Crypto++ 5.5 Benchmarks

Here are speed benchmarks for some of the most commonly used cryptographic algorithms. All were coded in C++, compiled with Microsoft Visual C++ 2005 SP1 (whole program optimization, optimize for speed), and ran on an Intel Core 2 1.83 GHz processor under Windows XP SP 2 in 32-bit mode. x86/MMX/SSE2 assembly language routines were used for integer arithmetic, AES, VMAC, Sosemanuk, Panama, Salsa20, SHA-512, Tiger, and Whirlpool. (OpenMP support was disabled so only one core of the CPU was used for this benchmark.) The results from the same binary running on an Intel Pentium 4 (Northwood) CPU are available here.

Also available are benchmarks that ran on an AMD Opteron 2.4 GHz processor under Linux 2.6.18. Those were compiled with GCC 4.1.1 using -O2 optimization, and x86-64/MMX/SSE2 assembly language routines were used for integer arithmetic, AES, VMAC, Sosemanuk, Panama, Salsa20, and Whirlpool.

Algorithm	MiB/Second	Cycles Per Byte	Microseconds to Setup Key and IV	Cycles to Setup Key and IV
VMAC(AES)-64	1489	1.2	3.066	5611
VMAC(AES)-128	794	2.2	3.597	6582
HMAC(SHA-1)	152	11.5	0.546	999
Two-Track-MAC	106	16.4	0.042	77
CBC-MAC/AES	86	20.2	0.287	525
DMAC/AES	82	21.3	1.739	3183
CRC-32	256	6.8
Adler-32	936	1.9
MD5	258	6.8
SHA-1	155	11.3
SHA-256	81	21.5
SHA-512	99	17.6
Tiger	217	8.0
Whirlpool	58	30.0
RIPEMD-160	108	16.1
RIPEMD-320	111	15.8
RIPEMD-128	155	11.3
RIPEMD-256	159	11.0
Panama-LE	851	2.1	1.605	2937
Panama-BE	405	4.3	2.639	4830
Salsa20	318	5.5	0.309	565
Salsa20/12	486	3.6	0.408	747
Salsa20/8	669	2.6	0.406	743
Sosemanuk	739	2.4	1.115	2040
MARC4	128	13.6	3.800	6954
SEAL-3.0-BE	372	4.7	60.103	109988
SEAL-3.0-LE	377	4.6	60.133	110043
WAKE-OFB-BE	206	8.5	2.790	5106
WAKE-OFB-LE	200	8.7	2.778	5084
AES/ECB (128-bit key)	99	17.7	0.248	454
AES/ECB (192-bit key)	86	20.2	0.242	443
AES/ECB (256-bit key)	77	22.6	0.312	572
AES/CTR (128-bit key)	96	18.1	0.620	1134
AES/OFB (128-bit key)	83	21.1	0.564	1033
AES/CFB (128-bit key)	69	25.3	0.746	1366
AES/CBC (128-bit key)	84	20.9	0.431	789
Camellia/ECB (128-bit key)	49	35.6	0.174	318
Camellia/ECB (256-bit key)	38	46.0	0.269	492
Twofish	62	28.1	7.077	12950
Serpent	33	53.3	0.736	1346
CAST-256	38	45.6	2.185	3998
RC6	103	16.9	2.379	4353
MARS	74	23.7	2.924	5351
SHACAL-2/ECB (128-bit key)	55	31.7	0.411	752
SHACAL-2/ECB (512-bit key)	55	32.0	0.436	798
DES	34	51.1	6.975	12765
DES-XEX3	28	61.4	7.128	13045
DES-EDE3	13	131.2	25.240	46190
IDEA	36	48.6	0.220	403
RC5 (r=16)	82	21.3	1.989	3640
Blowfish	61	28.6	61.035	111694
TEA/ECB	27	63.9	0.183	334
XTEA/ECB	27	64.3	0.183	335
CAST-128	57	30.5	0.497	910
SKIPJACK	10	173.8	2.991	5474

