Implementation notes: armeabi, cubie2, crypto_hash/simd256

Computer: cubie2
Architecture: armeabi
CPU ID: unknown CPU ID
SUPERCOP version: 20161026
Operation: crypto_hash
Primitive: simd256

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
192895	`sphlib`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
196299	`sphlib`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
199018	`sphlib`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
214734	`sphlib`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
218272	`sphlib`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
226591	`opt`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
227029	`opt`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
229854	`opt`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
230826	`sphlib-small`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
232173	`sphlib-small`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
238484	`sphlib-small`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
240905	`sphlib`	`gcc -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
251763	`sphlib-small`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
273588	`opt`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
277239	`sphlib-small`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
301080	`sphlib-small`	`gcc -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
467253	`opt`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
471890	`opt`	`gcc -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
3310514	`ref`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
3380930	`ref`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
3412202	`ref`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O3`	20161216	20161026
3498659	`ref`	`gcc -mcpu=native -mfpu=neon-vfpv4 -O2`	20161216	20161026
6803741	`ref`	`gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026
6818476	`ref`	`gcc -mcpu=native -mfpu=neon-vfpv4 -Os`	20161216	20161026

Compiler output

Implementation: crypto_hash/simd256/vect128
Compiler: gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2

vector.c: In file included from vector.c:5:0:
vector.c: vector.h:252:2: error: #error "I don't know how to vectorize on this architecture."
vector.c: #error "I don't know how to vectorize on this architecture."
vector.c: ^~~~~
vector.c: vector.h:344:25: error: array type has incomplete element type 'union cv'
vector.c: static const union cv FFT64_Twiddle[] = {
vector.c: ^~~~~~~~~~~~~
vector.c: vector.h:355:25: error: array type has incomplete element type 'union cv'
vector.c: static const union cv FFT128_Twiddle[] = {
vector.c: ^~~~~~~~~~~~~~
vector.c: ...
vector.c: ^~~~
vector.c: vector.c:25:23: error: storage size of 'V255' isn't known
vector.c: static const union cv V255 = CV(255);
vector.c: ^~~~
vector.c: vector.c:26:23: error: storage size of 'V257' isn't known
vector.c: static const union cv V257 = CV(257);
vector.c: ^~~~
vector.c: vector.c:27:25: error: storage size of 'V0' isn't known
vector.c: static const union cv8 V0 = CV(0);
vector.c: ^~

Number of similar (compiler,implementation) pairs: 6, namely:

Compiler	Implementations
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2	vect128
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3	vect128
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os	vect128
gcc -mcpu=native -mfpu=neon-vfpv4 -O2	vect128
gcc -mcpu=native -mfpu=neon-vfpv4 -O3	vect128
gcc -mcpu=native -mfpu=neon-vfpv4 -Os	vect128

Compiler output

Implementation: crypto_hash/simd256/vectarm
Compiler: gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2

vector_arm.c: vector_arm.c: In function 'fft64':
vector_arm.c: vector_arm.c:182:5: error: impossible constraint in 'asm'
vector_arm.c: __asm__ (
vector_arm.c: ^~~~~~~
vector_arm.c: vector_arm.c:217:5: error: impossible constraint in 'asm'
vector_arm.c: __asm__ (
vector_arm.c: ^~~~~~~

Number of similar (compiler,implementation) pairs: 6, namely:

Compiler	Implementations
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O2	vectarm
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -O3	vectarm
gcc -funroll-loops -mcpu=native -mfpu=neon-vfpv4 -Os	vectarm
gcc -mcpu=native -mfpu=neon-vfpv4 -O2	vectarm
gcc -mcpu=native -mfpu=neon-vfpv4 -O3	vectarm
gcc -mcpu=native -mfpu=neon-vfpv4 -Os	vectarm