[MLton-devel] profiling benchmarks
Stephen Weeks
MLton@mlton.org
Thu, 19 Dec 2002 18:02:40 -0800
Here are the benchmarks run with the new profiling scheme that I
checked in today. The overheads are acceptable (to me anyways) except
on peek, which I will look into.
MLton0 -- mlton -profile no
MLton1 -- mlton -profile alloc
MLton2 -- mlton -profile time
run time ratio
benchmark MLton1 MLton2
barnes-hut 1.03 1.04
boyer 0.92 0.87
checksum 1.00 0.98
count-graphs 1.35 1.11
DLXSimulator 1.13 1.04
fft 0.95 0.99
fib 1.48 1.49
hamlet 1.17 1.06
imp-for 1.00 1.36
knuth-bendix 1.18 1.13
lexgen 1.06 1.01
life 1.27 1.19
logic 1.09 1.05
mandelbrot 1.09 1.09
matrix-multiply 1.06 1.06
md5 1.20 1.32
merge 1.08 1.04
mlyacc 1.13 1.04
model-elimination 1.19 1.07
mpuz 1.07 1.10
nucleic 1.13 1.11
peek 5.44 5.34
psdes-random 0.98 1.04
ratio-regions 1.04 1.02
ray 1.13 1.03
raytrace 1.05 1.01
simple 1.17 1.07
smith-normal-form 1.00 0.99
tailfib 0.81 0.81
tak 1.51 1.51
tensor 0.97 0.98
tsp 1.01 0.99
tyan 1.20 1.07
vector-concat 1.00 0.99
vector-rev 1.03 1.04
vliw 1.18 1.06
wc-input1 1.08 1.07
wc-scanStream 1.18 1.31
zebra 1.27 0.96
zern 0.98 0.99
size
benchmark MLton0 MLton1 MLton2
barnes-hut 115,376 132,530 149,122
boyer 137,959 168,057 191,569
checksum 48,647 55,353 59,585
count-graphs 66,599 80,513 90,625
DLXSimulator 105,760 153,730 184,602
fft 57,379 66,181 75,125
fib 48,703 55,409 59,769
hamlet 1,238,304 1,884,464 2,318,368
imp-for 48,671 54,945 59,601
knuth-bendix 90,056 115,114 133,962
lexgen 169,549 230,901 277,869
life 68,815 77,217 84,705
logic 110,407 131,161 144,841
mandelbrot 48,791 55,249 59,609
matrix-multiply 49,223 56,217 60,777
md5 57,656 67,714 76,362
merge 50,055 56,849 61,441
mlyacc 510,397 645,613 779,613
model-elimination 627,808 883,322 1,056,210
mpuz 53,671 61,625 68,073
nucleic 196,319 207,697 214,969
peek 55,624 64,954 71,466
psdes-random 49,423 56,057 60,601
ratio-regions 67,215 87,617 110,705
ray 111,440 137,400 160,368
raytrace 278,925 315,029 346,445
simple 205,099 323,821 378,517
smith-normal-form 189,612 201,510 213,454
tailfib 48,479 54,793 59,089
tak 48,855 55,601 59,985
tensor 113,099 128,973 147,557
tsp 62,808 72,962 85,082
tyan 109,872 143,058 171,842
vector-concat 49,831 56,785 61,321
vector-rev 49,023 55,649 59,961
vliw 324,745 542,841 686,665
wc-input1 70,621 77,661 86,981
wc-scanStream 71,373 78,405 87,813
zebra 157,104 180,322 222,026
zern 54,650 63,836 70,540
compile time
benchmark MLton0 MLton1 MLton2
barnes-hut 3.84 4.11 4.98
boyer 8.00 9.26 9.21
checksum 1.03 1.28 1.38
count-graphs 2.55 2.89 3.00
DLXSimulator 5.98 7.67 8.39
fft 2.06 2.55 2.70
fib 0.97 1.26 1.28
hamlet 85.63 102.69 170.09
imp-for 1.03 1.30 1.38
knuth-bendix 3.47 4.14 4.34
lexgen 8.63 10.56 11.14
life 2.29 2.49 2.40
logic 4.56 5.82 6.01
mandelbrot 1.00 1.25 1.30
matrix-multiply 1.08 1.35 1.45
md5 1.75 2.11 2.27
merge 0.97 1.26 1.32
mlyacc 35.04 36.78 39.87
model-elimination 35.01 41.18 49.12
mpuz 1.32 1.57 1.70
nucleic 63.86 64.20 64.58
peek 1.45 1.79 1.87
psdes-random 1.05 1.29 1.36
ratio-regions 3.49 4.52 4.87
ray 5.55 6.39 6.48
raytrace 18.04 18.06 18.52
simple 10.07 14.47 14.98
smith-normal-form 10.20 10.96 11.11
tailfib 0.92 1.16 1.25
tak 0.92 1.20 1.25
tensor 4.09 4.74 5.08
tsp 2.31 2.70 2.82
tyan 5.44 6.46 6.73
vector-concat 1.06 1.33 1.39
vector-rev 0.96 1.21 1.30
vliw 20.28 28.08 31.83
wc-input1 2.33 2.47 2.63
wc-scanStream 2.37 2.55 2.67
zebra 9.96 10.48 10.64
zern 1.60 1.98 2.10
-------------------------------------------------------
This SF.NET email is sponsored by: Geek Gift Procrastinating?
Get the perfect geek gift now! Before the Holidays pass you by.
T H I N K G E E K . C O M http://www.thinkgeek.com/sf/
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel