1
0
mirror of https://github.com/henrydcase/nobs.git synced 2024-11-22 23:28:57 +00:00
nobs/dh/csidh/fp511_generic.go

124 lines
2.7 KiB
Go
Raw Permalink Normal View History

cSIDH-511: (#26) Implementation of Commutative Supersingular Isogeny Diffie Hellman, based on "A faster way to CSIDH" paper (2018/782). * For fast isogeny calculation, implementation converts a curve from Montgomery to Edwards. All calculations are done on Edwards curve and then converted back to Montgomery. * As multiplication in a field Fp511 is most expensive operation the implementation contains multiple multiplications. It has most performant, assembly implementation which uses BMI2 and ADOX/ADCX instructions for modern CPUs. It also contains slower implementation which will run on older CPUs * Benchmarks (Intel SkyLake): BenchmarkGeneratePrivate 6459 172213 ns/op 0 B/op 0 allocs/op BenchmarkGenerateKeyPair 25 45800356 ns/op 0 B/op 0 allocs/op BenchmarkValidate 297 3915983 ns/op 0 B/op 0 allocs/op BenchmarkValidateRandom 184683 6231 ns/op 0 B/op 0 allocs/op BenchmarkValidateGenerated 25 48481306 ns/op 0 B/op 0 allocs/op BenchmarkDerive 19 60928763 ns/op 0 B/op 0 allocs/op BenchmarkDeriveGenerated 8 137342421 ns/op 0 B/op 0 allocs/op BenchmarkXMul 2311 494267 ns/op 1 B/op 0 allocs/op BenchmarkXAdd 2396754 501 ns/op 0 B/op 0 allocs/op BenchmarkXDbl 2072690 571 ns/op 0 B/op 0 allocs/op BenchmarkIsom 78004 15171 ns/op 0 B/op 0 allocs/op BenchmarkFp512Sub 224635152 5.33 ns/op 0 B/op 0 allocs/op BenchmarkFp512Mul 246633255 4.90 ns/op 0 B/op 0 allocs/op BenchmarkCSwap 233228547 5.10 ns/op 0 B/op 0 allocs/op BenchmarkAddRdc 87348240 12.6 ns/op 0 B/op 0 allocs/op BenchmarkSubRdc 95112787 11.7 ns/op 0 B/op 0 allocs/op BenchmarkModExpRdc 25436 46878 ns/op 0 B/op 0 allocs/op BenchmarkMulBmiAsm 19527573 60.1 ns/op 0 B/op 0 allocs/op BenchmarkMulGeneric 7117650 164 ns/op 0 B/op 0 allocs/op * Go code has very similar performance when compared to C implementation. Results from sidh_torturer (4e2996e12d68364761064341cbe1d1b47efafe23) github.com:henrydcase/sidh-torture/csidh | TestName |Go | C | |------------------|----------|----------| |TestSharedSecret | 57.95774 | 57.91092 | |TestKeyGeneration | 62.23614 | 58.12980 | |TestSharedSecret | 55.28988 | 57.23132 | |TestKeyGeneration | 61.68745 | 58.66396 | |TestSharedSecret | 63.19408 | 58.64774 | |TestKeyGeneration | 62.34022 | 61.62539 | |TestSharedSecret | 62.85453 | 68.74503 | |TestKeyGeneration | 52.58518 | 58.40115 | |TestSharedSecret | 50.77081 | 61.91699 | |TestKeyGeneration | 59.91843 | 61.09266 | |TestSharedSecret | 59.97962 | 62.98151 | |TestKeyGeneration | 64.57525 | 56.22863 | |TestSharedSecret | 56.40521 | 55.77447 | |TestKeyGeneration | 67.85850 | 58.52604 | |TestSharedSecret | 60.54290 | 65.14052 | |TestKeyGeneration | 65.45766 | 58.42823 | On average Go implementation is 2% faster.
2019-11-24 03:39:35 +00:00
// +build noasm arm64
package csidh
import "math/bits"
2020-05-14 00:48:43 +01:00
// mul576 implements schoolbook multiplication of
// 64x512-bit integer. Returns result modulo 2^512.
