Implementation notes: armeabi, cubox, crypto_hash/cubehash512

Computer: cubox
Architecture: armeabi
CPU ID: unknown CPU ID
SUPERCOP version: 20161026
Operation: crypto_hash
Primitive: cubehash512

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
106520	`unrolled3`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
110492	`arm`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
110496	`arm`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
110537	`arm`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
110744	`arm`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
110748	`arm`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
110756	`arm`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
118004	`unrolled5`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
120640	`unrolled3`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
136223	`unrolled5`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
139508	`unrolled5`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
142436	`unrolled4`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
142898	`unrolled4`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
142987	`unrolled3`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
143212	`unrolled3`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
146181	`unrolled5`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
149692	`unrolled5`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
151345	`unrolled4`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
151436	`unrolled3`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
154476	`unrolled4`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
155144	`unrolled5`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
159418	`unrolled4`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
159992	`unrolled3`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
160436	`unrolled4`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
164685	`unrolled`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
171781	`simple`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
175604	`unrolled`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
175979	`unrolled`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
176455	`unrolled2`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
179593	`simple`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161208	20161026
187448	`unrolled2`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
187787	`unrolled2`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161208	20161026
188316	`unrolled`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
188844	`unrolled2`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
190992	`simple`	`gcc -mcpu=marvell-pj4 -O3`	20161208	20161026
192056	`unrolled`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
198656	`unrolled2`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
201636	`unrolled`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
206692	`unrolled2`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026
959185	`simple`	`gcc -mcpu=marvell-pj4 -O2`	20161208	20161026
979175	`simple`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161208	20161026
983752	`simple`	`gcc -mcpu=marvell-pj4 -Os`	20161208	20161026

Compiler output

Implementation: crypto_hash/cubehash512/armneon
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

hash.c: In file included from hash.c:2:0:
hash.c: hash.c: In function 'crypto_hash_cubehash512_armneon':
hash.c: /usr/lib/gcc/armv7l-unknown-linux-gnueabihf/6.2.1/include/arm_neon.h:9035:1: error: inlining failed in call to always_inline 'vld1q_u32': target specific option mismatch
hash.c: vld1q_u32 (const uint32_t * __a)
hash.c: ^~~~~~~~~
hash.c: hash.c:40:6: note: called from here
hash.c: x7 = vld1q_u32(constants + 28);
hash.c: ~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~
hash.c: In file included from hash.c:2:0:
hash.c: /usr/lib/gcc/armv7l-unknown-linux-gnueabihf/6.2.1/include/arm_neon.h:9035:1: error: inlining failed in call to always_inline 'vld1q_u32': target specific option mismatch
hash.c: ...
hash.c: hash.c:152:3: note: called from here
hash.c: vst1q_u32((uint32_t *) (out + 16),x1);
hash.c: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
hash.c: In file included from hash.c:2:0:
hash.c: /usr/lib/gcc/armv7l-unknown-linux-gnueabihf/6.2.1/include/arm_neon.h:9535:1: error: inlining failed in call to always_inline 'vst1q_u32': target specific option mismatch
hash.c: vst1q_u32 (uint32_t * __a, uint32x4_t __b)
hash.c: ^~~~~~~~~
hash.c: hash.c:151:3: note: called from here
hash.c: vst1q_u32((uint32_t *) out,x0);
hash.c: ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

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

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