Implementation notes: armeabi, h6dragon, crypto_core/salsa2012

Computer: h6dragon
Architecture: armeabi
CPU ID: unknown CPU ID
SUPERCOP version: 20141014
Operation: crypto_core
Primitive: salsa2012

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
1699	`armneon2`	`gcc -fno-schedule-insns -O -fomit-frame-pointer`	20141014	20141014
1699	`armneon2`	`gcc -mcpu=cortex-a5 -O2 -fomit-frame-pointer`	20141014	20141014
1699	`armneon`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
1699	`armneon2`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
1699	`armneon2`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
1702	`armneon`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
1721	`armneon2`	`gcc -O -fomit-frame-pointer`	20141014	20141014
1721	`armneon2`	`gcc -funroll-loops`	20141014	20141014
1721	`armneon2`	`gcc -mcpu=cortex-a9 -O -fomit-frame-pointer`	20141014	20141014
1723	`armneon2`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
1725	`armneon2`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`clang -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive`	20141014	20141014
1727	`armneon2`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive`	20141014	20141014
1727	`armneon2`	`clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -fno-schedule-insns -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -funroll-loops -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -funroll-loops -O -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=arm810 -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=arm810 -O -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=arm8 -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a5 -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a5 -O -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a9 -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
1727	`armneon`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=strongarm110 -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=strongarm -O3 -fomit-frame-pointer`	20141014	20141014
1727	`armneon2`	`gcc -mcpu=strongarm -O -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -fno-schedule-insns -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -funroll-loops -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -mcpu=arm810 -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
1729	`armneon`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -mcpu=strongarm1100 -Os -fomit-frame-pointer`	20141014	20141014
1729	`armneon2`	`gcc -mcpu=strongarm110 -O -fomit-frame-pointer`	20141014	20141014
1749	`armneon2`	`gcc -funroll-loops -O3 -fomit-frame-pointer`	20141014	20141014
1751	`armneon`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer`	20141014	20141014
1751	`armneon`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive`	20141014	20141014
1751	`armneon2`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`clang -O3 -fwrapv -mavx2 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`clang -O3 -fwrapv -mavx -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc`	20141014	20141014
1751	`armneon2`	`gcc -O2 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -O3 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -fno-schedule-insns -O3 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -funroll-loops -fno-schedule-insns -Os -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=arm810 -O3 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=arm8 -O2 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=arm8 -O -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=arm8 -Os -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=cortex-a5 -Os -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm1100 -O2 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm1100 -O3 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm1100 -O -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm110 -O2 -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm110 -Os -fomit-frame-pointer`	20141014	20141014
1751	`armneon2`	`gcc -mcpu=strongarm -Os -fomit-frame-pointer`	20141014	20141014
1752	`armneon2`	`cc`	20141014	20141014
1752	`armneon2`	`gcc -funroll-loops -fno-schedule-insns -O2 -fomit-frame-pointer`	20141014	20141014
1752	`armneon2`	`gcc -funroll-loops -fno-schedule-insns -O3 -fomit-frame-pointer`	20141014	20141014
1752	`armneon2`	`gcc -funroll-loops -fno-schedule-insns -O -fomit-frame-pointer`	20141014	20141014
1752	`armneon`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
1753	`armneon2`	`gcc -mcpu=cortex-a9 -O3 -fomit-frame-pointer`	20141014	20141014
1753	`armneon2`	`gcc -mcpu=cortex-a9 -Os -fomit-frame-pointer`	20141014	20141014
1753	`armneon`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
1753	`armneon2`	`gcc -mcpu=strongarm -O2 -fomit-frame-pointer`	20141014	20141014
1780	`armneon`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
1856	`ref`	`gcc -fno-schedule-insns -O2 -fomit-frame-pointer`	20141014	20141014