// r = m1*m2
cSIDH-511: (#26) Implementation of Commutative Supersingular Isogeny Diffie Hellman, based on "A faster way to CSIDH" paper (2018/782). * For fast isogeny calculation, implementation converts a curve from Montgomery to Edwards. All calculations are done on Edwards curve and then converted back to Montgomery. * As multiplication in a field Fp511 is most expensive operation the implementation contains multiple multiplications. It has most performant, assembly implementation which uses BMI2 and ADOX/ADCX instructions for modern CPUs. It also contains slower implementation which will run on older CPUs * Benchmarks (Intel SkyLake): BenchmarkGeneratePrivate 6459 172213 ns/op 0 B/op 0 allocs/op BenchmarkGenerateKeyPair 25 45800356 ns/op 0 B/op 0 allocs/op BenchmarkValidate 297 3915983 ns/op 0 B/op 0 allocs/op BenchmarkValidateRandom 184683 6231 ns/op 0 B/op 0 allocs/op BenchmarkValidateGenerated 25 48481306 ns/op 0 B/op 0 allocs/op BenchmarkDerive 19 60928763 ns/op 0 B/op 0 allocs/op BenchmarkDeriveGenerated 8 137342421 ns/op 0 B/op 0 allocs/op BenchmarkXMul 2311 494267 ns/op 1 B/op 0 allocs/op BenchmarkXAdd 2396754 501 ns/op 0 B/op 0 allocs/op BenchmarkXDbl 2072690 571 ns/op 0 B/op 0 allocs/op BenchmarkIsom 78004 15171 ns/op 0 B/op 0 allocs/op BenchmarkFp512Sub 224635152 5.33 ns/op 0 B/op 0 allocs/op BenchmarkFp512Mul 246633255 4.90 ns/op 0 B/op 0 allocs/op BenchmarkCSwap 233228547 5.10 ns/op 0 B/op 0 allocs/op BenchmarkAddRdc 87348240 12.6 ns/op 0 B/op 0 allocs/op BenchmarkSubRdc 95112787 11.7 ns/op 0 B/op 0 allocs/op BenchmarkModExpRdc 25436 46878 ns/op 0 B/op 0 allocs/op BenchmarkMulBmiAsm 19527573 60.1 ns/op 0 B/op 0 allocs/op BenchmarkMulGeneric 7117650 164 ns/op 0 B/op 0 allocs/op * Go code has very similar performance when compared to C implementation. Results from sidh_torturer (4e2996e12d68364761064341cbe1d1b47efafe23) github.com:henrydcase/sidh-torture/csidh | TestName |Go | C | |------------------|----------|----------| |TestSharedSecret | 57.95774 | 57.91092 | |TestKeyGeneration | 62.23614 | 58.12980 | |TestSharedSecret | 55.28988 | 57.23132 | |TestKeyGeneration | 61.68745 | 58.66396 | |TestSharedSecret | 63.19408 | 58.64774 | |TestKeyGeneration | 62.34022 | 61.62539 | |TestSharedSecret | 62.85453 | 68.74503 | |TestKeyGeneration | 52.58518 | 58.40115 | |TestSharedSecret | 50.77081 | 61.91699 | |TestKeyGeneration | 59.91843 | 61.09266 | |TestSharedSecret | 59.97962 | 62.98151 | |TestKeyGeneration | 64.57525 | 56.22863 | |TestSharedSecret | 56.40521 | 55.77447 | |TestKeyGeneration | 67.85850 | 58.52604 | |TestSharedSecret | 60.54290 | 65.14052 | |TestKeyGeneration | 65.45766 | 58.42823 | On average Go implementation is 2% faster.
