hwcap の変更点

• 追加された行はこの色です。
• 削除された行はこの色です。
• hwcap へ行く。

[[auxv]]

|cpuinfo   |location|bit    |Description|notes|
|fp        |HWCAP   |1 << 0 |Convert between fp32 and fp16|Probably required for all Armv8 and later|
|asimd     |HWCAP   |1 << 1 |Advanced SIMD|Probably required for all Armv8 and later|
|evtstrm   |HWCAP   |1 << 2 |Event stream ||
|aes       |HWCAP   |1 << 3 |AES crypto instructions||
|pmull     |HWCAP   |1 << 4 |Polynomial multiply instructions|Probably related to crypto graphic|
|sha1      |HWCAP   |1 << 5 |Compute SHA1 hash||
|sha2      |HWCAP   |1 << 6 |Compute SHA2 hash||
|crc32     |HWCAP   |1 << 7 |Compute CRC32 hash||
|atomics   |HWCAP   |1 << 8 |Atomic load/store instructions|https://cpufun.substack.com/p/atomics-in-aarch64|
|fphp      |HWCAP   |1 << 9 |Instruction to directly compute in half||
|asimdhp   |HWCAP   |1 << 10|Instruction to directly compute in half in vector||
|cpuid     |HWCAP   |1 << 11|||
|asimdrdm  |HWCAP   |1 << 12|Instruction SQRDMULH supported|Signed saturating Rounding Doubling Multiply returning High half https://developer.arm.com/documentation/ddi0596/2021-12/SIMD-FP-Instructions/SQRDMLAH--vector---Signed-Saturating-Rounding-Doubling-Multiply-Accumulate-returning-High-Half--vector--?lang=en|
|jscvt     |HWCAP   |1 << 13|Instruction specially for Javasciprt|https://qiita.com/mod_poppo/items/66663d9a05790579b5e4|
|fcma      |HWCAP   |1 << 14|||
|lrcpc     |HWCAP   |1 << 15|||
|dcpop     |HWCAP   |1 << 16|||
|sha3      |HWCAP   |1 << 17|||
|sm3       |HWCAP   |1 << 18|||
|sm4       |HWCAP   |1 << 19|||
|asimddp   |HWCAP   |1 << 20|Compute dot product of four int8 and add up in int32|Mainly for DL usage|
|sha512    |HWCAP   |1 << 21|Compute SHA512 hash||
|sve       |HWCAP   |1 << 22|Scalable Vector Engine||
|asimdfhm  |HWCAP   |1 << 23|||
|dit       |HWCAP   |1 << 24|||
|uscat     |HWCAP   |1 << 25|||
|ilrcpc    |HWCAP   |1 << 26|||
|flagm     |HWCAP   |1 << 27|||
|ssbs      |HWCAP   |1 << 28|||
|sb        |HWCAP   |1 << 29|||
|paca      |HWCAP   |1 << 30|||
|pacg      |HWCAP   |1 << 31|||
|dcpodp    |HWCAP2  |1 << 0 |||
|sve2      |HWCAP2  |1 << 1 |Scalable Vector Engine2||
|sveaes    |HWCAP2  |1 << 2 |||
|svepmull  |HWCAP2  |1 << 3 |||
|svebitperm|HWCAP2  |1 << 4 |||
|svesha3   |HWCAP2  |1 << 5 |||
|svesm4    |HWCAP2  |1 << 6 |||
|flagm2    |HWCAP2  |1 << 7 |||
|frint     |HWCAP2  |1 << 8 |||
|svei8mm   |HWCAP2  |1 << 9 |||
|svef32mm  |HWCAP2  |1 << 10|||
|svef64mm  |HWCAP2  |1 << 11|||
|svebf16   |HWCAP2  |1 << 12|||
|i8mm      |HWCAP2  |1 << 13|||
|bf16      |HWCAP2  |1 << 14|Instructions that treats bf16||
|dgh       |HWCAP2  |1 << 15|||
|rng       |HWCAP2  |1 << 16|||
|bti       |HWCAP2  |1 << 17|||
|mte       |HWCAP2  |1 << 18|||
|ecv       |HWCAP2  |1 << 19|||
|afp       |HWCAP2  |1 << 20|||
|rpres     |HWCAP2  |1 << 21|||
|mte3      |HWCAP2  |1 << 22|||

- https://elixir.bootlin.com/linux/latest/source/arch/arm64/include/uapi/asm/hwcap.h
- https://elixir.bootlin.com/linux/latest/source/arch/arm64/kernel/cpufeature.c#L2527
- https://documentation-service.arm.com/static/5fb7c9cdca04df4095c1d5e4?token=
- https://community.arm.com/arm-community-blogs/b/architectures-and-processors-blog/posts/armv8-a-architecture-2016-additions
- https://community.arm.com/arm-community-blogs/b/architectures-and-processors-blog/posts/arm-architecture-armv8-2-a-evolution-and-delivery
- https://en.wikichip.org/wiki/arm/armv8