Implementation notes: aarch64, supercoplxc, crypto_sign/picnicl3fs

Computer: supercoplxc
Architecture: aarch64
CPU ID: 410fd034
SUPERCOP version: 20190816
Operation: crypto_sign
Primitive: picnicl3fs
TimeObject sizeTest sizeImplementationCompilerBenchmark dateSUPERCOP version
221928800206493 2840 16248245 3832 1624optimizedct/cgcc_-funroll-loops_-O3_-fomit-frame-pointer2019102920190816
226157840205801 2840 16247493 3832 1624optimizedct/cgcc_-funroll-loops_-fno-schedule-insns_-O3_-fomit-frame-pointer2019102920190816
227589600193789 2840 16235125 3832 1600optimizedct/cgcc_-funroll-loops_-O2_-fomit-frame-pointer2019102920190816
229712560193569 2840 16234957 3832 1600optimizedct/cgcc_-funroll-loops_-fno-schedule-insns_-O2_-fomit-frame-pointer2019102920190816
237663680173642 2056 16212064 2912 1624optimizedct/cclang_-O3_-fwrapv_-mavx2_-fomit-frame-pointer_-Qunused-arguments2019102920190816
238173760173066 2056 16211440 2912 1624optimizedct/cclang_-O3_-fomit-frame-pointer_-Qunused-arguments2019102920190816
238287680173642 2056 16212064 2912 1624optimizedct/cclang_-O3_-fwrapv_-mavx_-maes_-mpclmul_-fomit-frame-pointer_-Qunused-arguments2019102920190816
238974960173642 2056 16212064 2912 1624optimizedct/cclang_-O3_-fwrapv_-mavx_-fomit-frame-pointer_-Qunused-arguments2019102920190816
240476400171453 2840 16211861 3832 1624optimizedct/cgcc_-O3_-fomit-frame-pointer2019102920190816
242998800175426 2056 16213856 2912 1624optimizedct/cclang_-mcpu=native_-mfpu=neon_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments2019102920190816
243914480171025 2840 16215453 3832 1624optimizedct/cgcc_-fno-schedule-insns_-O3_-fomit-frame-pointer2019102920190816
281713920161025 2840 16200325 3832 1600optimizedct/cgcc_-O2_-fomit-frame-pointer2019102920190816
296118640160873 2840 16200077 3832 1600optimizedct/cgcc_-fno-schedule-insns_-O2_-fomit-frame-pointer2019102920190816
313522960150982 2840 16189459 3816 1584optimizedct/cgcc_-funroll-loops_-fno-schedule-insns_-Os_-fomit-frame-pointer2019102920190816
315408240150982 2840 16189459 3816 1584optimizedct/cgcc_-funroll-loops_-Os_-fomit-frame-pointer2019102920190816
316048560193617 2840 16236285 3832 1600optimizedct/cgcc_-funroll-loops_-O_-fomit-frame-pointer2019102920190816
316107760193617 2840 16236285 3832 1600optimizedct/cgcc_-funroll-loops_-fno-schedule-insns_-O_-fomit-frame-pointer2019102920190816
320370800151014 2840 16189443 3816 1584optimizedct/cgcc_-fno-schedule-insns_-Os_-fomit-frame-pointer2019102920190816
322420640151014 2840 16189443 3816 1584optimizedct/cgcc_-Os_-fomit-frame-pointer2019102920190816
324392160162937 2840 16202373 3832 1600optimizedct/cgcc_-O_-fomit-frame-pointer2019102920190816
334842960162937 2840 16202373 3832 1600optimizedct/cgcc_-fno-schedule-insns_-O_-fomit-frame-pointer2019102920190816
896011840203226 2840 16247317 3816 1600optimizedct/cgcc_-funroll-loops2019102920190816
897320560203226 2840 16247317 3816 1600optimizedct/ccc2019102920190816
918438400203226 2840 16247317 3816 1600optimizedct/cgcc2019102920190816
21336507201056886 4 01094673 984 1592refgcc_-fno-schedule-insns_-O3_-fomit-frame-pointer2019102920190816
21456109601068232 4 01107033 984 1576refgcc_-funroll-loops_-O2_-fomit-frame-pointer2019102920190816
21655768801091782 4 01130969 984 1592refgcc_-funroll-loops_-O3_-fomit-frame-pointer2019102920190816
21767460001056974 4 01094801 984 1592refgcc_-O3_-fomit-frame-pointer2019102920190816
22275631201044384 4 01081129 984 1576refgcc_-O2_-fomit-frame-pointer2019102920190816
22734709601067980 4 01106841 984 1576refgcc_-funroll-loops_-fno-schedule-insns_-O2_-fomit-frame-pointer2019102920190816
22836241601091362 4 01130505 984 1592refgcc_-funroll-loops_-fno-schedule-insns_-O3_-fomit-frame-pointer2019102920190816
23147834401044264 4 01080937 984 1576refgcc_-fno-schedule-insns_-O2_-fomit-frame-pointer2019102920190816
24866048001044173 4 01081848 872 1608refclang_-O3_-fomit-frame-pointer_-Qunused-arguments2019102920190816
25145998401044569 4 01082328 872 1608refclang_-mcpu=native_-mfpu=neon_-O3_-fomit-frame-pointer_-fwrapv_-Qunused-arguments2019102920190816
25205110401044181 4 01081912 872 1608refclang_-O3_-fwrapv_-mavx_-fomit-frame-pointer_-Qunused-arguments2019102920190816
25668739201044181 4 01081912 872 1608refclang_-O3_-fwrapv_-mavx_-maes_-mpclmul_-fomit-frame-pointer_-Qunused-arguments2019102920190816
25757200801044181 4 01081912 872 1608refclang_-O3_-fwrapv_-mavx2_-fomit-frame-pointer_-Qunused-arguments2019102920190816
29382438401041356 4 01078185 984 1576refgcc_-fno-schedule-insns_-O_-fomit-frame-pointer2019102920190816
29523528001041356 4 01078185 984 1576refgcc_-O_-fomit-frame-pointer2019102920190816
29703557601037317 4 01073047 968 1568refgcc_-fno-schedule-insns_-Os_-fomit-frame-pointer2019102920190816
29895097601056620 4 01096681 984 1576refgcc_-funroll-loops_-O_-fomit-frame-pointer2019102920190816
30651496001037317 4 01073047 968 1568refgcc_-Os_-fomit-frame-pointer2019102920190816
30858625601037357 4 01073135 968 1568refgcc_-funroll-loops_-Os_-fomit-frame-pointer2019102920190816
30866358401037357 4 01073135 968 1568refgcc_-funroll-loops_-fno-schedule-insns_-Os_-fomit-frame-pointer2019102920190816
30927084801056620 4 01096681 984 1576refgcc_-funroll-loops_-fno-schedule-insns_-O_-fomit-frame-pointer2019102920190816

