PDA

View Full Version : hwbot multipi - multithreaded superpi (open beta)



RichBa5tard
08-14-2006, 12:32 PM
Hi guys,

I wrote a superpi app that calculates pi, but unlike superpi it dedects how many processor cores you have, and spreads the calculation over those available.

It works pretty well, but I need some feedback and comments on the length of calculation and stuff. It takes pretty long on a midrange processor (about 3 minutes on a 2ghz A64), but as it takes use of multiple cores, calculation on a decent dual core proc is about half. I think a 5ghz conroe may hit 30 seconds.

Some testruns i did:
Intel Pentium M 715 (1.5Ghz): 3m 18.815s
T2500 (mac osx), 2 cores: 1m 47.127s
T2500 (mac osx), 1 core: 3m 08.163s

I know some guys are going to nag because it's written in java, and you need a java vm installed, but as mac/linux user i'm tired of seeing windows-only benchmarks. This is one we can all enjoy. :)

If interested, you can betatest multipi v0.1. It runs on linux, mac, win32 and win64, but requires java 5 (1.5) to be installed. Just unzip and run multipi.bat or multipi.sh.

Comments and thoughts can be posted here:
http://www.hwbot.org/forums/viewtopic.php?id=196

Download:
http://www.hwbot.org/download/multipi-0.2.zip

\Karting_freak
08-14-2006, 12:47 PM
heh
happily my kentsfield arrived in time :)

RichBa5tard
08-14-2006, 12:53 PM
I'd love to see how a kentsfield would perform. Post yer results please. :)

\Karting_freak
08-14-2006, 01:39 PM
i'll post 4sure.
need to get some hw first :)

RichBa5tard
08-14-2006, 01:53 PM
Some interesting benches at OC Forums:
http://www.ocforums.com/showthread.php?t=472068

Core Duo T2500 1 core: 3m08s
A64 2.2Ghz: 3min
Core Duo T2500 2 cores: 1m45s
Woodcrest 3Ghz 4 cores: 32s

Kunaak
08-14-2006, 01:54 PM
51 seconds here.
nothing fancy here.
DFI 975X.
1 gig of corsair 5400UL cas 4-3-3-10, 670 mhz.
E6400 at 3.6 ghz.

Kunaak
08-14-2006, 02:01 PM
seems really reliable.
I only get a .05 second variation in time, when run over and over.

how about a really long version to play with?

like one that take 45 minutes or so?
32M is alittle too easy these days, since it's time has been cut in half with the introduction of these conroes and such.

RichBa5tard
08-14-2006, 02:02 PM
Hmmm, what do you think, more iterations by default? A quad core can do it 30 seconds without too much tweaking, maybe that's too easy.

An average A64 need 3 minutes though. If I make it much harder it will take a long time on slower systems... and a part of the success of superpi is the time it takes to get your result.



- edit -
45 minutes on what? A single core, dual core or quad core?

Kunaak
08-14-2006, 02:05 PM
well, how about alittle interactive menu then?

hit 1 for default.
hit 2 for long test.

then A64's have thier standard run, and Conroes have something alittle more challenging to play with.

RichBa5tard
08-14-2006, 02:11 PM
I will certainly add that option. The finished version will offer a simple gui, like superpi has. This is just a beta to test the algorithm.

Kunaak
08-14-2006, 02:12 PM
I look forward to trying it :)

[cTx]Philosophy
08-14-2006, 03:25 PM
SWEETTT!!! this is what ive always wanted to be compiled, this will give a far better show of power for dualies and quads..
And it begins..
Thnx, downloading it in a sec.....
I got backdoor trojan virus warning when I opened it, is this normal?

