NAME¶

mceliece-speed - run some libmceliece speed tests

SYNOPSIS¶

mceliece-speed
    

DESCRIPTION¶

mceliece-speed runs some tests of the speed of the libmceliece implementation of the Classic McEliece cryptosystem.

The mceliece-speed output format is subject to change. Examples of output lines with the current format, from a mceliece-speed run on an Intel Haswell CPU core, include the following:

kem_6960119f selected implementation avx compiler clang -Wall -fPIC -fwrapv -Qunused-arguments -O2 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -mavx -mbmi -mbmi2 -mpopcnt -mavx2 -mtune=haswell; Ubuntu clang version 14.0.0-1ubuntu1.1; Target: x86_64-pc-linux-gnu; Thread model: posix; InstalledDir: /usr/bin
kem_6960119f_keypair selected 1047319 239336111 -45926+75149-74501+77935+25323-18662-14170+21804-50286+79063+119473+24254+32692-86692-11325+11853-1641+85401+79900-36121+31647-46848-82935-77649+50771-61354+25615-19507-213-113509-15330+22405
kem_6960119f_enc selected 194 133429 +8665+1198+718-1682-1798-1955+3726+6311+3449+8172+423+6202-1740+68-2005-2473-2441+137-2698+6077+680-1746+1015-1769+3059+223+609+291+1042-2685-2438+534
kem_6960119f_dec selected 32 282786 +26439+2205+1168+72-226+118-56+122-334+23+193+46+1698+694+168-350+79-209+2-479+15+31-222-273-108+106-205+191-254+70-104-248
    

These lines say that the automatically selected implementation for the 6960119f parameter set was the avx implementation compiled with the clang options shown, taking an interquartile mean (IQM) of 239336111 CPU cycles to generate a key pair, 133429 CPU cycles to encapsulate a session key, and 282786 CPU cycles to decapsulate a session key. The numbers 282786 +26439+2205+... show the timings for a series of consecutive operations: the IQM was 282786 cycles, the first run took 26439 cycles more than the IQM, the second run took 2205 cycles more than the IQM, etc.

SEE ALSO¶

mceliece(1), mceliece(3)