1885	`ref`	`gcc -funroll-loops -fno-schedule-insns -O2 -fomit-frame-pointer`	20141014	20141014
1905	`ref`	`gcc -fno-schedule-insns -O3 -fomit-frame-pointer`	20141014	20141014
1906	`ref`	`gcc -funroll-loops -fno-schedule-insns -O3 -fomit-frame-pointer`	20141014	20141014
1957	`ref`	`clang -O3 -fomit-frame-pointer`	20141014	20141014
1958	`ref`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive`	20141014	20141014
1958	`ref`	`clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
1959	`ref`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive`	20141014	20141014
1959	`ref`	`clang -O3 -fwrapv -mavx2 -fomit-frame-pointer`	20141014	20141014
1959	`ref`	`clang -O3 -fwrapv -mavx -fomit-frame-pointer`	20141014	20141014
1984	`ref`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer`	20141014	20141014
1984	`ref`	`clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
1984	`ref`	`clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer`	20141014	20141014
1986	`ref`	`clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer`	20141014	20141014
1986	`ref`	`clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer`	20141014	20141014
2061	`ref`	`gcc -funroll-loops -fno-schedule-insns -Os -fomit-frame-pointer`	20141014	20141014
2087	`ref`	`gcc -mcpu=arm810 -Os -fomit-frame-pointer`	20141014	20141014
2091	`ref`	`gcc -funroll-loops -Os -fomit-frame-pointer`	20141014	20141014
2103	`ref`	`gcc -mcpu=strongarm110 -Os -fomit-frame-pointer`	20141014	20141014
2103	`ref`	`gcc -mcpu=strongarm -Os -fomit-frame-pointer`	20141014	20141014
2112	`ref`	`gcc -funroll-loops -O2 -fomit-frame-pointer`	20141014	20141014
2113	`ref`	`gcc -funroll-loops -O3 -fomit-frame-pointer`	20141014	20141014
2128	`ref`	`gcc -mcpu=arm8 -Os -fomit-frame-pointer`	20141014	20141014
2130	`ref`	`gcc -mcpu=strongarm1100 -Os -fomit-frame-pointer`	20141014	20141014
2140	`ref`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
2141	`ref`	`gcc -O3 -fomit-frame-pointer`	20141014	20141014
2162	`ref`	`gcc -mcpu=arm810 -O3 -fomit-frame-pointer`	20141014	20141014
2162	`ref`	`gcc -mcpu=strongarm110 -O3 -fomit-frame-pointer`	20141014	20141014
2162	`ref`	`gcc -mcpu=strongarm -O3 -fomit-frame-pointer`	20141014	20141014
2163	`ref`	`gcc -mcpu=arm810 -O2 -fomit-frame-pointer`	20141014	20141014
2163	`ref`	`gcc -mcpu=cortex-a9 -O2 -fomit-frame-pointer`	20141014	20141014
2164	`ref`	`gcc -O2 -fomit-frame-pointer`	20141014	20141014
2164	`ref`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
2164	`ref`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O2 -fomit-frame-pointer`	20141014	20141014
2164	`ref`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O3 -fomit-frame-pointer`	20141014	20141014
2164	`ref`	`gcc -mcpu=strongarm1100 -O2 -fomit-frame-pointer`	20141014	20141014
2165	`ref`	`gcc -Os -fomit-frame-pointer`	20141014	20141014
2166	`ref`	`gcc -mcpu=cortex-a9 -O3 -fomit-frame-pointer`	20141014	20141014
2177	`ref`	`gcc -fno-schedule-insns -Os -fomit-frame-pointer`	20141014	20141014
2179	`ref`	`gcc -mcpu=arm8 -O2 -fomit-frame-pointer`	20141014	20141014
2184	`ref`	`gcc -mcpu=strongarm1100 -O3 -fomit-frame-pointer`	20141014	20141014
2190	`ref`	`gcc -mcpu=arm8 -O3 -fomit-frame-pointer`	20141014	20141014
2190	`ref`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
2190	`ref`	`gcc -mcpu=strongarm110 -O2 -fomit-frame-pointer`	20141014	20141014
2190	`ref`	`gcc -mcpu=strongarm -O2 -fomit-frame-pointer`	20141014	20141014
2203	`ref`	`gcc -mcpu=cortex-a9 -Os -fomit-frame-pointer`	20141014	20141014
2210	`ref`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -Os -fomit-frame-pointer`	20141014	20141014
2216	`ref`	`gcc -mcpu=cortex-a5 -O3 -fomit-frame-pointer`	20141014	20141014
2229	`ref`	`gcc -mcpu=cortex-a5 -Os -fomit-frame-pointer`	20141014	20141014
2243	`ref`	`gcc -mcpu=cortex-a5 -O2 -fomit-frame-pointer`	20141014	20141014
3244	`ref`	`gcc -mcpu=strongarm -O -fomit-frame-pointer`	20141014	20141014
3245	`ref`	`gcc -mcpu=arm8 -O -fomit-frame-pointer`	20141014	20141014
3245	`ref`	`gcc -mcpu=strongarm1100 -O -fomit-frame-pointer`	20141014	20141014
3268	`ref`	`gcc -mcpu=arm810 -O -fomit-frame-pointer`	20141014	20141014
3272	`ref`	`gcc -mcpu=strongarm110 -O -fomit-frame-pointer`	20141014	20141014
3399	`ref`	`gcc -funroll-loops -fno-schedule-insns -O -fomit-frame-pointer`	20141014	20141014
3401	`ref`	`gcc -funroll-loops -O -fomit-frame-pointer`	20141014	20141014
3606	`ref`	`gcc -mcpu=cortex-a5 -O -fomit-frame-pointer`	20141014	20141014
3658	`ref`	`gcc -mcpu=cortex-a9 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
3709	`ref`	`gcc -O -fomit-frame-pointer`	20141014	20141014
3709	`ref`	`gcc -fno-schedule-insns -O -fomit-frame-pointer`	20141014	20141014
3709	`ref`	`gcc -mcpu=cortex-a9 -O -fomit-frame-pointer`	20141014	20141014
3733	`ref`	`gcc -mcpu=cortex-a8 -mfloat-abi=hard -mfpu=neon -O -fomit-frame-pointer`	20141014	20141014
8912	`ref`	`cc`	20141014	20141014
8912	`ref`	`gcc`	20141014	20141014
8940	`ref`	`gcc -funroll-loops`	20141014	20141014

