Implementation notes: amd64, gemini, crypto_aead/aes256gcmv1

Computer: gemini
Microarchitecture: amd64; Goldmont Plus (706a8)
Architecture: amd64
CPU ID: GenuineIntel-000706a8-bfebfbff
SUPERCOP version: 20240425
Operation: crypto_aead
Primitive: aes256gcmv1

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
12052	18582 0 0	39952 812 1016	`dolbeau/aesenc-int`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
12175	18851 0 0	42512 812 1016	`dolbeau/aesenc-int`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
12220	16403 0 0	35878 804 1016	`dolbeau/aesenc-int`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
12928	19032 0 0	41000 780 1080	`dolbeau/aesenc-int`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14062	894 0 0	21301 884 1016	`T:opensslnew`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
14068	926 0 0	22616 876 1048	`T:opensslnew`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14080	894 0 0	23279 892 1016	`T:opensslnew`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
14181	894 0 0	25583 892 1016	`T:opensslnew`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
14183	889 0 0	21741 884 1016	`T:opensslnew`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
14184	926 0 0	23800 876 1048	`T:opensslnew`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14242	878 0 0	20547 852 1048	`T:opensslnew`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14315	6181 0 0	25027 756 1048	`dolbeau/aesenc-int`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14550	11875 0 0	32696 780 1080	`dolbeau/aesenc-int`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
14821	12034 0 0	32295 772 1080	`dolbeau/aesenc-int`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
15729	8723 0 0	28694 804 1016	`dolbeau/aesenc-int`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
16465	893 0 0	24199 892 1016	`T:opensslnew`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
16666	946 0 0	22055 868 1048	`T:opensslnew`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
33753	17742 3040 0	1418662 148468 10936	`T:cryptopp`	`clang++_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
34097	37791 3040 0	1728678 148444 11032	`T:cryptopp`	`g++_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
34448	14673 2456 0	1412950 148460 10936	`T:cryptopp`	`clang++_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
34839	34666 3040 0	1724636 148452 11032	`T:cryptopp`	`g++_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
35264	33547 3040 0	1722062 148452 11032	`T:cryptopp`	`g++_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
39334	23337 7904 0	1424514 152108 10936	`T:cryptopp`	`clang++_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
42514	17242 3040 0	1415882 148468 10936	`T:cryptopp`	`clang++_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
44221	14290 4072 0	1695886 149268 11000	`T:cryptopp`	`g++_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
2359325	3594 0 16	26688 812 1048	`ref`	`clang_-mcpu=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
2409143	2829 0 16	27304 812 1048	`ref`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
2414136	2829 0 16	25000 812 1048	`ref`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
3286163	3927 0 16	26664 780 1112	`ref`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
4582166	1633 0 16	21926 804 1048	`ref`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
20955379	2269 0 16	23824 780 1112	`ref`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
27163691	1837 0 16	22590 804 1048	`ref`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20240429	20240425
27979472	2160 0 16	23167 772 1112	`ref`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425
29519614	1766 0 16	21355 756 1080	`ref`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20240429	20240425

Compiler output

Implementation: dolbeau/aesenc-int
Security model: constbranchindex
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

aesenc-int.c: aesenc-int.c:665:16: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_aesenc_int_constbranchindex_encrypt' that is compiled without support for 'ssse3'
aesenc-int.c: __m128i Hv = _mm_shuffle_epi8(_mm_loadu_si128((const __m128i*)H), rev);
aesenc-int.c: ^
aesenc-int.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/aesenc-int

Compiler output

Implementation: dolbeau/vaesenc-int
Security model: constbranchindex
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

vaesenc-int.c: vaesenc-int.c:990:18: error: '__builtin_ia32_inserti32x4_256' needs target feature avx512vl
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx512vlintrin.h:7887:12: note: expanded from macro '_mm256_inserti32x4'
vaesenc-int.c: (__m256i)__builtin_ia32_inserti32x4_256((__v8si)(__m256i)(A), \
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:37: error: '__builtin_ia32_inserti32x4_256' needs target feature avx512vl
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx512vlintrin.h:7887:12: note: expanded from macro '_mm256_inserti32x4'
vaesenc-int.c: (__m256i)__builtin_ia32_inserti32x4_256((__v8si)(__m256i)(A), \
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:56: error: always_inline function '_mm256_undefined_si256' requires target feature 'avx', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_vaesenc_int_constbranchindex_encrypt' that is compiled without support for 'avx'
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:56: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
vaesenc-int.c: vaesenc-int.c:991:31: error: always_inline function '_mm256_broadcast_i32x4' requires target feature 'avx512vl', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_vaesenc_int_constbranchindex_encrypt' that is compiled without support for 'avx512vl'
vaesenc-int.c: __m256i H34v = mulvv2(H12v, _mm256_broadcast_i32x4(H2v));
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:991:31: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
vaesenc-int.c: vaesenc-int.c:991:18: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
vaesenc-int.c: __m256i H34v = mulvv2(H12v, _mm256_broadcast_i32x4(H2v));
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:992:10: error: '__builtin_ia32_inserti64x4' needs target feature avx512f
vaesenc-int.c: vhv0 = _mm512_inserti64x4(_mm512_inserti64x4(_mm512_undefined_epi32(), H12v, 1), H34v, 0);
vaesenc-int.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/vaesenc-int
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/vaesenc-int
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/vaesenc-int
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/vaesenc-int

