Implementation notes: amd64, h3neo, crypto_aead/paeq80

Computer: h3neo
Microarchitecture: amd64; K10 45nm (100f63)
Architecture: amd64
CPU ID: AuthenticAMD-00100f63-078bfbff
SUPERCOP version: 20230530
Operation: crypto_aead
Primitive: paeq80

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
2651962	13084 32 65536	28792 916 66560	`T:ref`	`clang++_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230630	20230530
3340500	19436 28 65536	36405 840 66656	`T:ref`	`g++_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230630	20230530
3948319	12588 32 65536	27376 916 66560	`T:ref`	`clang++_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230630	20230530
10304785	6789 32 65536	19890 908 66560	`T:ref`	`clang++_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230630	20230530
10625508	6516 32 65536	20488 916 66560	`T:ref`	`clang++_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230630	20230530
17662291	7810 28 65536	23493 840 66656	`T:ref`	`g++_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230630	20230530
21189037	6392 28 65536	21212 832 66656	`T:ref`	`g++_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230630	20230530
26436274	5323 28 65536	18872 816 66624	`T:ref`	`g++_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230630	20230530

Compiler output

Implementation: T:aesni
Security model: timingleaks
Compiler: clang++ -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

encrypt.cpp: encrypt.cpp:166:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc0 = _mm_aesenc_si128(acc0, roundkey0);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:167:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc1 = _mm_aesenc_si128(acc1, roundkey1);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:168:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc2 = _mm_aesenc_si128(acc2, roundkey2);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:169:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc3 = _mm_aesenc_si128(acc3, roundkey3);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:171:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc4 = _mm_aesenc_si128(acc4, roundkey0);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:172:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc5 = _mm_aesenc_si128(acc5, roundkey1);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:173:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc6 = _mm_aesenc_si128(acc6, roundkey2);
encrypt.cpp: ^
encrypt.cpp: encrypt.cpp:174:11: error: always_inline function '_mm_aesenc_si128' requires target feature 'aes', but would be inlined into function 'DoubleFPermAsm' that is compiled without support for 'aes'
encrypt.cpp: acc7 = _mm_aesenc_si128(acc7, roundkey3);
encrypt.cpp: ^
encrypt.cpp: 8 errors generated.

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
clang++ -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:aesni
clang++ -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:aesni
clang++ -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:aesni
clang++ -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:aesni

Compiler output

Implementation: T:aesni
Security model: timingleaks
Compiler: g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

encrypt.cpp: In file included from encrypt.cpp:13:
encrypt.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/wmmintrin.h: In function 'void DoubleFPermAsm(const __m128i*, const __m128i*, __m128i*, __m128i*)':
encrypt.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/wmmintrin.h:61:1: error: inlining failed in call to 'always_inline' '__m128i _mm_aesenc_si128(__m128i, __m128i)': target specific option mismatch
encrypt.cpp: 61 | _mm_aesenc_si128 (__m128i __X, __m128i __Y)
encrypt.cpp: | ^~~~~~~~~~~~~~~~
encrypt.cpp: encrypt.cpp:174:48: note: called from here
encrypt.cpp: 174 | acc7 = _mm_aesenc_si128(acc7, roundkey3);
encrypt.cpp: | ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~
encrypt.cpp: In file included from encrypt.cpp:13:
encrypt.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/wmmintrin.h:61:1: error: inlining failed in call to 'always_inline' '__m128i _mm_aesenc_si128(__m128i, __m128i)': target specific option mismatch
encrypt.cpp: 61 | _mm_aesenc_si128 (__m128i __X, __m128i __Y)
encrypt.cpp: | ^~~~~~~~~~~~~~~~
encrypt.cpp: encrypt.cpp:173:48: note: called from here
encrypt.cpp: 173 | acc6 = _mm_aesenc_si128(acc6, roundkey2);
encrypt.cpp: | ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~
encrypt.cpp: In file included from encrypt.cpp:13:
encrypt.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/wmmintrin.h:61:1: error: inlining failed in call to 'always_inline' '__m128i _mm_aesenc_si128(__m128i, __m128i)': target specific option mismatch
encrypt.cpp: 61 | _mm_aesenc_si128 (__m128i __X, __m128i __Y)
encrypt.cpp: | ^~~~~~~~~~~~~~~~
encrypt.cpp: encrypt.cpp:172:48: note: called from here
encrypt.cpp: 172 | acc5 = _mm_aesenc_si128(acc5, roundkey1);
encrypt.cpp: | ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~
encrypt.cpp: In file included from encrypt.cpp:13:
encrypt.cpp: /usr/lib/gcc/x86_64-linux-gnu/11/include/wmmintrin.h:61:1: error: inlining failed in call to 'always_inline' '__m128i _mm_aesenc_si128(__m128i, __m128i)': target specific option mismatch
encrypt.cpp: 61 | _mm_aesenc_si128 (__m128i __X, __m128i __Y)
encrypt.cpp: ...

Number of similar (compiler,implementation) pairs: 4, namely:

Compiler	Implementations
g++ -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aesni
g++ -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aesni
g++ -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aesni
g++ -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:aesni