Implementation notes: armeabi, cubox, crypto_aead/lunarkeyakv2

Computer: cubox
Architecture: armeabi
CPU ID: unknown CPU ID
SUPERCOP version: 20161026
Operation: crypto_aead
Primitive: lunarkeyakv2

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
570184	`generic32`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
570407	`generic32lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
594498	`generic32`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
594660	`generic32lc`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
663917	`generic64`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
670323	`generic64lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
707614	`generic64`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
711868	`generic64lc`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
815378	`generic32lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
816109	`generic32`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
819123	`generic32`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
819263	`generic32lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
844922	`generic32lc`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
845918	`generic32`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
854054	`generic32lc`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
854918	`generic32`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
979798	`reference`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
1039447	`reference`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
1050019	`reference`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
1153058	`compact`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
1223738	`compact`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
1242176	`generic64`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
1244068	`generic64`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
1250340	`compact`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
1278520	`generic64lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
1291976	`generic64lc`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
1366130	`generic64`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
1368906	`generic64`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
1405996	`generic64lc`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
1426422	`generic64lc`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
2035750	`compact`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
2581887	`compact`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
2948176	`compact`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
3254624	`reference`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
6174187	`reference32bits`	`gcc -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
6824460	`reference`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
6852383	`reference`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
7668985	`reference32bits`	`gcc -mcpu=marvell-pj4 -O3`	20161215	20161026
10678043	`reference32bits`	`gcc -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
12550972	`refnew`	`g++ -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
12618272	`ref`	`g++ -mcpu=marvell-pj4 -O3`	20161215	20161026
12647464	`ref`	`g++ -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
12682524	`refnew`	`g++ -mcpu=marvell-pj4 -O3`	20161215	20161026
12727949	`ref`	`g++ -funroll-loops -mcpu=marvell-pj4 -O3`	20161215	20161026
12783575	`refnew`	`g++ -funroll-loops -mcpu=marvell-pj4 -O2`	20161215	20161026
13085989	`reference32bits`	`gcc -mcpu=marvell-pj4 -O2`	20161215	20161026
14069940	`ref`	`g++ -mcpu=marvell-pj4 -O2`	20161215	20161026
14241490	`refnew`	`g++ -mcpu=marvell-pj4 -O2`	20161215	20161026
16367029	`reference32bits`	`gcc -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
16428133	`reference32bits`	`gcc -mcpu=marvell-pj4 -Os`	20161215	20161026
50246169	`refnew`	`g++ -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
50280274	`refnew`	`g++ -mcpu=marvell-pj4 -Os`	20161215	20161026
50552847	`ref`	`g++ -funroll-loops -mcpu=marvell-pj4 -Os`	20161215	20161026
50927338	`ref`	`g++ -mcpu=marvell-pj4 -Os`	20161215	20161026

Test failure

Implementation: crypto_aead/lunarkeyakv2/ARMv6M
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	ARMv6M
gcc -funroll-loops -mcpu=marvell-pj4 -O3	ARMv6M
gcc -funroll-loops -mcpu=marvell-pj4 -Os	ARMv6M
gcc -mcpu=marvell-pj4 -O2	ARMv6M
gcc -mcpu=marvell-pj4 -O3	ARMv6M
gcc -mcpu=marvell-pj4 -Os	ARMv6M

Compiler output

Implementation: crypto_aead/lunarkeyakv2/ref
Compiler: g++ -funroll-loops -mcpu=marvell-pj4 -O2

Keyakv2.cpp: In file included from Keyakv2.h:21:0,
Keyakv2.cpp: from Keyakv2.cpp:16:
Keyakv2.cpp: Motorist.h:31:5: warning: 'templategt; class std::auto_ptr' is deprecated [-Wdeprecated-declarations]
Keyakv2.cpp: auto_ptrgt; state;
Keyakv2.cpp: ^~~~~~~~
Keyakv2.cpp: In file included from /usr/include/c++/6.2.1/memory:81:0,
Keyakv2.cpp: from Motorist.h:20,
Keyakv2.cpp: from Keyakv2.h:21,
Keyakv2.cpp: from Keyakv2.cpp:16:
Keyakv2.cpp: /usr/include/c++/6.2.1/bits/unique_ptr.h:49:28: note: declared here
Keyakv2.cpp: templategt; class auto_ptr;
Keyakv2.cpp: ^~~~~~~~
Motorist.cpp: In file included from Motorist.cpp:17:0:
Motorist.cpp: Motorist.h:31:5: warning: 'templategt; class std::auto_ptr' is deprecated [-Wdeprecated-declarations]
Motorist.cpp: auto_ptrgt; state;
Motorist.cpp: ^~~~~~~~
Motorist.cpp: In file included from /usr/include/c++/6.2.1/memory:81:0,
Motorist.cpp: from Motorist.h:20,
Motorist.cpp: from Motorist.cpp:17:
Motorist.cpp: /usr/include/c++/6.2.1/bits/unique_ptr.h:49:28: note: declared here
Motorist.cpp: templategt; class auto_ptr;
Motorist.cpp: ^~~~~~~~
encrypt.cpp: In file included from Keyakv2.h:21:0,
encrypt.cpp: from encrypt.cpp:22:
encrypt.cpp: Motorist.h:31:5: warning: 'templategt; class std::auto_ptr' is deprecated [-Wdeprecated-declarations]
encrypt.cpp: auto_ptrgt; state;
encrypt.cpp: ^~~~~~~~
encrypt.cpp: In file included from /usr/include/c++/6.2.1/memory:81:0,
encrypt.cpp: from Motorist.h:20,
encrypt.cpp: from Keyakv2.h:21,
encrypt.cpp: from encrypt.cpp:22:
encrypt.cpp: /usr/include/c++/6.2.1/bits/unique_ptr.h:49:28: note: declared here
encrypt.cpp: templategt; class auto_ptr;
encrypt.cpp: ^~~~~~~~

Number of similar (compiler,implementation) pairs: 12, namely:

Compiler	Implementations
g++ -funroll-loops -mcpu=marvell-pj4 -O2	ref refnew
g++ -funroll-loops -mcpu=marvell-pj4 -O3	ref refnew
g++ -funroll-loops -mcpu=marvell-pj4 -Os	ref refnew
g++ -mcpu=marvell-pj4 -O2	ref refnew
g++ -mcpu=marvell-pj4 -O3	ref refnew
g++ -mcpu=marvell-pj4 -Os	ref refnew

Compiler output

Implementation: crypto_aead/lunarkeyakv2/ARMv7A
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s: Assembler messages:
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:286: Error: selected FPU does not support instruction -- `vmov.i64 q0,#0'
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:287: Error: selected FPU does not support instruction -- `vmov.i64 q1,#0'
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:288: Error: selected FPU does not support instruction -- `vmov.i64 q2,#0'
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:289: Error: selected FPU does not support instruction -- `vmov.i64 q3,#0'
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:384: Error: selected FPU does not support instruction -- `vmov.i64 d0,#0'
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:525: Error: selected processor does not support `veor.64 d0,d0,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:527: Error: selected processor does not support `veor.64 d2,d2,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:529: Error: selected processor does not support `veor.64 d4,d4,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:531: Error: selected processor does not support `veor.64 d6,d6,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: ...
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:766: Error: selected processor does not support `veor.64 d9,d9,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:768: Error: selected processor does not support `veor.64 d7,d7,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:770: Error: selected processor does not support `veor.64 d5,d5,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:772: Error: selected processor does not support `veor.64 d3,d3,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:774: Error: selected processor does not support `veor.64 d1,d1,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:777: Error: selected processor does not support `veor.64 d8,d8,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:779: Error: selected processor does not support `veor.64 d6,d6,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:781: Error: selected processor does not support `veor.64 d4,d4,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:783: Error: selected processor does not support `veor.64 d2,d2,d30' in ARM mode
KeccakP-1600-armv7a-le-neon-gcc.s: KeccakP-1600-armv7a-le-neon-gcc.s:785: Error: selected processor does not support `veor.64 d0,d0,d30' in ARM mode

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	ARMv7A
gcc -funroll-loops -mcpu=marvell-pj4 -O3	ARMv7A
gcc -funroll-loops -mcpu=marvell-pj4 -Os	ARMv7A
gcc -mcpu=marvell-pj4 -O2	ARMv7A
gcc -mcpu=marvell-pj4 -O3	ARMv7A
gcc -mcpu=marvell-pj4 -Os	ARMv7A

Compiler output

Implementation: crypto_aead/lunarkeyakv2/ARMv8A
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s: Assembler messages:
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:257: Error: bad instruction `movi v0.2d,#0'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:258: Error: bad instruction `movi v1.2d,#0'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:259: Error: bad instruction `movi v2.2d,#0'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:260: Error: bad instruction `movi v3.2d,#0'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:261: Error: bad instruction `st4 { v0.2d,v1.2d,v2.2d,v3.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:262: Error: bad instruction `st4 { v0.2d,v1.2d,v2.2d,v3.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:263: Error: bad instruction `st4 { v0.2d,v1.2d,v2.2d,v3.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:264: Error: bad instruction `st1 { v0.d }[0],[x0],#8'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:265: Error: bad instruction `ret'
KeccakP-1600-armv8a-neon.s: ...
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:511: Error: ARM register expected -- `eor v25.16b,v25.16b,v7.16b'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:511: Error: ARM register expected -- `eor v26.16b,v26.16b,v4.16b'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:511: Error: bad instruction `ld1 { v16.d }[0],[x1],#8'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:511: Error: ARM register expected -- `eor v19.16b,v19.16b,v16.16b'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:512: Error: ARM register expected -- `subs w2,w2,#1'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:515: Error: bad instruction `st4 { v19.2d,v20.2d,v21.2d,v22.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:515: Error: bad instruction `st4 { v23.2d,v24.2d,v25.2d,v26.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:515: Error: bad instruction `st4 { v27.2d,v28.2d,v29.2d,v30.2d },[x0],#64'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:515: Error: bad instruction `st1 { v31.d }[0],[x0],#8'
KeccakP-1600-armv8a-neon.s: KeccakP-1600-armv8a-neon.s:516: Error: bad instruction `ret'

