For fun, I've successfully installed J on every computer / device laying
around the house and the two virtual private servers I run. That's 14
different instances of J.
I ran a modified version of the Joey Tuttle benchmark. Mine was 10 (6!:2)
'%. ?500 500$10000' to slow it down some on the faster devices.
In all cases it was J8 except where noted I also ran jconsole_for_android
My results (times in seconds):
Laptops:
1.) 0.330 2.8ghz i-7 2640M Dell
2.) 0.336 2.9ghz i-7 3520M Lenovo
3.) 0.427 2ghz Dell 1570
4.) 0.770 1.9ghz AMD T-58 Dell
5.) 2.520 1.6ghz Atom N55 Chromebook Cr48
Tablets:
first time is for jqt. second time is for jconsole_for_apk (jc)
5.) 1.269 iPad 4
6.) 1.823 Asus T-100 1.3ghz baytrail
7.) 13.73 jqt, 2.33 jc Nexus 4
8.) 31.42 jqt, 9.659 jc Kindle Fire v1
Phones:
9.) 4.2s linux chroot Droid 3
9.0 jconsole Droid 3
10.) 13.1 jqt, 3.33 jc Samsung Galaxy S4
11.) 90.1 jc LG Ally (old!)
VPS
12.) 0.96 2ghz , 4K bogomips Digital Ocean
13.) 0.58 2.2ghz 4k bogomips Prgmr.com
Other
14.) 16.56 Raspberry PI
Now, if I could only cluster them all together! (kidding)
Question: Is jqt for android compiled differently than jconsole? I noticed
significantly different times on all devices. Performance probably doesn't
matter so much on the tablet and phone but wanted to point it out
nonetheless.
For #9, I ran the first benchmark under a chrooted debian and the armhf
debian package if I remember right. I rooted my device and installed debian
on it through https://github.com/meefik/linuxdeploy. Running under a
chrooted linux instance was 2x as fast as the jconsole for android, which I
thought was interesting.
Forum links (many others) that reference the benchmark
[1] -
http://www.jsoftware.com/pipermail/programming/2007-December/009018.html
[2] - http://www.jsoftware.com/pipermail/beta/2012-May/006113.html
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