Opty 165 @ 2.9
2X1024 PDP @ 243 2.5-3-2-5 1T

k0nsl
08-14-2006, 03:43 PM
Pi calculated in: 03m 07.953s

Pentium D 805 @ 3.7GHz

:/

-k0nsl

Avman
08-14-2006, 06:07 PM
57 seconds - A64 X2 3800+ at 3.0Ghz

m0da
08-14-2006, 09:08 PM
02m 40.703s
specs in sig, ie/aim/services running in bg

dinos22
08-14-2006, 09:14 PM
it's pretty short in length..........try and make it at least 10x longer i reckon

dinos22
08-14-2006, 09:26 PM
Allendale @ 3.33GHz i run 24/7


http://img225.imageshack.us/img225/3746/untitleddl4.jpg (http://img225.imageshack.us/my.php?image=untitleddl4.jpg)

alpha0ne
08-14-2006, 10:17 PM
Nice proggy, many thanks RichBa5tard :toast:

http://i16.photobucket.com/albums/b29/alpha0ne/pi.jpg

m0da
08-14-2006, 10:36 PM
57 seconds - A64 X2 3800+ at 3.0Ghz

:eek: beat an e6600 @same clock by 4s!! how!!?

dinos22
08-14-2006, 10:37 PM
yeah there's something wrong there

RichBa5tard
08-15-2006, 12:42 AM
Pi calculated in: 03m 07.953s

Pentium D 805 @ 3.7GHz

:/

-k0nsl

3m08s should be on the spot if it where only using 1 processor, but a 805 has 2 so your time should be ~ 1m40s. Can you verify it used both cores? At the beginning of the calculation, it should say 'dedected 2 processor cores' and 'started calculation core 1' and 'started calculation core2'.

nn_step
08-15-2006, 12:48 AM
Thank you dear sir. Though I wish could look at the source to perhaps make a C version. So that it can run on the most stripped down OS

RichBa5tard
08-15-2006, 12:50 AM
SWEETTT!!! this is what ive always wanted to be compiled, this will give a far better show of power for dualies and quads..
And it begins..
Thnx, downloading it in a sec.....
I got backdoor trojan virus warning when I opened it, is this normal?

Not normal no. I think it's your virusscanner which is set to warn you when running bat files.

If your java vm is properly installed, you can also just double-click the jar file.

RichBa5tard
08-15-2006, 12:54 AM
Thank you dear sir. Though I wish could look at the source to perhaps make a C version. So that it can run on the most stripped down OS

Not being able to run on the most stripdown os is a drawback, i agree. But, i find the ability to run it too on a mac / linux / freebsd far more important, so i'm going to keep the program in java.

RichBa5tard
08-15-2006, 02:58 AM
Some early conclusions:
- current amount of iterations seems ok, maybe it needs to be a bit more.
- dual / quad core is almost 2/4 times as fast as single core... maybe tax the memory a bit more so it doesn't get a cpu-only test?
- memory gets barely taxed, heavier use on memory would make it a 'better' test (for dedecting memory faults), lower the advantage of dual/quad core a bit, and make it a more 'real life' test.
- finding the sweet spot between 'not too short on a oc'ed quad core' and 'not too boring on a midrange single core' is frigging hard. : )
- p4 dual core seems to perform equal to single core? need verification of this
- x2 equal as fast as conroe at same speed? need verification

RichBa5tard
08-15-2006, 06:35 AM
I made a second beta (0.2). Download link will be available shortly. Results will _not_ be comparable to beta 0.1.

I altered the algoritm to consume a lot more memory, and to check the memory for faults. The standard calculation now uses approx 200mb of ram.

The heavier use on memory makes the bonus for having additional cores smaller. I think a second core would give a +50% bonus, a 3th and 4th maybe 20 to 30%. The faster your memory, the more benefit you'll get from having additional cores.

Testrun on my macbook pro:
T2500 (1.86Ghz Core Duo) with 1 core: 3m 07.140s
T2500 (1.86Ghz Core Duo) with 2 cores: 2m 02.754s

download link:
http://www.hwbot.org/download/multipi-0.2.zip

Input highly appreciated!