Test failure

Implementation: ref
Security model: unknown
Compiler: cc
error 142
Alarm clock

Number of similar (compiler,implementation) pairs: 3, namely:
CompilerImplementations
cc ref
gcc ref
gcc -funroll-loops ref

Compiler output

Implementation: optimizedct/neon
Security model: unknown
Compiler: cc
picnic2_simulate_mul.c: picnic2_simulate_mul.c: In function 'transpose_64_64_s128':
picnic2_simulate_mul.c: picnic2_simulate_mul.c:316:24: warning: implicit declaration of function '_mm_set1_epi64x' [-Wimplicit-function-declaration]
picnic2_simulate_mul.c: word128 mask = _mm_set1_epi64x(TRANSPOSE_MASKS64[i]);
picnic2_simulate_mul.c: ^~~~~~~~~~~~~~~
picnic2_simulate_mul.c: picnic2_simulate_mul.c:316:24: error: incompatible types when initializing type 'word128' using type 'int'
picnic2_simulate_mul.c: In file included from mzd_additional.h:24,
picnic2_simulate_mul.c: from lowmc_pars.h:15,
picnic2_simulate_mul.c: from lowmc.h:13,
picnic2_simulate_mul.c: from picnic_impl.h:13,
picnic2_simulate_mul.c: from picnic2_impl.h:18,
picnic2_simulate_mul.c: from picnic2_types.h:16,
picnic2_simulate_mul.c: from picnic2_simulate_mul.h:13,
picnic2_simulate_mul.c: from picnic2_simulate_mul.c:14:
picnic2_simulate_mul.c: picnic2_simulate_mul.c:317:40: error: incompatible type for argument 2 of 'veorq_u64'
picnic2_simulate_mul.c: word128 inv_mask = mm128_xor(mask, _mm_set1_epi64x(UINT64_C(0xFFFFFFFFFFFFFFFF)));
picnic2_simulate_mul.c: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
picnic2_simulate_mul.c: simd.h:103:38: note: in definition of macro 'mm128_xor'
picnic2_simulate_mul.c: #define mm128_xor(l, r) veorq_u64(l, r)
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: In file included from simd.h:21,
picnic2_simulate_mul.c: from mzd_additional.h:24,
picnic2_simulate_mul.c: from lowmc_pars.h:15,
picnic2_simulate_mul.c: from lowmc.h:13,
picnic2_simulate_mul.c: from picnic_impl.h:13,
picnic2_simulate_mul.c: from picnic2_impl.h:18,
picnic2_simulate_mul.c: ...

