP packed RHS. Update unit tests to include support for new kernel Signed-off-by:
Emil Ohlsson <emil.ohlsson@arm.com> Reviewed-by:
Anton Bondarenko <anton.bondarenko@arm.com>
Reviewed-by: 
Jakub Sujak <jakub.sujak@arm.com>
-
- Feb 12, 2025
-
-
- Update build script - Add to unit test Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Reviewed-by: Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by: Anton Bondarenko <anton.bondarenko@arm.com>
-
- Feb 11, 2025
-
-
Resolves: #KLEIDIAI-485, #COMPMID-8024 Signed-off-by:
Anitha Raj <anitha.raj@arm.com> Reviewed-by:
Anton Bondarenko <anton.bondarenko@arm.com>
-
- Feb 06, 2025
-
-
- Update build script - Add to unit test Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Approved-by:
-
- Jan 31, 2025
-
-
- Added demonstration of klediai integration as external dependency Signed-off-by:
Mohammed Suhail Munshi <MohammedSuhail.Munshi@arm.com> Reviewed-by:
Mohammed Suhail Munshi <mohammedsuhail.munshi@arm.com> Reviewed-by:
-
- Jan 30, 2025
-
-
Signed-off-by:
Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by:
-
- Jan 29, 2025
-
-
Update all version indications Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Jan 28, 2025
-
-
Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Signed-off-by: Viet-Hoa Do <viet-hoa.do@arm.com> Reviewed-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
-
- Jan 27, 2025
-
-
In kai_matmul_clamp_f32_qai8dxp1vlx8_qsi4cxp4vlx8_1vlx4vl_sme2_mopa: * Fix the offset calculation * Fix pointer increments in the matmul Add new shapes to unit tests, to test n > 64 Resolves: #KLEIDIAI-405, #COMPMID-7918 Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com> -
Emil Ohlsson authored
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Jan 23, 2025
-
-
Emil Ohlsson authored
Downstream projects might not build with same toolchain, which might trigger warnings. Instead of treating these warnings as errors, using this patch they are only warnings. Also, add `-Werror` to the build flags to our testing pipeline, because we do not want to accidentally introduce new warnings by mistake Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Jan 22, 2025
-
-
Jakub Sujak authoredMention recent fixes for compiler warnings produced by enabling the `-Wcast-qual -Wmissing-prototypes -Wstrict-prototypes -Woverlength-strings` compiler options. Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by:
-
- Jan 16, 2025
-
-
Emil Ohlsson authored
Signed-off-by:
-
- Previously, we were dividing m_idx/n_idx by m_step/n_step. However, since m_step/n_step can be different from mr/nr, we should divide by mr/nr instead Signed-off-by:
Gian Marco Iodice <gianmarco.iodice@arm.com> Signed-off-by:
-
- Jan 10, 2025
-
-
Emil Ohlsson authored
Describe changes since 1.1.0 Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Dec 30, 2024
-
-
Jakub Sujak authored
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by:
-
- Dec 24, 2024
-
-
- Add unit test Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com> -
* GEMM and GEMV micro-kernels to compute the matrix multiplication of dynamically quantized 8-bit integer (QAI8DX) LHS matrix and quantized 4-bit integer (QSI4CX) RHS matrix and the accumulation of the result into a single-precision (F32) output, optimized for SME2 technology. Signed-off-by:
Mohamad Najem <mohamad.najem@arm.com> Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Signed-off-by: Thomas Bamelis <thomas.bamelis@arm.com> Reviewed-by:
Anton Bondarenko <anton.bondarenko@arm.com>
Reviewed-by: Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
-
- Dec 13, 2024
-
-
Emil Ohlsson authored
Changes reported release version, and updates changelog since last release Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com> -
There are a few items missing in the changelog since last relase. This commit updates the list with recent changes. Signed-off-by:
-
- Optimize the generic RHS packing NxK. The performance improvement is around ~1.5x Signed-off-by:
Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by: Anton Bondarenko <anton.bondarenko@arm.com> -
Signed-off-by:
alankelly <alankelly@google.com> Approved-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Dec 11, 2024
-
-
Jakub Sujak authored
The `__fp16` type is known to be troublesome for certain build environments. Instead use type `float` in the interface, and in the implementation use types `float16_t` (an alias for `__fp16`) for casting or `uint16_t` for pointer arithmetic. Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by:
-
- Dec 02, 2024
-
-
- GEMM and GEMV Micro-kernels to compute the matrix multiplication of dynamically quantized symmetric signed 8-bit integer with per-block quantization (QSI8D32) LHS matrix and quantized symmetric 4-bit signed integer with per-block quantization (QSI4C32) RHS matrix and the accumulation of the result into a single-precision (F32) output, optimized for SME2 technology. Signed-off-by:
Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
-
- Nov 29, 2024
-
-
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Anton Bondarenko <anton.bondarenko@arm.com>
-
- Nov 28, 2024
-
-
- Add GeMM-like micro-kernels - Add GeMV-like micro-kernels Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
-
- Nov 27, 2024
-
-
Signed-off-by:
Michael Kozlov <michael.kozlov@arm.com>
Signed-off-by: Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Anton Bondarenko <anton.bondarenko@arm.com>
-
- Nov 04, 2024
-
-
Jakub Sujak authored
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Anton Bondarenko <anton.bondarenko@arm.com>
Approved-by:
-
- Oct 16, 2024
-
-
Jakub Sujak authored
Compute the Vector-Matrix multiply of F32 inputs to produce an F32 matrix, optimized using SME instructions. Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by:
-
- Oct 14, 2024
-
-
Signed-off-by:
-
- Sep 27, 2024
-
-
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by: -
Signed-off-by:
-
Felix Johnny Thomasmathibalan authored
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Sep 25, 2024
-
-
This reverts commit f4f59599 The existing FP32 GEMM micro-kernel (matmul_clamp_f32_f32_f32p8x1biasf32_6x8x4_neon_mla) has a dedicated path for M=1 (a "GEMV" operation). Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by:
-
- Sep 12, 2024
-
-
- The LHS matrix is Quantized (Q) Asymmetric (A) Signed 8-bit (I8) with per-row (DX) quantization parameters - The RHS matrix is quantized (Q) Symmetric (S) Signed 4-bit (I4) with per-block quantization - The destination is F32 - Implement micro-kernels to perform the matrix multiplication - Implement a micro-kernel to pack the RHS matrix Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Reviewed-by: Max Ren <maxren@meta.com> Approved-by:
-
- Aug 30, 2024
-
-
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
-
- Aug 19, 2024
-
-
Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by:
-
- Aug 16, 2024
-
-
* The LHS matrix is Quantized (Q) Symmetric (S) Signed 8-bit (I8) with per-block quantization (D32) quantization parameters * The RHS matrix is Quantized (Q) Symmetric (S) Signed 4-bit (I4) with per-block quantization(C32) F16 scale factors, * The destination is F32 * Implement micro-kernels to perform the matrix multiplication * Implement a micro-kernel to pack the LHS and RHS matrices * Added unit tests Signed-off-by:
Gian Marco <Iodice gianmarco.iodice@arm.com> Signed-off-by:
Anitha <Raj Anitha.Raj@arm.com> Signed-off-by:
Jakub Sujak <jakub.sujak@arm.com>
Approved-by: Jakub Sujak <jakub.sujak@arm.com>
-