Operation	Milliseconds/Operation	Megacycles/Operation
RSA 1024 Encryption	0.07	0.13
RSA 1024 Decryption	1.52	2.78
LUC 1024 Encryption	0.07	0.13
LUC 1024 Decryption	2.32	4.25
DLIES 1024 Encryption	0.83	1.53
DLIES 1024 Encryption with precomputation	1.45	2.65
DLIES 1024 Decryption	3.06	5.60
LUCELG 512 Encryption	0.57	1.05
LUCELG 512 Encryption with precomputation	0.57	1.05
LUCELG 512 Decryption	0.62	1.13
RSA 2048 Encryption	0.15	0.28
RSA 2048 Decryption	5.95	10.89
LUC 2048 Encryption	0.17	0.32
LUC 2048 Decryption	9.71	17.77
DLIES 2048 Encryption	4.26	7.79
DLIES 2048 Encryption with precomputation	4.69	8.59
DLIES 2048 Decryption	18.84	34.48
LUCELG 1024 Encryption	1.86	3.40
LUCELG 1024 Encryption with precomputation	1.89	3.45
LUCELG 1024 Decryption	1.67	3.05
RSA 1024 Signature	1.42	2.60
RSA 1024 Verification	0.07	0.13
RW 1024 Signature	2.01	3.67
RW 1024 Verification	0.04	0.08
LUC 1024 Signature	2.37	4.34
LUC 1024 Verification	0.07	0.13
NR 1024 Signature	0.47	0.86
NR 1024 Signature with precomputation	0.41	0.75
NR 1024 Verification	0.50	0.92
NR 1024 Verification with precomputation	0.64	1.17
DSA 1024 Signature	0.47	0.85
DSA 1024 Signature with precomputation	0.41	0.76
DSA 1024 Verification	0.52	0.95
DSA 1024 Verification with precomputation	0.66	1.21
LUC-HMP 512 Signature	0.60	1.09
LUC-HMP 512 Signature with precomputation	0.59	1.08
LUC-HMP 512 Verification	0.62	1.13
LUC-HMP 512 Verification with precomputation	0.67	1.22
ESIGN 1023 Signature	0.21	0.38
ESIGN 1023 Verification	0.07	0.12
ESIGN 1536 Signature	0.38	0.69
ESIGN 1536 Verification	0.14	0.25
RSA 2048 Signature	5.95	10.89
RSA 2048 Verification	0.15	0.28
RW 2048 Signature	7.27	13.31
RW 2048 Verification	0.09	0.17
LUC 2048 Signature	9.71	17.77
LUC 2048 Verification	0.17	0.32
NR 2048 Signature	2.11	3.87
NR 2048 Signature with precomputation	0.93	1.71
NR 2048 Verification	2.35	4.31
NR 2048 Verification with precomputation	1.44	2.63
LUC-HMP 1024 Signature	1.87	3.42
LUC-HMP 1024 Signature with precomputation	1.88	3.45
LUC-HMP 1024 Verification	1.92	3.52
LUC-HMP 1024 Verification with precomputation	1.91	3.50
ESIGN 2046 Signature	0.46	0.84
ESIGN 2046 Verification	0.15	0.27
XTR-DH 171 Key-Pair Generation	0.77	1.40
XTR-DH 171 Key Agreement	1.52	2.77
XTR-DH 342 Key-Pair Generation	1.84	3.36
XTR-DH 342 Key Agreement	3.72	6.80
DH 1024 Key-Pair Generation	0.44	0.80
DH 1024 Key-Pair Generation with precomputation	0.74	1.36
DH 1024 Key Agreement	1.15	2.10
DH 2048 Key-Pair Generation	2.13	3.91
DH 2048 Key-Pair Generation with precomputation	2.38	4.35
DH 2048 Key Agreement	3.77	6.91
LUCDIF 512 Key-Pair Generation	0.30	0.54
LUCDIF 512 Key-Pair Generation with precomputation	0.30	0.54
LUCDIF 512 Key Agreement	0.62	1.13
LUCDIF 1024 Key-Pair Generation	0.94	1.71
LUCDIF 1024 Key-Pair Generation with precomputation	0.98	1.79
LUCDIF 1024 Key Agreement	1.68	3.07
MQV 1024 Key-Pair Generation	0.42	0.77
MQV 1024 Key-Pair Generation with precomputation	0.39	0.71
MQV 1024 Key Agreement	0.86	1.57
MQV 2048 Key-Pair Generation	2.09	3.83
MQV 2048 Key-Pair Generation with precomputation	0.91	1.66
MQV 2048 Key Agreement	3.85	7.04
ECIES over GF(p) 256 Encryption	5.57	10.19
ECIES over GF(p) 256 Encryption with precomputation	4.12	7.53
ECIES over GF(p) 256 Decryption	3.91	7.15
ECNR over GF(p) 256 Signature	2.82	5.17
ECNR over GF(p) 256 Signature with precomputation	2.13	3.90
ECNR over GF(p) 256 Verification	8.33	15.25
ECNR over GF(p) 256 Verification with precomputation	3.60	6.58
ECDHC over GF(p) 256 Key-Pair Generation	2.82	5.15
ECDHC over GF(p) 256 Key-Pair Generation with precomputation	2.06	3.78
ECDHC over GF(p) 256 Key Agreement	2.85	5.21
ECMQVC over GF(p) 256 Key-Pair Generation	2.79	5.11
ECMQVC over GF(p) 256 Key-Pair Generation with precomputation	2.07	3.80
ECMQVC over GF(p) 256 Key Agreement	8.55	15.64
ECIES over GF(2^n) 233 Encryption	21.51	39.35
ECIES over GF(2^n) 233 Encryption with precomputation	6.33	11.58
ECIES over GF(2^n) 233 Decryption	12.12	22.18
ECNR over GF(2^n) 233 Signature	10.70	19.57
ECNR over GF(2^n) 233 Signature with precomputation	3.13	5.72
ECNR over GF(2^n) 233 Verification	13.16	24.08
ECNR over GF(2^n) 233 Verification with precomputation	5.48	10.03
ECDHC over GF(2^n) 233 Key-Pair Generation	10.64	19.47
ECDHC over GF(2^n) 233 Key-Pair Generation with precomputation	3.06	5.60
ECDHC over GF(2^n) 233 Key Agreement	10.50	19.22
ECMQVC over GF(2^n) 233 Key-Pair Generation	10.64	19.47
ECMQVC over GF(2^n) 233 Key-Pair Generation with precomputation	3.18	5.82
ECMQVC over GF(2^n) 233 Key Agreement	13.51	24.73