Cranox
08-15-2006, 01:08 PM
2min 15 sec - A64 3800+ venice at 210 x 12 =2520 mhz ( 24/7 settings )

RichBa5tard
08-15-2006, 03:42 PM
A bit slower than a Core Duo at 1.86Ghz... seems about right. Thanks.

kiwi
08-15-2006, 03:52 PM
Any results Win vs Win64 vs linux? Anyway I will try and see :) Linux SMP is supposed to be better

EDIT:

OK Run on win, version 0.2
58.937s

e6300 @ 3.3Ghz


Well, I can tweak ram subtimings from Win but no way to do that from Linux :(
And DS3 bios is a bit buggy as for subtimings

[XC] leviathan18
08-15-2006, 04:41 PM
1m50.56s

t2400 stock with 1gb ram (laptop)

haPpydUde
08-15-2006, 09:29 PM
Its a nice little program you got there, xeon thats been on for days 2m on the dot, i was surprized it being so close to a c2d im guessing ht helps?

m0da
08-15-2006, 09:53 PM
02m 21.016s
specs in sig... still ha

RichBa5tard
08-16-2006, 01:02 AM
Its a nice little program you got there, xeon thats been on for days 2m on the dot, i was surprized it being so close to a c2d im guessing ht helps?

So you're running 2 xeons with HT on? 2m seems about right. Thank you.

RichBa5tard
08-16-2006, 01:03 AM
1m50.56s

t2400 stock with 1gb ram (laptop)

That's 10sec faster than my stock T2500, but I'm running Mac OSX with a lot of widgets running.... I'll see how it runs on windows.


02m 21.016s
specs in sig... still ha

That's too fast for a P4 at 3.9Ghz. You should be somewhere in the 2m45 - 3m range. Damn : /


The new beta 0.2 seems to produce better results:

6 minutes for a slow singlecore (Amd Turion 1.6Ghz)
4 minutes for a midrange singlecore (P4 3Ghz)
3 minutes for a fast singlecore (core solo 2ghz)
2 minutes for a midrange dual core (core duo 2ghz, X2 @ 2.5ghz, dual xeon 3Ghz)
1 minute for a fast dual core (core duo at 3.5ghz)
45s for a midrange quad core (woodcrest at 3ghz)

Anomalities:
P4 @ 3.9Ghz: 2m21s ? Thats too fast, it should be around 2m45 - 3m.
No P4 dual core results?

leejsmith
08-16-2006, 02:12 AM
would it be possible to run it for a set period of time and report on how many calculations performed in that period.
there is another thread someware with this idea and i am not sure who suggested it i thought it was a good idea.

RichBa5tard
08-16-2006, 02:22 AM
Yes that's possible, but I believe the popularity of SuperPi is mainly because it's so much fun trying to reach a certain time. You can even explain how superpi works to a computer illiterate.

alpha0ne
08-16-2006, 02:30 AM
1M 4.539s as per sig running @ 3GHz

Great program :clap:

hrvojezg00
08-16-2006, 03:06 AM
your download link doesn`t work, I`d like to try it on my dual dual opteron system
greats

RichBa5tard
08-16-2006, 03:09 AM
Sorry, new version, new link:

http://www.hwbot.org/download/multipi-0.2.zip

leejsmith
08-16-2006, 03:11 AM
Yes that's possible, but I believe the popularity of SuperPi is mainly because it's so much fun trying to reach a certain time. You can even explain how superpi works to a computer illiterate.

i belive it was based on the idea that soon 1m will be so quick and so 2m will be the new 1m. if it's based on lets say 60 secounds then no matter how fast the cpu the same benchmark can be run.

RichBa5tard
08-16-2006, 03:24 AM
i belive it was based on the idea that soon 1m will be so quick and so 2m will be the new 1m. if it's based on lets say 60 secounds then no matter how fast the cpu the same benchmark can be run.

