AArch64: Remove literal pools from native code

[hanno-becker]

This commit removes all literal pools from the native AArch64 assembly. Those literal pools are slightly easier to read, but impede verification using HOL-Light. Instead, constant vectors are prepared by loading immediates into GPRs and copying/broadcasting them into the target vector.

[oqs-bot]

Arm Cortex-A76 (Raspberry Pi 5) benchmarks

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	29054 cycles	28986 cycles	1.00
ML-KEM-512 encaps	35419 cycles	35399 cycles	1.00
ML-KEM-512 decaps	45892 cycles	45896 cycles	1.00
ML-KEM-768 keypair	49360 cycles	49364 cycles	1.00
ML-KEM-768 encaps	55620 cycles	55564 cycles	1.00
ML-KEM-768 decaps	70402 cycles	70315 cycles	1.00
ML-KEM-1024 keypair	72108 cycles	71989 cycles	1.00
ML-KEM-1024 encaps	80825 cycles	80746 cycles	1.00
ML-KEM-1024 decaps	100687 cycles	100615 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

AMD EPYC 3rd gen (c6a)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	18134 cycles	18102 cycles	1.00
ML-KEM-512 encaps	23045 cycles	23015 cycles	1.00
ML-KEM-512 decaps	30254 cycles	30251 cycles	1.00
ML-KEM-768 keypair	31159 cycles	31124 cycles	1.00
ML-KEM-768 encaps	33948 cycles	33998 cycles	1.00
ML-KEM-768 decaps	44584 cycles	44518 cycles	1.00
ML-KEM-1024 keypair	44727 cycles	44599 cycles	1.00
ML-KEM-1024 encaps	49953 cycles	49893 cycles	1.00
ML-KEM-1024 decaps	64420 cycles	64399 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton2

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	29051 cycles	28986 cycles	1.00
ML-KEM-512 encaps	35402 cycles	35425 cycles	1.00
ML-KEM-512 decaps	45900 cycles	45888 cycles	1.00
ML-KEM-768 keypair	49379 cycles	49378 cycles	1.00
ML-KEM-768 encaps	55642 cycles	55565 cycles	1.00
ML-KEM-768 decaps	70454 cycles	70311 cycles	1.00
ML-KEM-1024 keypair	72117 cycles	71969 cycles	1.00
ML-KEM-1024 encaps	80849 cycles	80763 cycles	1.00
ML-KEM-1024 decaps	100733 cycles	100630 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton3

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	18954 cycles	18959 cycles	1.00
ML-KEM-512 encaps	23562 cycles	23572 cycles	1.00
ML-KEM-512 decaps	30704 cycles	30660 cycles	1.00
ML-KEM-768 keypair	32312 cycles	32312 cycles	1
ML-KEM-768 encaps	35873 cycles	35886 cycles	1.00
ML-KEM-768 decaps	46023 cycles	46021 cycles	1.00
ML-KEM-1024 keypair	46558 cycles	46634 cycles	1.00
ML-KEM-1024 encaps	52442 cycles	52456 cycles	1.00
ML-KEM-1024 decaps	66253 cycles	66268 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

AMD EPYC 3rd gen (c6a) (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	43354 cycles	43336 cycles	1.00
ML-KEM-512 encaps	51813 cycles	51844 cycles	1.00
ML-KEM-512 decaps	67069 cycles	67031 cycles	1.00
ML-KEM-768 keypair	71673 cycles	71630 cycles	1.00
ML-KEM-768 encaps	82778 cycles	82693 cycles	1.00
ML-KEM-768 decaps	102937 cycles	103012 cycles	1.00
ML-KEM-1024 keypair	106866 cycles	106602 cycles	1.00
ML-KEM-1024 encaps	121161 cycles	121422 cycles	1.00
ML-KEM-1024 decaps	147008 cycles	146875 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton3 (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	39345 cycles	39365 cycles	1.00
ML-KEM-512 encaps	45463 cycles	45441 cycles	1.00
ML-KEM-512 decaps	57383 cycles	57491 cycles	1.00
ML-KEM-768 keypair	65845 cycles	65827 cycles	1.00
ML-KEM-768 encaps	73795 cycles	73817 cycles	1.00
ML-KEM-768 decaps	89870 cycles	89874 cycles	1.00
ML-KEM-1024 keypair	98974 cycles	98958 cycles	1.00
ML-KEM-1024 encaps	110054 cycles	110050 cycles	1.00
ML-KEM-1024 decaps	130823 cycles	130832 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton2 (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	60705 cycles	60686 cycles	1.00
ML-KEM-512 encaps	69882 cycles	69803 cycles	1.00
ML-KEM-512 decaps	88684 cycles	88744 cycles	1.00
ML-KEM-768 keypair	101829 cycles	101761 cycles	1.00
ML-KEM-768 encaps	113963 cycles	113894 cycles	1.00
ML-KEM-768 decaps	139384 cycles	139325 cycles	1.00
ML-KEM-1024 keypair	154338 cycles	154153 cycles	1.00
ML-KEM-1024 encaps	170098 cycles	169847 cycles	1.00
ML-KEM-1024 decaps	202520 cycles	202209 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[mkannwischer]

Changes look good. I'd like to see one complete benchmarking run after your last force push just to be sure.

[oqs-bot]

Intel Xeon 4th gen (c7i)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	13526 cycles	13975 cycles	0.97
ML-KEM-512 encaps	17268 cycles	17236 cycles	1.00
ML-KEM-512 decaps	22793 cycles	23052 cycles	0.99
ML-KEM-768 keypair	22474 cycles	22520 cycles	1.00
ML-KEM-768 encaps	24503 cycles	24524 cycles	1.00
ML-KEM-768 decaps	32399 cycles	32548 cycles	1.00
ML-KEM-1024 keypair	31487 cycles	31383 cycles	1.00
ML-KEM-1024 encaps	35012 cycles	34928 cycles	1.00
ML-KEM-1024 decaps	45965 cycles	45798 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Intel Xeon 4th gen (c7i) (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	33160 cycles	33236 cycles	1.00
ML-KEM-512 encaps	38697 cycles	38554 cycles	1.00
ML-KEM-512 decaps	50864 cycles	50857 cycles	1.00
ML-KEM-768 keypair	54850 cycles	54878 cycles	1.00
ML-KEM-768 encaps	60668 cycles	60646 cycles	1.00
ML-KEM-768 decaps	75867 cycles	75857 cycles	1.00
ML-KEM-1024 keypair	81837 cycles	81938 cycles	1.00
ML-KEM-1024 encaps	91696 cycles	91771 cycles	1.00
ML-KEM-1024 decaps	111498 cycles	111446 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Intel Xeon 3rd gen (c6i)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	20348 cycles	20350 cycles	1.00
ML-KEM-512 encaps	26955 cycles	26953 cycles	1.00
ML-KEM-512 decaps	36009 cycles	35746 cycles	1.01
ML-KEM-768 keypair	34900 cycles	34886 cycles	1.00
ML-KEM-768 encaps	38181 cycles	38182 cycles	1.00
ML-KEM-768 decaps	50956 cycles	50946 cycles	1.00
ML-KEM-1024 keypair	47961 cycles	47950 cycles	1.00
ML-KEM-1024 encaps	54110 cycles	54099 cycles	1.00
ML-KEM-1024 decaps	71576 cycles	71603 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

AMD EPYC 4th gen (c7a)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	14920 cycles	14915 cycles	1.00
ML-KEM-512 encaps	19650 cycles	19645 cycles	1.00
ML-KEM-512 decaps	26298 cycles	26297 cycles	1.00
ML-KEM-768 keypair	25599 cycles	25589 cycles	1.00
ML-KEM-768 encaps	28060 cycles	28078 cycles	1.00
ML-KEM-768 decaps	37929 cycles	37792 cycles	1.00
ML-KEM-1024 keypair	35679 cycles	35753 cycles	1.00
ML-KEM-1024 encaps	40966 cycles	40945 cycles	1.00
ML-KEM-1024 decaps	54445 cycles	54417 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Intel Xeon 3rd gen (c6i) (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	51411 cycles	51419 cycles	1.00
ML-KEM-512 encaps	59575 cycles	59547 cycles	1.00
ML-KEM-512 decaps	76550 cycles	76562 cycles	1.00
ML-KEM-768 keypair	84259 cycles	84278 cycles	1.00
ML-KEM-768 encaps	94964 cycles	94991 cycles	1.00
ML-KEM-768 decaps	117113 cycles	117180 cycles	1.00
ML-KEM-1024 keypair	124532 cycles	124782 cycles	1.00
ML-KEM-1024 encaps	138737 cycles	138756 cycles	1.00
ML-KEM-1024 decaps	167364 cycles	167416 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

