Rob Pike
e3634aad9d
test/bench: update timings; moving to new machine.
...
Interesting comparisons between old and new machine,
and relationship between gccgo and gc.
R=golang-dev, rsc1
CC=golang-dev
https://golang.org/cl/4430045
2011-04-18 09:50:20 -07:00
Rob Pike
65ece70812
test/bench: update timings for new GC
...
Some significant improvements; e.g. pidigits goes from 8.33 to 6.37
Also add gccgo for regex-dna.
R=rsc
CC=golang-dev
https://golang.org/cl/4160056
2011-02-16 10:46:29 -08:00
Rob Pike
aa78cec69c
test/bench: update numbers for regex-dna after speedup to regexp
...
R=rsc
CC=golang-dev
https://golang.org/cl/3634042
2010-12-16 12:38:14 -08:00
Anthony Martin
85f5bb8216
test/bench: a couple fixes to timing.sh
...
- Skip the gccgo tests if we don't have it
- Add -lm so nbody.c will compile
- Pass 2098 to the meteor test (cf. the shootout site)
R=rsc, r, iant
CC=golang-dev
https://golang.org/cl/3619042
2010-12-15 09:47:13 -05:00
Rob Pike
f57f8b6f68
test/bench: update numbers
...
R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/2631041
2010-10-20 17:43:57 -07:00
Rob Pike
7a5852b50e
update timing logs for benchmarks.
...
some significant improvements:
reverse-complement
nbody (gccgo only)
regex-dna
pidigits
one mysterious drop:
binary-tree.go (gccgo only)
R=rsc
CC=golang-dev
https://golang.org/cl/181158
2010-01-06 19:36:28 -08:00
Rob Pike
eb33c0bb35
timing log update for new version of chameneos
...
R=rsc
CC=golang-dev, rog
https://golang.org/cl/160054
2009-11-24 14:19:43 -08:00
Rob Pike
0263103a05
move evaluation of null-matching instructions one iteration earlier.
...
performance hit of about 20% but more intuitive results for submatches.
we need a good regexp package at some point.
Fixes #110 .
R=rsc
CC=golang-dev
https://golang.org/cl/152131
2009-11-14 12:23:24 -08:00
Rob Pike
f448057b8a
Sqrt in assembler is worth 3-4X for nbody
...
R=rsc
DELTA=10 (10 added, 0 deleted, 0 changed)
OCL=34987
CL=34987
2009-09-24 14:47:34 -07:00
Rob Pike
b9493ded12
updated numbers after more slice inlining
...
R=rsc
DELTA=57 (57 added, 0 deleted, 0 changed)
OCL=34896
CL=34901
2009-09-22 14:13:40 -07:00
Rob Pike
7f0dd8b1b7
new benchmark results.
...
main changes:
gccgo is much slower on binary-tree and fannkuch
6g seems to struggle with alignment of loops in nbody (dignosis uncertain)
6g is a little faster at regexps and concurrency
R=rsc
DELTA=93 (93 added, 0 deleted, 0 changed)
OCL=34320
CL=34322
2009-09-03 11:41:01 -07:00
Robert Griesemer
0dbd8971a2
- use in-place bignum operations where available
...
- runs approx. 30% faster
R=r
DELTA=24 (10 added, 2 deleted, 12 changed)
OCL=32984
CL=33005
2009-08-10 17:44:46 -07:00
Rob Pike
ea8197cb45
new timing data after compiler updates (fp regs, integer div and mod)
...
R=rsc
DELTA=90 (90 added, 0 deleted, 0 changed)
OCL=32993
CL=33001
2009-08-10 17:04:59 -07:00
Rob Pike
c12ccabb9c
chameneos
...
R=rsc
DELTA=514 (513 added, 0 deleted, 1 changed)
OCL=32898
CL=32910
2009-08-07 15:28:46 -07:00
Russ Cox
0ab8dbef76
annotations
...