I fully agree, but it wouldn't be as much fun. There are dozens of apps that bench your cpu for x time, but they're not populare, are they? :)

leejsmith
08-16-2006, 03:55 AM
I fully agree, but it wouldn't be as much fun. There are dozens of apps that bench your cpu for x time, but they're not populare, are they? :)

i know it was awsome when 10 secounds was broke in 1m pi.
perhaps as an option if you have nothing better to do.

m0da
08-16-2006, 09:06 AM
P4 550 has HT as well, maybe that's why the time is a bit lower:

http://img146.imageshack.us/img146/32/multipiiv9.th.jpg (http://img146.imageshack.us/my.php?image=multipiiv9.jpg)

Vapor
08-16-2006, 10:10 AM
Any chance of a non-java coming out, even if it means no cross-platform comparos? Maybe call it MultiPiW? It'd need to be slightly re-worked to accomodate for Java's slowness, but I imagine it'd really increase popularity.

dig412
08-16-2006, 10:22 AM
Any chance of a non-java coming out, even if it means no cross-platform comparos? Maybe call it MultiPiW? It'd need to be slightly re-worked to accomodate for Java's slowness, but I imagine it'd really increase popularity.
Speed doesnt matter as much, as long as it's consistent across platforms. Java makes it much easier to compare a linux and windows system for example.

leejsmith
08-16-2006, 10:52 AM
my time 56s
http://www.ljsnet.co.uk/core2/miltipi3400.jpg

RichBa5tard
08-16-2006, 11:21 AM
Any chance of a non-java coming out, even if it means no cross-platform comparos? Maybe call it MultiPiW? It'd need to be slightly re-worked to accomodate for Java's slowness, but I imagine it'd really increase popularity.

Not planning a non java version, no. Why would it increase popularity?

Vapor
08-16-2006, 11:35 AM
By eliminating the need to install Java runtime enviro on a bench partition....people already throw fits installing .NET framework for CCC.

I like the cross-platform comparos, but a cut-and-dry version in C would be great for the benching partition. (and we'd all know it couldn't be compared across platforms, but it'd still be great for just our purposes)


Speed doesnt matter as much, as long as it's consistent across platforms. Java makes it much easier to compare a linux and windows system for example.Of course, but this exact same algorithm in C would be too fast for benching, probably even faster than SPi 1M. Hence the necessary tweaks (if a benching version were made).

btw:
http://img129.imageshack.us/img129/9618/screenshot816200624219pmmg4.png

RichBa5tard
08-16-2006, 12:52 PM
Still not convinced me not to do it in java. Okay, installing .net sucks. If you really hate to install java that much I can make a bundled version. It wouldn't require a java vm installed, but it would weigh a heavy 30mb, but my stats (hwbot.org) tells me 97.5% of the visitors already have it installed.

Cross platform is a holy grail for me, i'm tired of not being able to bench my linux / mac boxes.

Kunaak
08-16-2006, 01:10 PM
I say let them throw fits.
every damn time I been apart of testing some program like this, 90% of overclockers want to strip the program down to nothing, so it fits thier needs alone.

I had no problem downloading java for this...
I dont see why anyone else would.
all I want is functionality.
this is doing great so far... thats all I want to see :)

one of the things that killed PI-Fast, was that there was about 20 different versions out there, and no version was clearly marked, so if you didn't like your time, just download a different version and try again....
it was frustrating as hell back then to find versions from japan were like 10 seconds faster, and looked exactly like the version I was using...
once people saw all that crap, that benchmark pretty much died.

I think your on the right track so far.

Gautam
08-16-2006, 02:23 PM
Frederik and Kunaak-

This isn't an issue about people being upset over extra stuff to install. It's a question of which language will work best to create a taxing benchmark. Java is inherently too inefficient to be optimal for the task at hand. This current version fails to even stress the CPU 100% at realtime. (Check task manager) Java is bloated and a little clunky with memory allocation. It has a lot of nice uses, but IMHO, it is less than optimal to be used in such an application. I too feel that C/C++ would be better choices, with seperate versions compiled for Windows, Linux and MacOS. The loss of portability is unfortunate, but I highly doubt that even a Java version would result in identical performance across all operating systems given the same hardware.

dig412
08-16-2006, 02:52 PM
The 100% cpu thing is odd, but it uses 98% on my (single core) pc. The problem with having the c versions for every platform is maintaining it. RichBa5tard has already said its too much hassle, and if it works well enough in java, which i think it does, then its no problem.