Number of similar (compiler,implementation) pairs: 19, namely:
CompilerImplementations
cc optimizedct/neon
gcc optimizedct/neon
gcc -O2 -fomit-frame-pointer optimizedct/neon
gcc -O3 -fomit-frame-pointer optimizedct/neon
gcc -O -fomit-frame-pointer optimizedct/neon
gcc -Os -fomit-frame-pointer optimizedct/neon
gcc -fno-schedule-insns -O2 -fomit-frame-pointer optimizedct/neon
gcc -fno-schedule-insns -O3 -fomit-frame-pointer optimizedct/neon
gcc -fno-schedule-insns -O -fomit-frame-pointer optimizedct/neon
gcc -fno-schedule-insns -Os -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops optimizedct/neon
gcc -funroll-loops -O2 -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -O3 -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -O -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -Os -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -fno-schedule-insns -O2 -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -fno-schedule-insns -O3 -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -fno-schedule-insns -O -fomit-frame-pointer optimizedct/neon
gcc -funroll-loops -fno-schedule-insns -Os -fomit-frame-pointer optimizedct/neon

Compiler output

Implementation: optimizedct/neon
Security model: unknown
Compiler: clang -O3 -fomit-frame-pointer -Qunused-arguments
picnic2_simulate_mul.c: picnic2_simulate_mul.c:316:24: warning: implicit declaration of function '_mm_set1_epi64x' is invalid in C99 [-Wimplicit-function-declaration]
picnic2_simulate_mul.c: word128 mask = _mm_set1_epi64x(TRANSPOSE_MASKS64[i]);
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: picnic2_simulate_mul.c:316:13: error: initializing 'word128' (aka 'uint64x2_t') with an expression of incompatible type 'int'
picnic2_simulate_mul.c: word128 mask = _mm_set1_epi64x(TRANSPOSE_MASKS64[i]);
picnic2_simulate_mul.c: ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
picnic2_simulate_mul.c: picnic2_simulate_mul.c:329:57: warning: implicit declaration of function '_mm_srli_epi64' is invalid in C99 [-Wimplicit-function-declaration]
picnic2_simulate_mul.c: out128[i1 / 2] = mm128_xor(mm128_and(t1, mask), _mm_srli_epi64(mm128_and(t2, mask), width));
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: picnic2_simulate_mul.c:331:48: warning: implicit declaration of function '_mm_slli_epi64' is invalid in C99 [-Wimplicit-function-declaration]
picnic2_simulate_mul.c: mm128_xor(mm128_and(t2, inv_mask), _mm_slli_epi64(mm128_and(t1, inv_mask), width));
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: picnic2_simulate_mul.c:331:48: error: passing 'int' to parameter of incompatible type 'uint64x2_t' (vector of 2 'uint64_t' values)
picnic2_simulate_mul.c: mm128_xor(mm128_and(t2, inv_mask), _mm_slli_epi64(mm128_and(t1, inv_mask), width));
picnic2_simulate_mul.c: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
picnic2_simulate_mul.c: ./simd.h:103:38: note: expanded from macro 'mm128_xor'
picnic2_simulate_mul.c: #define mm128_xor(l, r) veorq_u64(l, r)
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: /usr/lib/llvm-7/lib/clang/7.0.1/include/arm_neon.h:5421:55: note: passing argument to parameter '__p1' here
picnic2_simulate_mul.c: __ai uint64x2_t veorq_u64(uint64x2_t __p0, uint64x2_t __p1) {
picnic2_simulate_mul.c: ^
picnic2_simulate_mul.c: 3 warnings and 2 errors generated.

Number of similar (compiler,implementation) pairs: 5, namely:
CompilerImplementations
clang -O3 -fomit-frame-pointer -Qunused-arguments optimizedct/neon
clang -O3 -fwrapv -mavx2 -fomit-frame-pointer -Qunused-arguments optimizedct/neon
clang -O3 -fwrapv -mavx -fomit-frame-pointer -Qunused-arguments optimizedct/neon
clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer -Qunused-arguments optimizedct/neon
clang -mcpu=native -mfpu=neon -O3 -fomit-frame-pointer -fwrapv -Qunused-arguments optimizedct/neon