AMD EPYC 4th gen (c7a) (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	39585 cycles	39375 cycles	1.01
ML-KEM-512 encaps	45641 cycles	45586 cycles	1.00
ML-KEM-512 decaps	59101 cycles	59050 cycles	1.00
ML-KEM-768 keypair	64510 cycles	64596 cycles	1.00
ML-KEM-768 encaps	73138 cycles	72846 cycles	1.00
ML-KEM-768 decaps	91091 cycles	91403 cycles	1.00
ML-KEM-1024 keypair	96044 cycles	95969 cycles	1.00
ML-KEM-1024 encaps	107178 cycles	107130 cycles	1.00
ML-KEM-1024 decaps	130762 cycles	130669 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton4

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	18126 cycles	18116 cycles	1.00
ML-KEM-512 encaps	22175 cycles	22178 cycles	1.00
ML-KEM-512 decaps	28835 cycles	28839 cycles	1.00
ML-KEM-768 keypair	30559 cycles	30560 cycles	1.00
ML-KEM-768 encaps	33617 cycles	33637 cycles	1.00
ML-KEM-768 decaps	43168 cycles	43158 cycles	1.00
ML-KEM-1024 keypair	44169 cycles	44163 cycles	1.00
ML-KEM-1024 encaps	49653 cycles	49653 cycles	1
ML-KEM-1024 decaps	62606 cycles	62642 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Graviton4 (no-opt)

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	37985 cycles	38042 cycles	1.00
ML-KEM-512 encaps	43306 cycles	43379 cycles	1.00
ML-KEM-512 decaps	55588 cycles	55553 cycles	1.00
ML-KEM-768 keypair	63189 cycles	63014 cycles	1.00
ML-KEM-768 encaps	70437 cycles	70323 cycles	1.00
ML-KEM-768 decaps	86874 cycles	86776 cycles	1.00
ML-KEM-1024 keypair	94505 cycles	94468 cycles	1.00
ML-KEM-1024 encaps	105259 cycles	105175 cycles	1.00
ML-KEM-1024 decaps	126500 cycles	126797 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Arm Cortex-A55 (Snapdragon 888) benchmarks

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	58346 cycles	58327 cycles	1.00
ML-KEM-512 encaps	65747 cycles	65795 cycles	1.00
ML-KEM-512 decaps	84554 cycles	84608 cycles	1.00
ML-KEM-768 keypair	98987 cycles	99034 cycles	1.00
ML-KEM-768 encaps	110142 cycles	110325 cycles	1.00
ML-KEM-768 decaps	137284 cycles	137025 cycles	1.00
ML-KEM-1024 keypair	150235 cycles	150344 cycles	1.00
ML-KEM-1024 encaps	166495 cycles	166740 cycles	1.00
ML-KEM-1024 decaps	202438 cycles	202805 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Bananapi bpi-f3 benchmarks

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	334919 cycles	334942 cycles	1.00
ML-KEM-512 encaps	443623 cycles	443716 cycles	1.00
ML-KEM-512 decaps	591660 cycles	591749 cycles	1.00
ML-KEM-768 keypair	559257 cycles	559270 cycles	1.00
ML-KEM-768 encaps	697687 cycles	697687 cycles	1
ML-KEM-768 decaps	890117 cycles	890201 cycles	1.00
ML-KEM-1024 keypair	828054 cycles	828159 cycles	1.00
ML-KEM-1024 encaps	1000522 cycles	999913 cycles	1.00
ML-KEM-1024 decaps	1231962 cycles	1232943 cycles	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[oqs-bot]

Arm Cortex-A72 (Raspberry Pi 4) benchmarks

Benchmark suite	Current: 7d15fdb	Previous: e4ff720	Ratio
ML-KEM-512 keypair	51624 cycles	51568 cycles	1.00
ML-KEM-512 encaps	58057 cycles	58005 cycles	1.00
ML-KEM-512 decaps	74302 cycles	74147 cycles	1.00
ML-KEM-768 keypair	88316 cycles	87902 cycles	1.00
ML-KEM-768 encaps	97274 cycles	96043 cycles	1.01
ML-KEM-768 decaps	119245 cycles	119330 cycles	1.00
ML-KEM-1024 keypair	131103 cycles	131867 cycles	0.99
ML-KEM-1024 encaps	144551 cycles	145139 cycles	1.00
ML-KEM-1024 decaps	177206 cycles	176050 cycles	1.01

This comment was automatically generated by workflow using github-action-benchmark.