[gambit-list] Comparing Raspberry Pi 2 and 3 B when running Chudnovsky's algorithm for 𝜋; also, some comparisons with GMP

It appears that the pi benchmark in CLN, which uses GMP, uses exactly the same algorithm as chud2.scm, which has been posted here before.

So it’s possible to compare GMP bignum performance with Gambit bignum performance rather directly on both the RPi2B and the RPi3B. Both programs convert the final answer to a string.

On my Raspberry Pi 3 B:

CLN pi with GMP:

sweety-pi:/proc> time pi 1000000 > /dev/null 15.430u 0.130s 0:15.56 100.0% 0+0k 0+0io 0pf+0w

Gambit, configured with:

sweety-pi:~> gsc -v v4.8.4 20160308184356 armv7l-unknown-linux-gnueabihf "./configure 'CC=gcc -mcpu=cortex-a53 -mfpu=neon-vfpv4 -mfloat-abi=hard' '--enable-single-host' '--enable-shared' '--enable-multiple-versions'"

Chudnovsky's algorithm using binary splitting in Gambit Scheme: digits 1000000, CPU time: 55.45.

On my Raspberry Pi 2 B

CLN pi with GMP:

cutie-pi:~> time pi 1000000 > /dev/null 25.560u 0.160s 0:25.75 99.8% 0+0k 0+0io 0pf+0w

Gambit, configured with:

cutie-pi:~> gsc -v v4.8.4 20160308184356 armv7l-unknown-linux-gnueabihf "./configure 'CC=gcc -mcpu=cortex-a7 -mfpu=neon-vfpv4 -mfloat-abi=hard' '--enable-single-host' '--enable-shared' '--enable-multiple-versions' '--disable-gnu-gcc-no-strict-aliasing'"

Chudnovsky's algorithm using binary splitting in Gambit Scheme: digits 1000000, CPU time: 105.64.