OpenBLAS

Commit Graph

Author	SHA1	Message	Date
Chris Sidebottom	ea2faf0c9a	Add optimized BGEMM for NEOVERSEN2 target This re-uses the existing NEOVERSEN2 8x4 `sbgemm` kernel to implement `bgemm`.	6 months ago
Annop Wongwathanarat	e11744a411	Use SVE kernel for S/DGEMVN for SVE machines	9 months ago
Annop Wongwathanarat	ec146157d3	Use SVE kernel for S/DGEMVT for SVE machines	10 months ago
Ye Tao	38ee7c9301	Add dispatch of SBGEMVNKERNEL for NEOVERSEN2 and NEOVERSEV2	11 months ago
Annop Wongwathanarat	edaf51dd99	Add sbgemv_t_bfdot kernel for ARM64 This improves performance for sbgemv_t by up to 100x on NEOVERSEV1. The geometric mean speedup is ~61x for M=N=[2,512].	11 months ago
Martin Kroeker	3345007d8f	retire the thunderx2 NRM2 kernels due to reported inaccuracies and NAN	1 year ago
Martin Kroeker	09ace3cf23	Merge pull request #3846 from lilh9598/sbgemm_opt Improve the performance of sbgemm_tcopy on neoversen2	2 years ago
Chris Sidebottom	fd4f52c797	Add SVE implementation for sdot/ddot This adds an SVE implementation to sdot/ddot when available, falling back to the previous Advanced SIMD kernel where there's no SVE implementation for the kernel. All the targets were essentially treating `dot_thunderx2t99.c` as the Advanced SIMD implementation so I've renamed it to better fit with the feature detection.	3 years ago
lilianhuang	fdac8a97c1	Add sbgemm_ncopy_8 and sbgemm_tcopy_4	3 years ago
Honglin Zhu	79066b6bf3	Change file name to match the norm and delete useless code.	3 years ago
Honglin Zhu	b00d5b9746	New sbgemm implementation for Neoverse N2 1. Use UZP instructions but not gather load and scatter store instructions to get lower latency. 2. Padding k to a power of 4.	3 years ago
Honglin Zhu	55d686d41e	neoverse n2 sbgemm: implement ncopy tcopy kernel_8x4	3 years ago
Honglin Zhu	04593bb27c	neoverse n2 sbgemm: init file	3 years ago
Sunita Nadampalli	19c8f615dc	OpenBLAS: aarch64: Add neoverse-v1/n2 architecture specifics	4 years ago

14 Commits (develop)