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: cc

core.c: In file included from core.c:9:0:
core.c: /usr/lib/gcc/arm-linux-gnueabihf/4.7/include/arm_neon.h:32:2: error: #error You must enable NEON instructions (e.g. -mfloat-abi=softfp -mfpu=neon) to use arm_neon.h
core.c: core.c: In function 'crypto_core_salsa2012_armneon':
core.c: core.c:21:3: error: unknown type name 'uint32x4_t'
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:23:3: error: unknown type name 'uint32x4_t'
core.c: core.c:23:26: error: 'uint32x4_t' undeclared (first use in this function)
core.c: core.c:23:26: note: each undeclared identifier is reported only once for each function it appears in
core.c: core.c:23:38: error: expected ',' or ';' before 'vld1q_u8'
core.c: core.c:24:14: error: expected ';' before 'k4k5k6k7'
core.c: core.c:25:14: error: expected ';' before 'c0c1c2c3'
core.c: core.c:26:14: error: expected ';' before 'n0n1n2n3'
core.c: core.c:28:14: error: expected ';' before 'n1n2n3n0'
core.c: core.c:29:3: error: unknown type name 'uint32x2_t'
core.c: core.c:29:34: error: 'n1n2n3n0' undeclared (first use in this function)
core.c: core.c:30:3: error: unknown type name 'uint32x2_t'
core.c: core.c:31:3: error: unknown type name 'uint32x2_t'
core.c: core.c:32:3: error: unknown type name 'uint32x2_t'
core.c: core.c:33:3: error: unknown type name 'uint32x2_t'
core.c: core.c:33:34: error: 'k4k5k6k7' undeclared (first use in this function)
core.c: ...

Number of similar (compiler,implementation) pairs: 27, namely:

Compiler	Implementations
cc	armneon
gcc	armneon
gcc -O2 -fomit-frame-pointer	armneon
gcc -O3 -fomit-frame-pointer	armneon
gcc -O -fomit-frame-pointer	armneon
gcc -Os -fomit-frame-pointer	armneon
gcc -fno-schedule-insns -O2 -fomit-frame-pointer	armneon
gcc -fno-schedule-insns -O3 -fomit-frame-pointer	armneon
gcc -fno-schedule-insns -O -fomit-frame-pointer	armneon
gcc -fno-schedule-insns -Os -fomit-frame-pointer	armneon
gcc -funroll-loops	armneon
gcc -funroll-loops -O2 -fomit-frame-pointer	armneon
gcc -funroll-loops -O3 -fomit-frame-pointer	armneon
gcc -funroll-loops -O -fomit-frame-pointer	armneon
gcc -funroll-loops -Os -fomit-frame-pointer	armneon
gcc -funroll-loops -fno-schedule-insns -O2 -fomit-frame-pointer	armneon
gcc -funroll-loops -fno-schedule-insns -O3 -fomit-frame-pointer	armneon
gcc -funroll-loops -fno-schedule-insns -O -fomit-frame-pointer	armneon
gcc -funroll-loops -fno-schedule-insns -Os -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a5 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a5 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a5 -O -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a5 -Os -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a9 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a9 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a9 -O -fomit-frame-pointer	armneon
gcc -mcpu=cortex-a9 -Os -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fomit-frame-pointer

core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 2, namely:

Compiler	Implementations
clang -O3 -fomit-frame-pointer	armneon
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fomit-frame-pointer

core.s: clang: warning: argument unused during compilation: '-O3'
core.s: clang: warning: argument unused during compilation: '-fomit-frame-pointer'
core.s: clang: warning: argument unused during compilation: '-D SUPERCOP'
core.s: clang: warning: argument unused during compilation: '-I .'
core.s: clang: warning: argument unused during compilation: '-I /home/djb/benchmarking/supercop-20141014/supercop-data/h6dragon/armeabi/include'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 2, namely:

Compiler	Implementations
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer	armneon2
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive

Number of similar (compiler,implementation) pairs: 3, namely:

Compiler	Implementations
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive	armneon ref
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive

core.s: clang: warning: argument unused during compilation: '-O3'
core.s: clang: warning: argument unused during compilation: '-fwrapv'
core.s: clang: warning: argument unused during compilation: '-fomit-frame-pointer'
core.s: clang: warning: argument unused during compilation: '-fpolly'
core.s: clang: warning: argument unused during compilation: '-funroll-loops'
core.s: clang: warning: argument unused during compilation: '-fvectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.s: clang: warning: argument unused during compilation: '-D SUPERCOP'
core.s: clang: warning: argument unused during compilation: '-I .'
core.s: clang: warning: argument unused during compilation: '-I /home/djb/benchmarking/supercop-20141014/supercop-data/h6dragon/armeabi/include'

Number of similar (compiler,implementation) pairs: 2, namely:

Compiler	Implementations
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=neon -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive	armneon2
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive	armneon2

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive

core.c: clang: warning: argument unused during compilation: '-fpolly'
core.c: clang: warning: argument unused during compilation: '-fvectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -march=armv7-a -mfloat-abi=hard -mfpu=vfpv3-d16 -fomit-frame-pointer -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive	armneon

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx2 -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx2'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx2 -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx2'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx2 -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx2'
core.c: clang: warning: argument unused during compilation: '-fpolly'
core.c: clang: warning: argument unused during compilation: '-fvectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.s: clang: warning: argument unused during compilation: '-O3'
core.s: clang: warning: argument unused during compilation: '-fwrapv'
core.s: clang: warning: argument unused during compilation: '-mavx2'
core.s: clang: warning: argument unused during compilation: '-fpolly'
core.s: clang: warning: argument unused during compilation: '-funroll-loops'
core.s: clang: warning: argument unused during compilation: '-fvectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.s: clang: warning: argument unused during compilation: '-fomit-frame-pointer'
core.s: clang: warning: argument unused during compilation: '-D SUPERCOP'
core.s: clang: warning: argument unused during compilation: '-I .'
core.s: clang: warning: argument unused during compilation: '-I /home/djb/benchmarking/supercop-20141014/supercop-data/h6dragon/armeabi/include'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx2 -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'
core.c: clang: warning: argument unused during compilation: '-fpolly'
core.c: clang: warning: argument unused during compilation: '-fvectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.s: clang: warning: argument unused during compilation: '-O3'
core.s: clang: warning: argument unused during compilation: '-fwrapv'
core.s: clang: warning: argument unused during compilation: '-mavx'
core.s: clang: warning: argument unused during compilation: '-fpolly'
core.s: clang: warning: argument unused during compilation: '-funroll-loops'
core.s: clang: warning: argument unused during compilation: '-fvectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.s: clang: warning: argument unused during compilation: '-fomit-frame-pointer'
core.s: clang: warning: argument unused during compilation: '-D SUPERCOP'
core.s: clang: warning: argument unused during compilation: '-I .'
core.s: clang: warning: argument unused during compilation: '-I /home/djb/benchmarking/supercop-20141014/supercop-data/h6dragon/armeabi/include'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'
core.c: clang: warning: argument unused during compilation: '-maes'
core.c: clang: warning: argument unused during compilation: '-mpclmul'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'
core.c: clang: warning: argument unused during compilation: '-maes'
core.c: clang: warning: argument unused during compilation: '-mpclmul'
core.c: clang: warning: argument unused during compilation: '-fpolly'
core.c: clang: warning: argument unused during compilation: '-fvectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.c: clang: warning: argument unused during compilation: '-mavx'
core.c: clang: warning: argument unused during compilation: '-maes'
core.c: clang: warning: argument unused during compilation: '-mpclmul'
core.c: clang: warning: argument unused during compilation: '-fpolly'
core.c: clang: warning: argument unused during compilation: '-fvectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.c: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.c: In file included from core.c:9:
core.c: /usr/include/clang/3.2/include/arm_neon.h:28:2: error: "NEON support not enabled"
core.c: #error "NEON support not enabled"
core.c: ^
core.c: 1 error generated.