dr_sharp
08-16-2006, 05:52 PM
2:32.938 - AXP-M @ 2420MHz

These times are still too quick IMO. If this program is to test the next generation of cpus... its going to need to hold conroes to 2min+

Gautam
08-16-2006, 09:38 PM
I think the duration is all right...takes me around 40 secs. A bit slower would be nice though to take into account quad scores. Here's my score for what its worth.

http://mysite.verizon.net/gautamb/multipi.JPG

RichBa5tard
08-16-2006, 11:44 PM
Frederik and Kunaak-

This isn't an issue about people being upset over extra stuff to install. It's a question of which language will work best to create a taxing benchmark. Java is inherently too inefficient to be optimal for the task at hand. This current version fails to even stress the CPU 100% at realtime. (Check task manager) Java is bloated and a little clunky with memory allocation. It has a lot of nice uses, but IMHO, it is less than optimal to be used in such an application. I too feel that C/C++ would be better choices, with seperate versions compiled for Windows, Linux and MacOS. The loss of portability is unfortunate, but I highly doubt that even a Java version would result in identical performance across all operating systems given the same hardware.

Gautam,

The goal of superpi/multipi is not to be the most efficient Pi calculation app. The algoritm is filled with checks for bad memory / calculation errors to dedect unstable overclocks, that's why it's so 'slow'.

As for the not stressing the processors 100%, that's intentional too. The first version (0.1) stressed all the cores at 100%, but that gave such a huge performance gain on a quad core, it just wouldn't be fun. A mediocre 2ghz A64 needed 4 minutes, a quad core 30 seconds... those differences are just way too high to make it fun for everyone. v0.2 taxes the memory a lot more, which causes a 100% cpuload with one core, and ~80% with 2 cores. I think this is much more fun as now your memory speed matters too.

I am, in no way, claiming that I've used the most effecient algoritm to calculate pi, nor that I've used the most efficient language to do it.

My goals are:
- fun to benchmark
- advantage with multiple cores, but single core should still be 'fun' and quad core should not obliterate all at stock speed
- cross platform
- easy integration with hwbot.org

My goals are not:
- calculate pi in an efficient way

SuperPi doesn't uses the most efficient algorithm either... i haven't seen anyone complaining about that either.


One last thing Gautam, do you know how much time it takes to develop it in C++ for 3 different platforms properly? I'm with Kunaak on this, i've seen too many great benchmark apps go down the pooper because the community was demanding too much. I can write it in java in my spare time, I can not do it in C++ without taking a break from work.

hrvojezg00
08-17-2006, 03:01 AM
54 sec on dual dual 265 opterons

Gautam
08-17-2006, 07:47 AM
I didn't realize that that was deliberate...so you ended up "cheating" ultimately. ;) But you know that I'm in support for leveling the playing field, so I retract what I said earlier in that case. Good concept. :up:

I wasn't expecting this to be such a sore subject. You know I appreciate all of your hard work, but I always try to share my thoughts as well. The efficiency isn't a big deal like I said...just the stressing of both the CPU and memory subsystem. If you feel Java is the best path to take then no worries. :toast:

RichBa5tard
08-17-2006, 08:16 AM
Yeah, cross platform is a sore subject for me. I don't want to install windows just for benching! Give us mac/linux users a break! :D

Gautam
08-17-2006, 08:21 AM
Riddle me this though, do you think that *nix results will be comparable with Windows results?

Anyone with a Core-based Mac run this yet?

M.Beier
08-17-2006, 08:37 AM
RichBa5tard, you have PM.

EDIT:
Got a 53 sec time on my 3.6ghz conroe.

RichBa5tard
08-17-2006, 08:48 AM
Riddle me this though, do you think that *nix results will be comparable with Windows results?

Anyone with a Core-based Mac run this yet?

I honestly think so, yes.

I've got a core based mac, with dual boot windows. I'll compare it tonight.

