Implementation notes: amd64, gemini, crypto_stream/simon6496ctr

Computer: gemini
Microarchitecture: amd64; Goldmont Plus (706a8)
Architecture: amd64
CPU ID: GenuineIntel-000706a8-bfebfbff
SUPERCOP version: 20231107
Operation: crypto_stream
Primitive: simon6496ctr

Time	Object size	Test size	Implementation	Compiler	Benchmark date	SUPERCOP version
7735	103552 0 0	118232 780 1016	`T:sse4`	`gcc_-march=native_-mtune=native_-O3_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230523	20221122
8363	102852 0 0	116760 812 952	`T:sse4`	`clang_-march=native_-O2_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230523	20221122
8370	103316 0 0	119528 812 952	`T:sse4`	`clang_-march=native_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230523	20221122
12205	107975 0 0	120935 772 1016	`T:sse4`	`gcc_-march=native_-mtune=native_-O_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230523	20221122
12284	103187 0 0	116648 780 1016	`T:sse4`	`gcc_-march=native_-mtune=native_-O2_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230523	20221122
12724	94462 0 0	106235 756 984	`T:sse4`	`gcc_-march=native_-mtune=native_-Os_-fomit-frame-pointer_-fwrapv_-fPIC_-fPIE`	20230523	20221122
13065	110679 0 0	123222 804 952	`T:sse4`	`clang_-march=native_-O_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230523	20221122
13321	106533 0 0	118590 804 952	`T:sse4`	`clang_-march=native_-Os_-fomit-frame-pointer_-fwrapv_-Qunused-arguments_-fPIC_-fPIE`	20230523	20221122

Compiler output

Implementation: T:avx2
Security model: timingleaks
Compiler: clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

stream.c: stream.c:148:3: error: always_inline function '_mm256_set_epi32' requires target feature 'avx', but would be inlined into function 'Encrypt' that is compiled without support for 'avx'
stream.c: SET1(X[0],nonce[1]); SET8(Y[0],nonce[0]);
stream.c: ^
stream.c: ./Intrinsics_AVX2_64block.h:25:22: note: expanded from macro 'SET1'
stream.c: #define SET1(X,c) (X=SET(c,c,c,c,c,c,c,c))
stream.c: ^
stream.c: ./Intrinsics_AVX2_64block.h:24:13: note: expanded from macro 'SET'
stream.c: #define SET _mm256_set_epi32
stream.c: ^
stream.c: stream.c:148:3: error: AVX vector return of type '__m256i' (vector of 4 'long long' values) without 'avx' enabled changes the ABI
stream.c: ./Intrinsics_AVX2_64block.h:25:22: note: expanded from macro 'SET1'
stream.c: #define SET1(X,c) (X=SET(c,c,c,c,c,c,c,c))
stream.c: ^
stream.c: ./Intrinsics_AVX2_64block.h:24:13: note: expanded from macro 'SET'
stream.c: #define SET _mm256_set_epi32
stream.c: ^
stream.c: stream.c:148:24: error: always_inline function '_mm256_set_epi32' requires target feature 'avx', but would be inlined into function 'Encrypt' that is compiled without support for 'avx'
stream.c: SET1(X[0],nonce[1]); SET8(Y[0],nonce[0]);
stream.c: ^
stream.c: ./Intrinsics_AVX2_64block.h:26:22: note: expanded from macro 'SET8'
stream.c: #define SET8(X,c) (X=SET(c,c,c,c,c,c,c,c), X=ADD(X,_q))
stream.c: ^
stream.c: ./Intrinsics_AVX2_64block.h:24:13: note: expanded from macro 'SET'
stream.c: #define SET _mm256_set_epi32
stream.c: ^
stream.c: ...

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

Compiler	Implementations
clang -march=native -O2 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2
clang -march=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2
clang -march=native -O -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2
clang -march=native -Os -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:avx2

Compiler output

Implementation: T:avx2
Security model: timingleaks
Compiler: gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE

stream.c: In file included from Simon6496AVX2.h:17,
stream.c: from stream.c:20:
stream.c: stream.c: In function 'Encrypt':
stream.c: Intrinsics_AVX2_64block.h:25:21: warning: AVX vector return without AVX enabled changes the ABI [-Wpsabi]
stream.c: 25 | #define SET1(X,c) (X=SET(c,c,c,c,c,c,c,c))
stream.c: | ~~^~~~~~~~~~~~~~~~~~~~~~
stream.c: stream.c:148:3: note: in expansion of macro 'SET1'
stream.c: 148 | SET1(X[0],nonce[1]); SET8(Y[0],nonce[0]);
stream.c: | ^~~~
stream.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
stream.c: from Intrinsics_AVX2_64block.h:5,
stream.c: from Simon6496AVX2.h:17,
stream.c: from stream.c:20:
stream.c: stream.c: In function 'Transpose':
stream.c: /usr/lib/gcc/x86_64-linux-gnu/10/include/avx2intrin.h:913:1: error: inlining failed in call to 'always_inline' '_mm256_xor_si256': target specific option mismatch
stream.c: 913 | _mm256_xor_si256 (__m256i __A, __m256i __B)
stream.c: | ^~~~~~~~~~~~~~~~
stream.c: In file included from Simon6496AVX2.h:17,
stream.c: from stream.c:20:
stream.c: Intrinsics_AVX2_64block.h:15:13: note: called from here
stream.c: 15 | #define XOR _mm256_xor_si256
stream.c: stream.c:462:10: note: in expansion of macro 'XOR'
stream.c: 462 | M[6] = XOR(M[6], W[3]);
stream.c: | ^~~
stream.c: In file included from /usr/lib/gcc/x86_64-linux-gnu/10/include/immintrin.h:53,
stream.c: ...

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

Compiler	Implementations
gcc -march=native -mtune=native -O2 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2
gcc -march=native -mtune=native -O3 -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2
gcc -march=native -mtune=native -O -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2
gcc -march=native -mtune=native -Os -fomit-frame-pointer -fwrapv -fPIC -fPIE	T:avx2

Compiler output

Implementation: T:sse4
Security model: timingleaks
Compiler: clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE

stream.c: stream.c:341:3: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ExpandKeyBS' that is compiled without support for 'ssse3'
stream.c: EKBS(rk);
stream.c: ^
stream.c: ./Simon6496SSE4.h:59:19: note: expanded from macro 'EKBS'
stream.c: #define EKBS(rk) (RKBS(rk,3,_D), RKBS(rk,4,_C), RKBS(rk,5,_D), RKBS(rk,6,_C), RKBS(rk,7,_D), RKBS(rk,8,_D), \
stream.c: ^
stream.c: ./Simon6496SSE4.h:50:52: note: expanded from macro 'RKBS'
stream.c: #define RKBS(rk,r,_V) (rk[r][7]= _D ^ rk[r-3][7] ^ ROR8(rk[r-1][2] ^ rk[r-1][3]), \
stream.c: ^
stream.c: ./Intrinsics_SSE4_64block.h:40:19: note: expanded from macro 'ROR8'
stream.c: #define ROR8(X) (SHFL(X,R8))
stream.c: ^
stream.c: ./Intrinsics_SSE4_64block.h:35:14: note: expanded from macro 'SHFL'
stream.c: #define SHFL _mm_shuffle_epi8
stream.c: ^
stream.c: stream.c:341:3: error: always_inline function '_mm_shuffle_epi8' requires target feature 'ssse3', but would be inlined into function 'ExpandKeyBS' that is compiled without support for 'ssse3'
stream.c: ./Simon6496SSE4.h:59:19: note: expanded from macro 'EKBS'
stream.c: #define EKBS(rk) (RKBS(rk,3,_D), RKBS(rk,4,_C), RKBS(rk,5,_D), RKBS(rk,6,_C), RKBS(rk,7,_D), RKBS(rk,8,_D), \
stream.c: ^
stream.c: ./Simon6496SSE4.h:51:52: note: expanded from macro 'RKBS'
stream.c: rk[r][6]= _D ^ rk[r-3][6] ^ ROR8(rk[r-1][1] ^ rk[r-1][2]), \
stream.c: ^
stream.c: ./Intrinsics_SSE4_64block.h:40:19: note: expanded from macro 'ROR8'
stream.c: #define ROR8(X) (SHFL(X,R8))
stream.c: ^
stream.c: ...

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

Compiler	Implementations
clang -mcpu=native -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments -fPIC -fPIE	T:sse4