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	ARMv8A
gcc -funroll-loops -mcpu=marvell-pj4 -O3	ARMv8A
gcc -funroll-loops -mcpu=marvell-pj4 -Os	ARMv8A
gcc -mcpu=marvell-pj4 -O2	ARMv8A
gcc -mcpu=marvell-pj4 -O3	ARMv8A
gcc -mcpu=marvell-pj4 -Os	ARMv8A

Compiler output

Implementation: crypto_aead/lunarkeyakv2/AVR8
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s: Assembler messages:
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:2: Error: bad instruction `implementation by the Keccak,Keyak and Ketje Teams,namely,Guido Bertoni,'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:3: Error: bad instruction `joan Daemen,Michaël Peeters,Gilles Van Assche and Ronny Van Keer,hereby'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:4: Error: bad instruction `denoted as "the implementer".'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:6: Error: bad instruction `for more information,feedback or questions,please refer to our websites:'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:11: Error: bad instruction `to the extent possible under law,the implementer has waived all copyright'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:12: Error: ARM register expected -- `and related or neighboring rights to the source code in this file.'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:16: Error: bad instruction `tested on ATmega1280 simulator'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:18: Error: bad instruction `registers used in all routines'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:26: Error: junk at end of line, first unrecognized character is `-'
KeccakP-1600-avr8-fast.s: ...
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1069: Error: bad instruction `rjmp KeccakRhoPi_PiStore'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1072: Error: bad instruction `ld rTempBis+7,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1073: Error: bad instruction `ld rTempBis+0,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1074: Error: bad instruction `ld rTempBis+1,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1075: Error: bad instruction `ld rTempBis+2,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1076: Error: bad instruction `ld rTempBis+3,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1077: Error: bad instruction `ld rTempBis+4,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1078: Error: bad instruction `ld rTempBis+5,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1079: Error: bad instruction `ld rTempBis+6,Y+'
KeccakP-1600-avr8-fast.s: KeccakP-1600-avr8-fast.s:1080: Error: bad instruction `rjmp KeccakRhoPi_PiStore'

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	AVR8
gcc -funroll-loops -mcpu=marvell-pj4 -O3	AVR8
gcc -funroll-loops -mcpu=marvell-pj4 -Os	AVR8
gcc -mcpu=marvell-pj4 -O2	AVR8
gcc -mcpu=marvell-pj4 -O3	AVR8
gcc -mcpu=marvell-pj4 -Os	AVR8

Compiler output

Implementation: crypto_aead/lunarkeyakv2/ARMv7M
Compiler: gcc -funroll-loops -mcpu=marvell-pj4 -O2

KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s: Assembler messages:
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:406: Error: bad arguments to instruction -- `uxth r9'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:406: Error: bad arguments to instruction -- `uxth r10'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:406: Error: bad arguments to instruction -- `uxth r11'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:406: Error: bad arguments to instruction -- `uxth r12'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:406: Error: bad arguments to instruction -- `uxth lr'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:408: Error: bad arguments to instruction -- `uxth r9'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:408: Error: bad arguments to instruction -- `uxth r10'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:408: Error: bad arguments to instruction -- `uxth r11'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:408: Error: bad arguments to instruction -- `uxth r12'
KeccakP-400-armv7m-le-gcc.s: KeccakP-400-armv7m-le-gcc.s:408: Error: bad arguments to instruction -- `uxth lr'

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

Compiler	Implementations
gcc -funroll-loops -mcpu=marvell-pj4 -O2	ARMv7M
gcc -funroll-loops -mcpu=marvell-pj4 -O3	ARMv7M
gcc -funroll-loops -mcpu=marvell-pj4 -Os	ARMv7M
gcc -mcpu=marvell-pj4 -O2	ARMv7M
gcc -mcpu=marvell-pj4 -O3	ARMv7M
gcc -mcpu=marvell-pj4 -Os	ARMv7M