2019-11-24 03:39:35 +00:00
func mul512(r, m1 *fp, m2 uint64) {
var c, h, l uint64
c, r[0] = bits.Mul64(m2, m1[0])
h, l = bits.Mul64(m2, m1[1])
r[1], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[2])
r[2], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[3])
r[3], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[4])
r[4], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[5])
r[5], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[6])
r[6], c = bits.Add64(l, c, 0)
c = h + c
2020-05-14 00:48:43 +01:00
_, l = bits.Mul64(m2, m1[7])
cSIDH-511: (#26) Implementation of Commutative Supersingular Isogeny Diffie Hellman, based on "A faster way to CSIDH" paper (2018/782). * For fast isogeny calculation, implementation converts a curve from Montgomery to Edwards. All calculations are done on Edwards curve and then converted back to Montgomery. * As multiplication in a field Fp511 is most expensive operation the implementation contains multiple multiplications. It has most performant, assembly implementation which uses BMI2 and ADOX/ADCX instructions for modern CPUs. It also contains slower implementation which will run on older CPUs * Benchmarks (Intel SkyLake): BenchmarkGeneratePrivate 6459 172213 ns/op 0 B/op 0 allocs/op BenchmarkGenerateKeyPair 25 45800356 ns/op 0 B/op 0 allocs/op BenchmarkValidate 297 3915983 ns/op 0 B/op 0 allocs/op BenchmarkValidateRandom 184683 6231 ns/op 0 B/op 0 allocs/op BenchmarkValidateGenerated 25 48481306 ns/op 0 B/op 0 allocs/op BenchmarkDerive 19 60928763 ns/op 0 B/op 0 allocs/op BenchmarkDeriveGenerated 8 137342421 ns/op 0 B/op 0 allocs/op BenchmarkXMul 2311 494267 ns/op 1 B/op 0 allocs/op BenchmarkXAdd 2396754 501 ns/op 0 B/op 0 allocs/op BenchmarkXDbl 2072690 571 ns/op 0 B/op 0 allocs/op BenchmarkIsom 78004 15171 ns/op 0 B/op 0 allocs/op BenchmarkFp512Sub 224635152 5.33 ns/op 0 B/op 0 allocs/op BenchmarkFp512Mul 246633255 4.90 ns/op 0 B/op 0 allocs/op BenchmarkCSwap 233228547 5.10 ns/op 0 B/op 0 allocs/op BenchmarkAddRdc 87348240 12.6 ns/op 0 B/op 0 allocs/op BenchmarkSubRdc 95112787 11.7 ns/op 0 B/op 0 allocs/op BenchmarkModExpRdc 25436 46878 ns/op 0 B/op 0 allocs/op BenchmarkMulBmiAsm 19527573 60.1 ns/op 0 B/op 0 allocs/op BenchmarkMulGeneric 7117650 164 ns/op 0 B/op 0 allocs/op * Go code has very similar performance when compared to C implementation. Results from sidh_torturer (4e2996e12d68364761064341cbe1d1b47efafe23) github.com:henrydcase/sidh-torture/csidh | TestName |Go | C | |------------------|----------|----------| |TestSharedSecret | 57.95774 | 57.91092 | |TestKeyGeneration | 62.23614 | 58.12980 | |TestSharedSecret | 55.28988 | 57.23132 | |TestKeyGeneration | 61.68745 | 58.66396 | |TestSharedSecret | 63.19408 | 58.64774 | |TestKeyGeneration | 62.34022 | 61.62539 | |TestSharedSecret | 62.85453 | 68.74503 | |TestKeyGeneration | 52.58518 | 58.40115 | |TestSharedSecret | 50.77081 | 61.91699 | |TestKeyGeneration | 59.91843 | 61.09266 | |TestSharedSecret | 59.97962 | 62.98151 | |TestKeyGeneration | 64.57525 | 56.22863 | |TestSharedSecret | 56.40521 | 55.77447 | |TestKeyGeneration | 67.85850 | 58.52604 | |TestSharedSecret | 60.54290 | 65.14052 | |TestKeyGeneration | 65.45766 | 58.42823 | On average Go implementation is 2% faster.
2019-11-24 03:39:35 +00:00
r[7], _ = bits.Add64(l, c, 0)
}
2020-05-14 00:48:43 +01:00
// mul576 implements schoolbook multiplication of
// 64x512-bit integer. Returns 576-bit result of
// multiplication.
