RSA KeyGeneration slower than OpenPGP

[jochenonline]

When comparing RSA-Key-Generation between asmcrypto and OpenPGP on the same machine and in same browser and with the same bit-length (4096) OpenPGP is almost twice as fast as asmcrypto (average time 13,3 sec vs 20,2 sec). Do you see any obvious reason for that or am I comparing apples with oranges?

[vibornoff]

Is your result statistically significant? I mean did you calculate Standard deviation along with the average time?

[jochenonline]

Standard deviations:

• asmcrypto: 9,7
• OpenPGP: 2,2

So the times of OpenPGP are also much more stable. Theses are the times we have measured:

asmcrypto:

• 35,7
• 25,7
• 9,6
• 21
• 11,7
• 5,1
• 32,8
• 16,2
• 19,5
• 29,3

OpenPGP:

• 12,7
• 11,6
• 10,8
• 13,9
• 11,2
• 15,2
• 15,9
• 17,1
• 10,4
• 13,8

[vibornoff]

Having a brief look to the opengpg.js source I suppose it's because your browser supports WebCrypto API :)