Common LISP benchmarking
This requires much further investigation, but as an example,
using sbcl --eval
with the following…
(progn (ql:quickload :the-cost-of-nothing) (asdf:test-system :the-cost-of-nothing) (quit))
… gave me such puzzling numbers:
my-computer | linode | advin | |
---|---|---|---|
implementation, sbcl | 2.1.9.23-3614a905b | 1.4.16.debian | 2.1.1.debian |
machine | Ryzen 7 3800X 8c | EPYC 7601 32c | Ryzen 9 5900X 12c |
alloc a cons cell | 2.89 ns | 35.16 ns | -0.00 |
GC | 2.91 ms | 3.57 ms | 1.07 ms |
full GC | 60.42 ms | 40.31 ms | 11.34 ms |
FUNCALL 0 mand. args | 1.79 ns | 5.89 ns | 0 |
FUNCALL 1 mand. arg | 1.79 ns | 11.66 ns | 1.55 ns |
FUNCALL 2 mand. args | 2.03 ns | 5.36 ns | 0 |
FUNCALL 3 mand. args | 1.79 ns | 6.54 ns | 0 |
FUNCALL 4 mand. args | 2.02 ns | 6.26 ns | 0 |
FUNCALL 5 mand. args | 2.07 ns | 4.33 ns | 0 |
FUNCALL 6 mand. args | 2.02 ns | 4.55 ns | 0 |
FUNCALL 0 opt. args | 1.79 ns | 5.89 ns | 0 |
FUNCALL 1 opt. arg | 2.02 ns | 6.74 ns | 0 |
FUNCALL 2 opt. args | 2.74 ns | 4.58 ns | 1.89 ns |
FUNCALL 3 opt. args | 2.50 ns | 4.53 ns | 676.41 ps |
FUNCALL 4 opt. args | 2.26 ns | 4.84 ns | 0 |
FUNCALL 5 opt. args | 2.51 ns | 6.06 ns | 0 |
FUNCALL 6 opt. args | 2.50 ns | 5.06 ns | 0 |
FUNCALL 0 keyw. args | 1.79 ns | 5.89 ns | 0 |
FUNCALL 1 keyw. arg | 4.24 ns | 7.29 ns | 0 |
FUNCALL 2 keyw. args | 5.63 ns | 12.44 ns | 0 |
FUNCALL 3 keyw. args | 7.15 ns | 14.61 ns | 0 |
FUNCALL 4 keyw. args | 8.59 ns | 18.71 ns | 0 |
FUNCALL 5 keyw. args | 11.49 ns | 21.70 ns | 1.84 ns |
FUNCALL 6 keyw. args | 13.07 ns | 32.53 ns | 2.27 ns |
FUNCALL 0 rest args | 1.79 ns | 5.89 ns | 0 |
FUNCALL 1 rest arg | 1.55 ns | 3.27 ns | 0 |
FUNCALL 2 rest args | 1.79 ns | 6.46 ns | 152.55 ps |
FUNCALL 3 rest args | 1.79 ns | 4.62 ns | 248.34 ps |
FUNCALL 4 rest args | 2.02 ns | 5.41 ns | 1.40 ns |
FUNCALL 5 rest args | 1.78 ns | 3.97 ns | 1.66 ns |
FUNCALL 6 rest args | 2.04 ns | 4.78 ns | 0 |
Flops (single-float): | 4.64 giga | 2.27 giga | 8.14 giga |
Flops (double-float): | 4.63 giga | 2.14 giga | 8.30 giga |