Notes

Crypto++ 5.5 Benchmark Average for this platform: 395.11. This number is computed by taking the geometric mean of the number of MB/second of each cipher, hash function, and MAC, and operations/second of each asymmetric operation listed above.
RSA and LUC use 17 as the public exponent.
DH and ElGamal encryption and decryption use short exponents to save time. The size of the secret exponents were chosen so that a meet-in-the-middle attack would be slower than the general discrete log algorithm (NFS). The sizes used were:

modulus exponent
1024 164
2048 226
EC means elliptic curve. Operations in GF(2^n) are implemented using trinomial basis. Note that compared to other algorithms listed here, Crypto++'s implementation of EC over GF(2^n) is less optimized.
All tests were done by repeating the crypto operations over small blocks of random data. In practice you will likely see slower speeds because time is needed to transfer data to and from memory.
For the ciphers that specify big endian byte order, the timing data listed include time needed to convert to and from little endian order.
The RSA, RW, DH, MQV, and elliptic curve schemes come from the IEEE P1363 standard. For more info see http://grouper.ieee.org/groups/1363/index.html.
Precomputation means using a table of 16 precomputed powers of each fixed base to speed up exponentiation.
Source code for these benchmarks is available as part of the Crypto++ library.

modulus	exponent
1024	164
2048	226

Written by: Wei Dai Last modified: 5/5/2007