RichBa5tard
08-18-2006, 06:53 AM
Well, I've tested 0.2 thoroughly and there seems to be a difference between platforms I think it has mainly to do with the amount of open processors primarely caused by the window manager. The heavier/fancier your GUI, the slower multipi gets.

Ranked by speed:
linux terminal > linux KDE > windows XP > MacOSX Tiger

The gain for a linux terminal was too much to call it fair, so I started tweaking the algorithm again.

The new algorithm (0.3, to be released early next week), will tax memory less, but will be comparable between platforms. I haven't had the chance to test it on Mac OSX yet, but results on a linux terminal and windows gui are nearly identical (I did 5 runs on each and fastest run on linux was not more than 200 milliseconds faster than slowest run on windows xp). If I get the same, identical results on macosx i think we have a winner. :)

It doesn't tax the memory as much anymore though, but I find consistency between platforms much more important than taxing the memory.

Gautam
08-18-2006, 07:28 AM
Yeah, I was expecting mostly the same. In case you haven't noticed, all experienced Pi benchers use Windows Server 2003. Despite using the same kernel and otherwise being virtually identical to XP, its much leaner and as a result will nearly almost win, albeit by a tiny amount. Linux would be one step further.

Once again this is quite a feat- a multithreaded benchmark comparable across virtually any platform. Quite unprecedented I believe. :)

RichBa5tard
08-18-2006, 07:35 AM
w00t :)

I'll see if I can get hold of a Win2003 somewhere. ;)

Gautam
08-18-2006, 07:52 AM
180 day eval free from Microsoft (http://www.microsoft.com/windowsserver2003/evaluation/trial/default.mspx). :D ;)

RichBa5tard
08-18-2006, 07:56 AM
Thanks!

kiwi
08-18-2006, 03:18 PM
Sorry what do you mean by "taxing" memory?

I can see there is a switch to VM how much ram to use but not sure what is optimal

dinos22
08-18-2006, 04:32 PM
Well, I've tested 0.2 thoroughly and there seems to be a difference between platforms I think it has mainly to do with the amount of open processors primarely caused by the window manager. The heavier/fancier your GUI, the slower multipi gets.

Ranked by speed:
linux terminal > linux KDE > windows XP > MacOSX Tiger

The gain for a linux terminal was too much to call it fair, so I started tweaking the algorithm again.

The new algorithm (0.3, to be released early next week), will tax memory less, but will be comparable between platforms. I haven't had the chance to test it on Mac OSX yet, but results on a linux terminal and windows gui are nearly identical (I did 5 runs on each and fastest run on linux was not more than 200 milliseconds faster than slowest run on windows xp). If I get the same, identical results on macosx i think we have a winner. :)

It doesn't tax the memory as much anymore though, but I find consistency between platforms much more important than taxing the memory.lol you just gave me a good idea

*searches for my copy of linux that installs in RAM heh :)

dig412
08-19-2006, 02:59 AM
lol you just gave me a good idea

*searches for my copy of linux that installs in RAM heh :)
What you need is a live cd, they dont use the hard drive at all :P
Try something like SLAX (http://www.slax.org), its pretty easy to use, and it can copy itself entirely to ram.

AAdjuster
08-19-2006, 05:28 AM
1:54.254 sec's. On my Newark at 3060 ghz.

Peen
08-20-2006, 02:08 PM
I havent posted here in like a year but decided to try this out

P4 Prescott w/HT 3.6ghz - 2m 31s

P4 Prescott w/o HT 3.6ghz - 3m 11s

Nice benchmark!

