126 commits

Author	SHA1	Message	Date
Andrea Mazzoleni	2db6d7d6ba	Refactor and reorganize parity generation functions using inline templates	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	21739d1d53	Add AArch64 NEON parity generation and recovery functions	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	f5442bda18	Use vpternlogq in raid_gen3-6_avx512bw to fuse sequential vpxorq instructions into single 3-way XORs	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	05b7bcb6b9	Gate AVX-512BW gen1/gen2 behind slow_avx512() to avoid double-pumped shuffle penalties	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	1f95857e80	Remove the unused number of data disks from the vector allocation ... The same addressing rules are applied to the parity disks because they are used as data disks during recovery operations.	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	cf3ef036de	Inject variable page-offset noise to disk buffer displacement to mitigate prefetcher contention	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	cf3ef337a7	Use the optimal displacement depending on the number of disks	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	28f61c11bd	Change RAID AVX512/GFNI functions to use forward iterations	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	a4e8a68805	Remove the special case of a single data disk when not strictly required	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	102eee70c6	Remove unused level 0 from the pshufb table	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	f838c7434a	Change the AVX512BW generation to use pshufb also for Q parity	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	b2f88ade14	Extend the pshufb table to include the Q parity coefficients	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	826cbe70bc	New displacement benchmarks	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	d15da0172a	Simplify GFNI implementations	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	d4a3ea54ff	Always build RAID test with USE_RAID_AES	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	0c1580767e	Assume the assembler supports all x86 SSE/AVX instructions	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	affb3790ab	Extend the RAID regression test to detect unset parity	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	d8beca84c5	Add support for AVX2 GFNI	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	96005850dd	Disable AVX512 on CPUs with narrow shuffle ports or double-pumped 256-bit datapath	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	b5d5234847	Revise the RAID library Makefile	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	0753bedb30	Be more clear when GFNI is really used	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	2865662b2c	Optimize AVX512 gen1 parity with the vpternlogq instruction	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	53d10db9fd	Add tables for the AES Polynomial	2026-06-10 15:28:54 +02:00
Andrea Mazzoleni	d6db8355ef	Comment about the RAID optimizations	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	d4a3179d44	Use a simpler approach in the new AVX512 generation functions	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	9eca9623b3	Preliminary support for AVX512GFNI	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	5956ecbbd1	Change commenting style	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	825f02d45a	Add detection of AVX2GFNI	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	2621d74fcf	Add more AVX512 functions for parity	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	f660338084	Preliminary support for ADVX512	2026-06-10 15:28:53 +02:00
Christoph Hellwig	9135379dab	Make all RAID forwarders static and local using a new registration interface ... Co-authored-by: Andrea Mazzoleni <amadvance@gmail.com>	2026-06-10 15:28:53 +02:00
Andrea Mazzoleni	ae9b472898	Add detection of AVX512BW and AVX512GFNI	2026-04-16 00:13:51 +02:00
Andrea Mazzoleni	b709371d2d	Fix typos in the comment of the raid library	2026-04-15 10:53:25 +02:00
Andrea Mazzoleni	0d0431a907	Simpler Copyright string	2026-04-09 17:56:49 +02:00
Christoph Hellwig	60269a1236	Use SPDX (https://spdx.dev/) tags to identify the license used for each ... file, and a COPYING file explaining the licensing scheme, with the actual licenses listed in the LICENSES/ directory.	2026-04-02 16:01:37 +02:00
Andrea Mazzoleni	e80734fc4c	Fix typos	2026-01-22 16:13:25 +01:00
Andrea Mazzoleni	f7798b2357	Update performance test of raid.c	2025-10-14 23:26:34 +02:00
Andrea Mazzoleni	13a4872e69	Revise the top comment of raid.c	2025-10-13 17:32:55 +02:00
Andrea Mazzoleni	56c4076e23	Avoid a warning about function pointer conversion. No functional changes	2025-01-23 15:51:59 +01:00
Andrea Mazzoleni	607564c5fe	Fix some typos	2023-12-09 10:45:26 +01:00
Jens Kühnel	6b3ac2a6cb	Update address of FSF in GPLv2	2023-08-09 14:44:48 +02:00
Andrea Gelmini	2dacb15516	Fix typos ... Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>	2022-07-21 14:48:48 +02:00
Andrea Gelmini	986e9b4131	Fix typos	2020-05-26 12:50:15 +02:00
Andrea Mazzoleni	21241c5078	Preliminary advise support	2016-05-15 14:50:59 +02:00
Andrea Mazzoleni	9430106c71	Allow to disable inline assembly from configure ... This is required for the -fsanitize options that don't support inline assembly statements.	2016-01-09 11:19:52 +01:00
Andrea Mazzoleni	cffa2f0c10	Better Atom prediction of fastest RAID implementation	2015-11-21 09:08:53 +01:00
Andrea Mazzoleni	77117b1cd5	Add some more warning checks, and fix the ones reported	2015-08-05 09:42:11 +02:00
Andrea Mazzoleni	1c06674d1a	Some comments about x86 optimizations	2015-06-20 16:47:50 +02:00
Andrea Mazzoleni	2ae9d8d41a	Add recognition of new Atom CPUs	2015-06-04 22:38:32 +02:00
Andrea Mazzoleni	35b1b08b6c	Clean the cpu.h header	2015-03-30 23:36:36 +02:00