TBR=r
OCL=32896
CL=32896
2009-08-07 14:18:30 -07:00
Rob Pike
5005fb46a2
simplify threadring slightly; slight speed improvement
...
R=rsc
DELTA=7 (3 added, 3 deleted, 1 changed)
OCL=32885
CL=32889
2009-08-07 13:30:20 -07:00
Rob Pike
c30d81bd23
threadring
...
more interesting than most
R=rsc
DELTA=132 (131 added, 0 deleted, 1 changed)
OCL=32876
CL=32881
2009-08-07 12:53:51 -07:00
Rob Pike
e98dde27ef
timing updates after 6g change for efficient division by powers of two
...
R=rsc
DELTA=25 (23 added, 0 deleted, 2 changed)
OCL=32873
CL=32875
2009-08-07 10:39:45 -07:00
Rob Pike
2f9c48f6d1
timings for pidigits
...
TBR=rsc
OCL=32857
CL=32857
2009-08-06 18:27:15 -07:00
Rob Pike
7655588438
meteor-contest
...
R=rsc
DELTA=1276 (1275 added, 0 deleted, 1 changed)
OCL=32851
CL=32854
2009-08-06 18:04:48 -07:00
Rob Pike
c68398617b
mandelbrot
...
R=rsc
DELTA=147 (145 added, 0 deleted, 2 changed)
OCL=32840
CL=32845
2009-08-06 14:21:21 -07:00
Rob Pike
ae3939cb78
k-nucleotide
...
R=rsc
DELTA=367 (366 added, 0 deleted, 1 changed)
OCL=32832
CL=32836
2009-08-06 13:00:26 -07:00
Rob Pike
158f3f6440
parallel spectral-norm
...
R=rsc
DELTA=85 (84 added, 1 deleted, 0 changed)
OCL=32810
CL=32810
2009-08-05 18:44:49 -07:00
Rob Pike
8764ebee95
spectral-norm
...
make regexp-dna use bytes not strings (no significant timing change)
R=rsc
DELTA=149 (138 added, 1 deleted, 10 changed)
OCL=32804
CL=32807
2009-08-05 17:25:38 -07:00
Rob Pike
d9a0bc9b58
regex-dna
...
R=rsc
DELTA=243 (242 added, 0 deleted, 1 changed)
OCL=32786
CL=32791
2009-08-05 13:03:46 -07:00
Rob Pike
6b7bf38487
fannkuch
...
R=rsc
DELTA=240 (239 added, 0 deleted, 1 changed)
OCL=32783
CL=32785
2009-08-05 11:33:59 -07:00
Rob Pike
3a23e53387
binary tree
...
R=rsc
DELTA=324 (323 added, 0 deleted, 1 changed)
OCL=32759
CL=32768
2009-08-04 19:38:08 -07:00
Rob Pike
abaf604648
add info about Sqrt instruction
...
couple of fixes to timing.sh
R=rsc
DELTA=10 (5 added, 0 deleted, 5 changed)
OCL=32742
CL=32756
2009-08-04 17:31:19 -07:00
Rob Pike
5b1fb14371
nbody benchmark
...
timing.sh improvements
R=rsc
DELTA=334 (319 added, 0 deleted, 15 changed)
OCL=32734
CL=32736
2009-08-04 14:27:46 -07:00
Rob Pike
0444da91c0
improve myrandom() in fasta.
...
add script to automate timing
R=rsc
DELTA=78 (68 added, 0 deleted, 10 changed)
OCL=32729
CL=32732
2009-08-04 14:07:17 -07:00
Rob Pike
9ae42ee87f
another benchmark: reverse-complement
...
R=rsc
DELTA=682 (511 added, 171 deleted, 0 changed)
OCL=32725
CL=32727
2009-08-04 13:04:37 -07:00
Rob Pike
417683c3d3
clean up the code a bit
...
start a log of progress
R=rsc
DELTA=222 (185 added, 17 deleted, 20 changed)
OCL=32701
CL=32718
2009-08-04 10:13:58 -07:00