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer

core.s: clang: warning: argument unused during compilation: '-O3'
core.s: clang: warning: argument unused during compilation: '-fwrapv'
core.s: clang: warning: argument unused during compilation: '-mavx'
core.s: clang: warning: argument unused during compilation: '-maes'
core.s: clang: warning: argument unused during compilation: '-mpclmul'
core.s: clang: warning: argument unused during compilation: '-fpolly'
core.s: clang: warning: argument unused during compilation: '-funroll-loops'
core.s: clang: warning: argument unused during compilation: '-fvectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize'
core.s: clang: warning: argument unused during compilation: '-fslp-vectorize-aggressive'
core.s: clang: warning: argument unused during compilation: '-fomit-frame-pointer'
core.s: clang: warning: argument unused during compilation: '-D SUPERCOP'
core.s: clang: warning: argument unused during compilation: '-I .'
core.s: clang: warning: argument unused during compilation: '-I /home/djb/benchmarking/supercop-20141014/supercop-data/h6dragon/armeabi/include'

Number of similar (compiler,implementation) pairs: 1, namely:

Compiler	Implementations
clang -O3 -fwrapv -mavx -maes -mpclmul -fpolly -funroll-loops -fvectorize -fslp-vectorize -fslp-vectorize-aggressive -fomit-frame-pointer	armneon2

Compiler output

Implementation: crypto_core/salsa2012/armneon
Compiler: gcc -mcpu=arm810 -O2 -fomit-frame-pointer

core.c: core.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
core.c: In file included from core.c:9:0:
core.c: /usr/lib/gcc/arm-linux-gnueabihf/4.7/include/arm_neon.h:32:2: error: #error You must enable NEON instructions (e.g. -mfloat-abi=softfp -mfpu=neon) to use arm_neon.h
core.c: core.c: In function 'crypto_core_salsa2012_armneon':
core.c: core.c:21:3: error: unknown type name 'uint32x4_t'
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:21:3: warning: excess elements in scalar initializer [enabled by default]
core.c: core.c:21:3: warning: (near initialization for 'abab') [enabled by default]
core.c: core.c:23:3: error: unknown type name 'uint32x4_t'
core.c: core.c:23:26: error: 'uint32x4_t' undeclared (first use in this function)
core.c: core.c:23:26: note: each undeclared identifier is reported only once for each function it appears in
core.c: core.c:23:38: error: expected ',' or ';' before 'vld1q_u8'
core.c: core.c:24:14: error: expected ';' before 'k4k5k6k7'
core.c: core.c:25:14: error: expected ';' before 'c0c1c2c3'
core.c: core.c:26:14: error: expected ';' before 'n0n1n2n3'
core.c: core.c:28:14: error: expected ';' before 'n1n2n3n0'
core.c: core.c:29:3: error: unknown type name 'uint32x2_t'
core.c: core.c:29:34: error: 'n1n2n3n0' undeclared (first use in this function)
core.c: core.c:30:3: error: unknown type name 'uint32x2_t'
core.c: core.c:31:3: error: unknown type name 'uint32x2_t'
core.c: core.c:32:3: error: unknown type name 'uint32x2_t'
core.c: core.c:33:3: error: unknown type name 'uint32x2_t'
core.c: ...

Number of similar (compiler,implementation) pairs: 20, namely:

Compiler	Implementations
gcc -mcpu=arm810 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=arm810 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=arm810 -O -fomit-frame-pointer	armneon
gcc -mcpu=arm810 -Os -fomit-frame-pointer	armneon
gcc -mcpu=arm8 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=arm8 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=arm8 -O -fomit-frame-pointer	armneon
gcc -mcpu=arm8 -Os -fomit-frame-pointer	armneon
gcc -mcpu=strongarm1100 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm1100 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm1100 -O -fomit-frame-pointer	armneon
gcc -mcpu=strongarm1100 -Os -fomit-frame-pointer	armneon
gcc -mcpu=strongarm110 -O2 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm110 -O3 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm110 -O -fomit-frame-pointer	armneon
gcc -mcpu=strongarm110 -Os -fomit-frame-pointer	armneon
gcc -mcpu=strongarm -O2 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm -O3 -fomit-frame-pointer	armneon
gcc -mcpu=strongarm -O -fomit-frame-pointer	armneon
gcc -mcpu=strongarm -Os -fomit-frame-pointer	armneon

Compiler output

Implementation: crypto_core/salsa2012/ref
Compiler: gcc -mcpu=arm810 -O2 -fomit-frame-pointer

core.c: core.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
measure.c: measure.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
measure.c: measure-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]

Number of similar (compiler,implementation) pairs: 20, namely:

Compiler	Implementations
gcc -mcpu=arm810 -O2 -fomit-frame-pointer	ref
gcc -mcpu=arm810 -O3 -fomit-frame-pointer	ref
gcc -mcpu=arm810 -O -fomit-frame-pointer	ref
gcc -mcpu=arm810 -Os -fomit-frame-pointer	ref
gcc -mcpu=arm8 -O2 -fomit-frame-pointer	ref
gcc -mcpu=arm8 -O3 -fomit-frame-pointer	ref
gcc -mcpu=arm8 -O -fomit-frame-pointer	ref
gcc -mcpu=arm8 -Os -fomit-frame-pointer	ref
gcc -mcpu=strongarm1100 -O2 -fomit-frame-pointer	ref
gcc -mcpu=strongarm1100 -O3 -fomit-frame-pointer	ref
gcc -mcpu=strongarm1100 -O -fomit-frame-pointer	ref
gcc -mcpu=strongarm1100 -Os -fomit-frame-pointer	ref
gcc -mcpu=strongarm110 -O2 -fomit-frame-pointer	ref
gcc -mcpu=strongarm110 -O3 -fomit-frame-pointer	ref
gcc -mcpu=strongarm110 -O -fomit-frame-pointer	ref
gcc -mcpu=strongarm110 -Os -fomit-frame-pointer	ref
gcc -mcpu=strongarm -O2 -fomit-frame-pointer	ref
gcc -mcpu=strongarm -O3 -fomit-frame-pointer	ref
gcc -mcpu=strongarm -O -fomit-frame-pointer	ref
gcc -mcpu=strongarm -Os -fomit-frame-pointer	ref

Compiler output

Implementation: crypto_core/salsa2012/armneon2
Compiler: gcc -mcpu=arm810 -O2 -fomit-frame-pointer

try.c: try.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
try.c: try-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
measure.c: measure.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]
measure.c: measure-anything.c:1:0: warning: target CPU does not support THUMB instructions [enabled by default]

Number of similar (compiler,implementation) pairs: 20, namely:

Compiler	Implementations
gcc -mcpu=arm810 -O2 -fomit-frame-pointer	armneon2
gcc -mcpu=arm810 -O3 -fomit-frame-pointer	armneon2
gcc -mcpu=arm810 -O -fomit-frame-pointer	armneon2
gcc -mcpu=arm810 -Os -fomit-frame-pointer	armneon2
gcc -mcpu=arm8 -O2 -fomit-frame-pointer	armneon2
gcc -mcpu=arm8 -O3 -fomit-frame-pointer	armneon2
gcc -mcpu=arm8 -O -fomit-frame-pointer	armneon2
gcc -mcpu=arm8 -Os -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm1100 -O2 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm1100 -O3 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm1100 -O -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm1100 -Os -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm110 -O2 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm110 -O3 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm110 -O -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm110 -Os -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm -O2 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm -O3 -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm -O -fomit-frame-pointer	armneon2
gcc -mcpu=strongarm -Os -fomit-frame-pointer	armneon2