Compiler output

Implementation: dolbeau/vaesenc-int
Security model: constbranchindex
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

vaesenc-int.c: vaesenc-int.c:985:16: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_vaesenc_int_constbranchindex_encrypt' that is compiled without support for 'ssse3'
vaesenc-int.c: __m128i Hv = _mm_shuffle_epi8(_mm_load_si128((const __m128i*)H), rev);
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:18: error: '__builtin_ia32_inserti32x4_256' needs target feature avx512vl
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx512vlintrin.h:7887:12: note: expanded from macro '_mm256_inserti32x4'
vaesenc-int.c: (__m256i)__builtin_ia32_inserti32x4_256((__v8si)(__m256i)(A), \
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:37: error: '__builtin_ia32_inserti32x4_256' needs target feature avx512vl
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: /usr/lib/llvm-11/lib/clang/11.0.1/include/avx512vlintrin.h:7887:12: note: expanded from macro '_mm256_inserti32x4'
vaesenc-int.c: (__m256i)__builtin_ia32_inserti32x4_256((__v8si)(__m256i)(A), \
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:56: error: always_inline function '_mm256_undefined_si256' requires target feature 'avx', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_vaesenc_int_constbranchindex_encrypt' that is compiled without support for 'avx'
vaesenc-int.c: __m256i H12v = _mm256_inserti32x4(_mm256_inserti32x4(_mm256_undefined_si256(), Hv, 1), H2v, 0);
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:990:56: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
vaesenc-int.c: vaesenc-int.c:991:31: error: always_inline function '_mm256_broadcast_i32x4' requires target feature 'avx512vl', but would be inlined into function 'crypto_aead_aes256gcmv1_dolbeau_vaesenc_int_constbranchindex_encrypt' that is compiled without support for 'avx512vl'
vaesenc-int.c: __m256i H34v = mulvv2(H12v, _mm256_broadcast_i32x4(H2v));
vaesenc-int.c: ^
vaesenc-int.c: vaesenc-int.c:991:31: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
vaesenc-int.c: vaesenc-int.c:991:18: warning: AVX vector argument of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI [-Wpsabi]
vaesenc-int.c: __m256i H34v = mulvv2(H12v, _mm256_broadcast_i32x4(H2v));
vaesenc-int.c: ...

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	dolbeau/vaesenc-int

Compiler output

Implementation: dolbeau/vaesenc-int
Security model: constbranchindex
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

vaesenc-int.c: vaesenc-int.c: In function 'mulvv2':
vaesenc-int.c: vaesenc-int.c:392:9: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
vaesenc-int.c: 392 | __m256i B) {
vaesenc-int.c: | ^~~~~~~
vaesenc-int.c: vaesenc-int.c: In function 'mulvv4':
vaesenc-int.c: vaesenc-int.c:435:9: warning: AVX512F vector return without AVX512F enabled changes the ABI [-Wpsabi]
vaesenc-int.c: 435 | __m512i B) {
vaesenc-int.c: | ^~~~~~~
vaesenc-int.c: vaesenc-int.c: In function 'mulvv2':
vaesenc-int.c: vaesenc-int.c:391:23: note: the ABI for passing parameters with 32-byte alignment has changed in GCC 4.6
vaesenc-int.c: 391 | static inline __m256i mulvv2(__m256i A,
vaesenc-int.c: | ^~~~~~
vaesenc-int.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:55,
vaesenc-int.c: from vaesenc-int.c:11:
vaesenc-int.c: vaesenc-int.c: In function 'aesni_key256_expand':
vaesenc-int.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avx512fintrin.h:4186:1: error: inlining failed in call to 'always_inline' '_mm512_broadcast_i32x4': target specific option mismatch
vaesenc-int.c: 4186 | _mm512_broadcast_i32x4 (__m128i __A)
vaesenc-int.c: | ^~~~~~~~~~~~~~~~~~~~~~
vaesenc-int.c: vaesenc-int.c:95:18: note: called from here
vaesenc-int.c: 95 | rkeys[idx++] = _mm512_broadcast_i32x4(temp0);
vaesenc-int.c: | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
vaesenc-int.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:55,
vaesenc-int.c: from vaesenc-int.c:11:
vaesenc-int.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avx512fintrin.h:4186:1: error: inlining failed in call to 'always_inline' '_mm512_broadcast_i32x4': target specific option mismatch
vaesenc-int.c: 4186 | _mm512_broadcast_i32x4 (__m128i __A)
vaesenc-int.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	dolbeau/vaesenc-int
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	dolbeau/vaesenc-int
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	dolbeau/vaesenc-int
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	dolbeau/vaesenc-int

Compiler output

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

decrypt.c: decrypt.c:13:18: error: variable has incomplete type 'EVP_CIPHER_CTX' (aka 'struct evp_cipher_ctx_st')
decrypt.c: EVP_CIPHER_CTX x;
decrypt.c: ^
decrypt.c: /usr/include/openssl/ossl_typ.h:90:16: note: forward declaration of 'struct evp_cipher_ctx_st'
decrypt.c: typedef struct evp_cipher_ctx_st EVP_CIPHER_CTX;
decrypt.c: ^
decrypt.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 5, namely:

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

Compiler output

Implementation: T:openssl
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

decrypt.c: decrypt.c: In function 'crypto_aead_aes256gcmv1_openssl_timingleaks_decrypt':
decrypt.c: decrypt.c:13:18: error: storage size of 'x' isn't known
decrypt.c: 13 | EVP_CIPHER_CTX x;
decrypt.c: | ^

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:openssl
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:openssl
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:openssl
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:openssl