Implementation notes: armeabi, cubox, crypto_hash/blake256

Computer: cubox
Architecture: armeabi
CPU ID: unknown CPU ID
SUPERCOP version: 20161026
Operation: crypto_hash
Primitive: blake256

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
51161	`arm11`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
51184	`arm11`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
51232	`arm11`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
51264	`arm11`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
51496	`arm11`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
51496	`arm11`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
53220	`arm_thumb2`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
53226	`arm_thumb2`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
53248	`arm_thumb2`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
53296	`arm_thumb2`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
53512	`arm_thumb2`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
53516	`arm_thumb2`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
56162	`bswap`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
57481	`sphlib`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
57518	`sphlib`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
57916	`regs`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
58478	`sphlib`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
58555	`sphlib`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
59178	`bswap`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
61588	`sphlib`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
61741	`regs`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
63569	`bswap`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
63608	`bswap`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
63851	`sphlib`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
64984	`bswap`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
65216	`bswap`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
66379	`regs`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
66435	`regs`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
68360	`regs`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
68512	`regs`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
85218	`sphlib-small`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
88568	`sphlib-small`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
98092	`ref`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
102732	`ref`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
107348	`ref`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
112074	`sphlib-small`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
112494	`sphlib-small`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
119160	`sphlib-small`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
119276	`ref`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
120229	`sphlib-small`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
120417	`ref`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
126324	`ref`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026

Test failure

Implementation: crypto_hash/blake256/vect128-neon
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

error 111

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	vect128-neon
gcc -funroll-loops -mcpu=marvell-pj4 -O3	vect128-neon
gcc -funroll-loops -mcpu=marvell-pj4 -Os	vect128-neon
gcc -mcpu=marvell-pj4 -O2	vect128-neon
gcc -mcpu=marvell-pj4 -O3	vect128-neon
gcc -mcpu=marvell-pj4 -Os	vect128-neon

Compiler output

Implementation: crypto_hash/blake256/vect128-mmxhack
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

vector.c: In file included from vector.c:7:0:
vector.c: vector.h:968:2: 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:982:3: error: unknown type name 'v16'
vector.c: v16 v16;
vector.c: ^~~
vector.c: vector.h:987:3: error: unknown type name 'v8'
vector.c: v8 v8;
vector.c: ^~
vector.c: ...
vector.c: static const union cv8 v0 = CV(0);
vector.c: ^~
vector.c: In file included from vector.c:140:0:
vector.c: /usr/lib/gcc/armv7l-unknown-linux-gnueabihf/6.2.1/include/mmintrin.h:28:2: error: #error mmintrin.h included without enabling WMMX/WMMX2 instructions (e.g. -march=iwmmxt or -march=iwmmxt2)
vector.c: #error mmintrin.h included without enabling WMMX/WMMX2 instructions (e.g. -march=iwmmxt or -march=iwmmxt2)
vector.c: ^~~~~
vector.c: vector.c:141:23: fatal error: emmintrin.h: No such file or directory
vector.c: #include gt;
vector.c: ^
vector.c: compilation terminated.

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	vect128-mmxhack
gcc -funroll-loops -mcpu=marvell-pj4 -O3	vect128-mmxhack
gcc -funroll-loops -mcpu=marvell-pj4 -Os	vect128-mmxhack
gcc -mcpu=marvell-pj4 -O2	vect128-mmxhack
gcc -mcpu=marvell-pj4 -O3	vect128-mmxhack
gcc -mcpu=marvell-pj4 -Os	vect128-mmxhack

Compiler output

Implementation: crypto_hash/blake256/vect128
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

vector.c: In file included from vector.c:7:0:
vector.c: vector.h:968:2: 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:982:3: error: unknown type name 'v16'
vector.c: v16 v16;
vector.c: ^~~
vector.c: vector.h:987:3: error: unknown type name 'v8'
vector.c: v8 v8;
vector.c: ^~
vector.c: ...
vector.c: vector.c: In function 'blake256_update':
vector.c: vector.c:887:15: error: 'v32' undeclared (first use in this function)
vector.c: round256((v32*)S->gt;buf, (v32*)S->gt;h,
vector.c: ^~~
vector.c: vector.c:887:19: error: expected expression before ')' token
vector.c: round256((v32*)S->gt;buf, (v32*)S->gt;h,
vector.c: ^
vector.c: vector.c:902:19: error: expected expression before ')' token
vector.c: round256((v32*)data, (v32*)S->gt;h,
vector.c: ^

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	vect128
gcc -funroll-loops -mcpu=marvell-pj4 -O3	vect128
gcc -funroll-loops -mcpu=marvell-pj4 -Os	vect128
gcc -mcpu=marvell-pj4 -O2	vect128
gcc -mcpu=marvell-pj4 -O3	vect128
gcc -mcpu=marvell-pj4 -Os	vect128