Implementation notes: mipso32, loongson, crypto_scalarmult/curve25519

Computer: loongson
Architecture: mipso32
CPU ID: unknown CPU ID
SUPERCOP version: 20161026
Operation: crypto_scalarmult
Primitive: curve25519

Time	Implementation	Compiler	Benchmark date	SUPERCOP version
1981836	`donna_c64`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -Os`	20161220	20161026
2024424	`donna_c64`	`gcc -march=loongson2f -mabi=n32 -Os`	20161220	20161026
2517398	`donna_c64`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O2`	20161220	20161026
2588566	`donna_c64`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O3`	20161220	20161026
2623252	`donna_c64`	`gcc -march=loongson2f -mabi=n32 -O3`	20161220	20161026
2673922	`donna_c64`	`gcc -march=loongson2f -mabi=n32 -O2`	20161220	20161026
7454528	`ref10`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O2`	20161220	20161026
7533246	`ref10`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O3`	20161220	20161026
7846808	`ref10`	`gcc -march=loongson2f -mabi=n32 -O2`	20161220	20161026
7850032	`ref10`	`gcc -march=loongson2f -mabi=n32 -O3`	20161220	20161026
8325248	`ref10`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -Os`	20161220	20161026
8448662	`ref10`	`gcc -march=loongson2f -mabi=n32 -Os`	20161220	20161026
62084408	`ref`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O3`	20161220	20161026
63735358	`ref`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -O2`	20161220	20161026
69289278	`ref`	`gcc -march=loongson2f -mabi=n32 -O3`	20161220	20161026
69422026	`ref`	`gcc -march=loongson2f -mabi=n32 -O2`	20161220	20161026
85477470	`ref`	`gcc -funroll-loops -march=loongson2f -mabi=n32 -Os`	20161220	20161026
89810990	`ref`	`gcc -march=loongson2f -mabi=n32 -Os`	20161220	20161026

Compiler output

Implementation: crypto_scalarmult/curve25519/donna_c64
Compiler: gcc -funroll-loops -march=loongson2f -mabi=n32 -O2

smult.c: smult.c:149:1: warning: always_inline function might not be inlinable [-Wattributes]
smult.c: fsquare_times(felem output, const felem in, limb count) {
smult.c: ^
smult.c: smult.c:100:1: warning: always_inline function might not be inlinable [-Wattributes]
smult.c: fmul(felem output, const felem in2, const felem in) {
smult.c: ^
smult.c: smult.c:70:1: warning: always_inline function might not be inlinable [-Wattributes]
smult.c: fscalar_product(felem output, const felem in, const limb scalar) {
smult.c: ^
smult.c: smult.c:56:1: warning: always_inline function might not be inlinable [-Wattributes]
smult.c: fdifference_backwards(felem out, const felem in) {
smult.c: ^
smult.c: smult.c:41:1: warning: always_inline function might not be inlinable [-Wattributes]
smult.c: fsum(limb *output, const limb *in) {
smult.c: ^

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

Compiler	Implementations
gcc -funroll-loops -march=loongson2f -mabi=n32 -O2	donna_c64
gcc -funroll-loops -march=loongson2f -mabi=n32 -O3	donna_c64
gcc -funroll-loops -march=loongson2f -mabi=n32 -Os	donna_c64
gcc -march=loongson2f -mabi=n32 -O2	donna_c64
gcc -march=loongson2f -mabi=n32 -O3	donna_c64
gcc -march=loongson2f -mabi=n32 -Os	donna_c64