// r = m1*m2
cSIDH-511: (#26) Implementation of Commutative Supersingular Isogeny Diffie Hellman, based on "A faster way to CSIDH" paper (2018/782). * For fast isogeny calculation, implementation converts a curve from Montgomery to Edwards. All calculations are done on Edwards curve and then converted back to Montgomery. * As multiplication in a field Fp511 is most expensive operation the implementation contains multiple multiplications. It has most performant, assembly implementation which uses BMI2 and ADOX/ADCX instructions for modern CPUs. It also contains slower implementation which will run on older CPUs * Benchmarks (Intel SkyLake): BenchmarkGeneratePrivate 6459 172213 ns/op 0 B/op 0 allocs/op BenchmarkGenerateKeyPair 25 45800356 ns/op 0 B/op 0 allocs/op BenchmarkValidate 297 3915983 ns/op 0 B/op 0 allocs/op BenchmarkValidateRandom 184683 6231 ns/op 0 B/op 0 allocs/op BenchmarkValidateGenerated 25 48481306 ns/op 0 B/op 0 allocs/op BenchmarkDerive 19 60928763 ns/op 0 B/op 0 allocs/op BenchmarkDeriveGenerated 8 137342421 ns/op 0 B/op 0 allocs/op BenchmarkXMul 2311 494267 ns/op 1 B/op 0 allocs/op BenchmarkXAdd 2396754 501 ns/op 0 B/op 0 allocs/op BenchmarkXDbl 2072690 571 ns/op 0 B/op 0 allocs/op BenchmarkIsom 78004 15171 ns/op 0 B/op 0 allocs/op BenchmarkFp512Sub 224635152 5.33 ns/op 0 B/op 0 allocs/op BenchmarkFp512Mul 246633255 4.90 ns/op 0 B/op 0 allocs/op BenchmarkCSwap 233228547 5.10 ns/op 0 B/op 0 allocs/op BenchmarkAddRdc 87348240 12.6 ns/op 0 B/op 0 allocs/op BenchmarkSubRdc 95112787 11.7 ns/op 0 B/op 0 allocs/op BenchmarkModExpRdc 25436 46878 ns/op 0 B/op 0 allocs/op BenchmarkMulBmiAsm 19527573 60.1 ns/op 0 B/op 0 allocs/op BenchmarkMulGeneric 7117650 164 ns/op 0 B/op 0 allocs/op * Go code has very similar performance when compared to C implementation. Results from sidh_torturer (4e2996e12d68364761064341cbe1d1b47efafe23) github.com:henrydcase/sidh-torture/csidh | TestName |Go | C | |------------------|----------|----------| |TestSharedSecret | 57.95774 | 57.91092 | |TestKeyGeneration | 62.23614 | 58.12980 | |TestSharedSecret | 55.28988 | 57.23132 | |TestKeyGeneration | 61.68745 | 58.66396 | |TestSharedSecret | 63.19408 | 58.64774 | |TestKeyGeneration | 62.34022 | 61.62539 | |TestSharedSecret | 62.85453 | 68.74503 | |TestKeyGeneration | 52.58518 | 58.40115 | |TestSharedSecret | 50.77081 | 61.91699 | |TestKeyGeneration | 59.91843 | 61.09266 | |TestSharedSecret | 59.97962 | 62.98151 | |TestKeyGeneration | 64.57525 | 56.22863 | |TestSharedSecret | 56.40521 | 55.77447 | |TestKeyGeneration | 67.85850 | 58.52604 | |TestSharedSecret | 60.54290 | 65.14052 | |TestKeyGeneration | 65.45766 | 58.42823 | On average Go implementation is 2% faster.
2019-11-24 03:39:35 +00:00
func mul576(r *[9]uint64, m1 *fp, m2 uint64) {
var c, h, l uint64
c, r[0] = bits.Mul64(m2, m1[0])
h, l = bits.Mul64(m2, m1[1])
r[1], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[2])
r[2], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[3])
r[3], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[4])
r[4], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[5])
r[5], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[6])
r[6], c = bits.Add64(l, c, 0)
c = h + c
h, l = bits.Mul64(m2, m1[7])
r[7], c = bits.Add64(l, c, 0)
r[8], c = bits.Add64(h, c, 0)
r[8] += c
}
2020-05-14 00:48:43 +01:00
// cswap512 implements constant time swap operation.
// If choice = 0, leave x,y unchanged. If choice = 1, set x,y = y,x.
// If choice is neither 0 nor 1 then behaviour is undefined.