CCUABIDExORxDIE
08-23-2006, 10:00 PM
http://img.photobucket.com/albums/v642/TerrorXepidemic1/mutlipi.png

conroe really doesnt dominate here for the mhz. i wonder what an x2 at like 3.6ghz would do.

massman
08-24-2006, 12:56 PM
conroe really doesnt dominate here for the mhz. i wonder what an x2 at like 3.6ghz would do.

certainly not 40s. I guess about 50s

dsl
08-24-2006, 03:44 PM
DC Opty 175 @ 2.25GHz - 1m 39.266s
24/7 OS, a lot of apps running in background

bsodmike
09-11-2006, 01:21 PM
This doesn't work right in OS X 10.4.7... it's really screwed up. ran as sh file.sh



[01:53:54] kurosaki-ichigo:~/Desktop/multipi michaeldesilva$ sh multipi.sh
**** MULTIPI 0.2: a multithreaded pi calculator ****

USE: /MultiPi <numberofiterations> <workunitsize>
<numberofiterations>: total number of iterations to find pi - DEFAULT: 100000000
<workunitsize>: divise iterations into blocks of this size to spread work over cores - DEFAULT: 5000000

WARNING: this is a beta version, do not distribute unless you're a betatester. Results may vary in future versions.

Starting Pi calculation
Available cores dedected: 2
Calculation on core 1 started.
Calculation on core 2 started.
30m 08.295s - Loop 1 of 20 finished
30m 16.048s - Loop 3 of 20 finished
^Z
[1]+ Stopped sh multipi.sh


it goes towards 60secs, and ends up with a 32min end time..it's messed up ;)

RichBa5tard
09-11-2006, 01:59 PM
That's odd, as I have developped it on MacOSX 10.4.7 :D

dr_sharp
09-11-2006, 03:52 PM
Oh snap, forgot about this thread...

53.391 - E6600 @ 400x9 (3.6G)

Ace-a-Rue
09-12-2006, 08:02 PM
57.375 s

specs below

eshbach
09-12-2006, 08:23 PM
here's an idea...

i've been playing with the JNI a bit at work because we have some java apps and some C++ apps that need to run each other's code.

so, it occurred to me that you could use the JNI for something like this to improve performance while keeping cross-platform interoperability.

basically you'd do the algorithm in C or C++ (probably C), and then you could either leave the console interface and the rest of the functionality in java (possibly even the threading, depending on how you've written the algorithm), or you could write a Super-Pi style swing gui for it, or whatever.

then you'd just need to compile the C code on each platform (not that big a deal, since there wouldn't be any platform-specific code in there), and then have the java portion detect the OS and pick the right native assembly to use... or you could just have seperate versions for each OS, if you'd prefer.

so, if you think that could be beneficial, let me know. i'd be glad to help out if you want, but the JNI is pretty straightforward stuff, there would just be the matter of converting the algorithm, which could be difficult, or could be a breeze, i can't say without looking at the code.

wittekakker
09-13-2006, 04:41 AM
Celeron D @ 3,2GHz:

http://home.scarlet.be/wittekakker/multipis.jpg

Heineken01
10-12-2006, 09:44 AM
My 2 cents :toast:

http://home.concepts.nl/~pc119488/multipi.jpg

Cheers!

Lancer33
10-14-2006, 10:55 PM
Dead stock.

Zardokk
10-15-2006, 07:45 AM
Oh rly? How 'bout this (4400 at 2.5GHz):

http://www.kingsasquatch.com/images/multipi.jpg

m0da
10-15-2006, 10:14 PM
my conroe, finally!
http://img243.imageshack.us/img243/946/untitledgi7.jpg (http://imageshack.us)

\= 49s

RichBa5tard
10-16-2006, 01:42 AM
here's an idea...
...
so, it occurred to me that you could use the JNI for something like this to improve performance while keeping cross-platform interoperability.

basically you'd do the algorithm in C or C++ (probably C), and then you could either leave the console interface and the rest of the functionality in java ...

I haven't played with JNI before but I will certainly appreciate the offer. Do you know whether you can control the thread priority of the forked C++ thread?

RichBa5tard
10-16-2006, 02:00 AM
I'm troubled having to choose between a scientifical horrible method to calculate PI (MC), which does provide consistent benchmark results and is fun to bench (multiple cores give a nice boost, but still fun with single core), and using a solid scientific method to calculate PI effeciently, but which gives multi/quad cores a tremendous advantage.

What's the most important, a fun to bench benchapp that does a rather onorthodox computation, or a more scientific benchapp that may not be as fun, but actually calculates Pi to a gazillion numbers effeciently?

Thickbrit
10-16-2006, 03:19 AM
Keep it real man.Accuracy,i say.

RichBa5tard
10-16-2006, 03:41 AM
What defines 'real'? if i'd use a scientific way to accuratly calculate pi, quadcores would completly obliterate all scores. It would take like 20 sec on a kentsfield and 4 minutes on a A64 3000+. This doesn't reflect real world performance, nor would it be fun to bench (imho).

mtzki
10-16-2006, 03:45 AM
Random number simulations are contemporary science if anything.

On the other hand, some of the desired effects might be possible also for the sum rules by clever indexing.

Anyways, i would vote for staying with mc. Taxing the cpu in a consistent manner is all that matters here imo. The results seemed to be surprisingly good in this respect.

Heineken01
10-17-2006, 08:35 AM
http://home.concepts.nl/~pc119488/multipi1.jpg

Cheers!

odin318
11-01-2006, 10:22 AM
I have:

Java 1.5_09, do I need another version? Im running Win XP x64. MultiPi says I dont have Java :confused:

Vapor
11-01-2006, 05:18 PM
http://img176.imageshack.us/img176/1812/hwpi1me8.png

Core 1 crashes out instantly....and I have more than enough memory in the system....maybe a Java setting?

The Ghost
11-04-2006, 08:33 PM
here is a thread with a pair of opterons dual core systems

http://www.amdzone.com/index.php?name=PNphpBB2&file=viewtopic&p=117419&sid=ff9b5df39fe5bc702f1a750e6d8c2d0c#117419

wwwww
11-04-2006, 09:35 PM
http://www.wprime.net/multipi1.png

StuffRippx0r
11-24-2006, 01:01 PM
Here you go. Shows how insecure JAVA is.

- edit by rb -
Bashing java because it's easy to decompile does not belong in this topic.

rjw
12-12-2006, 09:02 PM
35.656 seconds on qx6700 @4.29Ghz

Second run was oveer 45 seconds

taskmanager in xp 32 bit states that java.exe is only using between 50 and 65% of cpu

Sometimes it starts core 2 before core 1 ???

strange results.....:banana:

pois0n
01-21-2007, 11:07 PM
Something is wrong with this :stick:

http://img212.imageshack.us/img212/8495/superpi6fs.jpg

Blang
01-29-2007, 08:18 PM
http://img176.imageshack.us/img176/1812/hwpi1me8.png

Core 1 crashes out instantly....and I have more than enough memory in the system....maybe a Java setting?

You need to increase the amount of memory Java is using. If you open the .sh file included with multipi there is "java -Xmx512m -jar multipi-0.2.jar" -Xmx512m increases the maximum amount of memory Java is using and it'll stop that other core from crashing out.

I had trouble getting it working in the end I ran it through Terminal using the following;

java -Xmx512m -jar /users/james/multipi/multipi-0.2.jar

On my Dual 2Ghz G5 I got 1m 58.481s

Postal Dude
07-08-2007, 06:21 AM
Hey guys. I kno this is none pc related but i thought it would be interesting in this thread. I just downloaded a java virtual machine for the PSP which should hopefully be able to run this program. Will post results on how a 333Mhz psp cpu fairs up against C2D's and such

RichBa5tard
07-09-2007, 08:15 AM
well, multipi is no longer under development (it has been replaced with wPrime, but this in windows only), but i'd still be interesting to see whether it runs on a PSP. :D

Postal Dude
07-13-2007, 11:55 AM
Will let you kno as soon as i get java to work. Its gettin annoying lol. Hopefully i'll get it all to work and do it in under 10 minutes lol

Java on psp only supports MIDP 2.0 stuff. Like mobile phone games and such. Dont think this will run on it at all.

Off topic. iv found a prog to calculate Pi on the PSP. maximum of 30000 DPs tho. Think its a memory limitation. Considerin the 256K calculation takes i dunno, less than a second on a decent pc. I think the psp is gonna be considerably slower!