cSIDH-511: (#26) Implementation of Commutative Supersingular Isogeny Diffie Hellman, based on "A faster way to CSIDH" paper (2018/782). * For fast isogeny calculation, implementation converts a curve from Montgomery to Edwards. All calculations are done on Edwards curve and then converted back to Montgomery. * As multiplication in a field Fp511 is most expensive operation the implementation contains multiple multiplications. It has most performant, assembly implementation which uses BMI2 and ADOX/ADCX instructions for modern CPUs. It also contains slower implementation which will run on older CPUs * Benchmarks (Intel SkyLake): BenchmarkGeneratePrivate 6459 172213 ns/op 0 B/op 0 allocs/op BenchmarkGenerateKeyPair 25 45800356 ns/op 0 B/op 0 allocs/op BenchmarkValidate 297 3915983 ns/op 0 B/op 0 allocs/op BenchmarkValidateRandom 184683 6231 ns/op 0 B/op 0 allocs/op BenchmarkValidateGenerated 25 48481306 ns/op 0 B/op 0 allocs/op BenchmarkDerive 19 60928763 ns/op 0 B/op 0 allocs/op BenchmarkDeriveGenerated 8 137342421 ns/op 0 B/op 0 allocs/op BenchmarkXMul 2311 494267 ns/op 1 B/op 0 allocs/op BenchmarkXAdd 2396754 501 ns/op 0 B/op 0 allocs/op BenchmarkXDbl 2072690 571 ns/op 0 B/op 0 allocs/op BenchmarkIsom 78004 15171 ns/op 0 B/op 0 allocs/op BenchmarkFp512Sub 224635152 5.33 ns/op 0 B/op 0 allocs/op BenchmarkFp512Mul 246633255 4.90 ns/op 0 B/op 0 allocs/op BenchmarkCSwap 233228547 5.10 ns/op 0 B/op 0 allocs/op BenchmarkAddRdc 87348240 12.6 ns/op 0 B/op 0 allocs/op BenchmarkSubRdc 95112787 11.7 ns/op 0 B/op 0 allocs/op BenchmarkModExpRdc 25436 46878 ns/op 0 B/op 0 allocs/op BenchmarkMulBmiAsm 19527573 60.1 ns/op 0 B/op 0 allocs/op BenchmarkMulGeneric 7117650 164 ns/op 0 B/op 0 allocs/op * Go code has very similar performance when compared to C implementation. Results from sidh_torturer (4e2996e12d68364761064341cbe1d1b47efafe23) github.com:henrydcase/sidh-torture/csidh | TestName |Go | C | |------------------|----------|----------| |TestSharedSecret | 57.95774 | 57.91092 | |TestKeyGeneration | 62.23614 | 58.12980 | |TestSharedSecret | 55.28988 | 57.23132 | |TestKeyGeneration | 61.68745 | 58.66396 | |TestSharedSecret | 63.19408 | 58.64774 | |TestKeyGeneration | 62.34022 | 61.62539 | |TestSharedSecret | 62.85453 | 68.74503 | |TestKeyGeneration | 52.58518 | 58.40115 | |TestSharedSecret | 50.77081 | 61.91699 | |TestKeyGeneration | 59.91843 | 61.09266 | |TestSharedSecret | 59.97962 | 62.98151 | |TestKeyGeneration | 64.57525 | 56.22863 | |TestSharedSecret | 56.40521 | 55.77447 | |TestKeyGeneration | 67.85850 | 58.52604 | |TestSharedSecret | 60.54290 | 65.14052 | |TestKeyGeneration | 65.45766 | 58.42823 | On average Go implementation is 2% faster.
2019-11-24 03:39:35 +00:00
func cswap512(x, y *fp, choice uint8) {
var tmp uint64
mask64 := 0 - uint64(choice)
for i := 0; i < numWords; i++ {
tmp = mask64 & (x[i] ^ y[i])
x[i] = tmp ^ x[i]
y[i] = tmp ^ y[i]
}
}
// mulRdc performs montgomery multiplication r = x * y mod P.
// Returned result r is already reduced and in Montgomery domain.
func mulRdc(r, x, y *fp) {
var t fp
var c uint64
mulGeneric(r, x, y)
// if p <= r < 2p then r = r-p
t[0], c = bits.Sub64(r[0], p[0], 0)
t[1], c = bits.Sub64(r[1], p[1], c)
t[2], c = bits.Sub64(r[2], p[2], c)
t[3], c = bits.Sub64(r[3], p[3], c)
t[4], c = bits.Sub64(r[4], p[4], c)
t[5], c = bits.Sub64(r[5], p[5], c)
t[6], c = bits.Sub64(r[6], p[6], c)
t[7], c = bits.Sub64(r[7], p[7], c)
var w = uint64(0 - uint64(c))
r[0] = ctPick64(w, r[0], t[0])
r[1] = ctPick64(w, r[1], t[1])
r[2] = ctPick64(w, r[2], t[2])
r[3] = ctPick64(w, r[3], t[3])
r[4] = ctPick64(w, r[4], t[4])
r[5] = ctPick64(w, r[5], t[5])
r[6] = ctPick64(w, r[6], t[6])
r[7] = ctPick64(w, r[7], t[7])
}