MilkyWay

[Ben]

Does anyone know if the optimized app for v0.17 has been released? If so does anyone have the link, i'm getting a bit confused now about which app is which!  :roll:

[Temujin]

Unbelievable is all i can say... A MW member has developed a first test application on his own ATI GPU that does a MW app in 8seconds!!  :shock:

oh gawd, just when I've gone and got an expensive collection of nvidia cards, bugg3r

It does look to be an amazing app and should make all those ATI users very happy

hmmn, I do have 1 machine without a GPU

[Ben]

Unbelievable is all i can say... A MW member has developed a first test application on his own ATI GPU that does a MW app in 8seconds!!  :shock:

oh gawd, just when I've gone and got an expensive collection of nvidia cards, bugg3r

It does look to be an amazing app and should make all those ATI users very happy

hmmn, I do have 1 machine without a GPU

World domination coming your way John    :lol:

P.S - And as for the machine without a GPU? Not for long i think  :lol:

[Temujin]

Does anyone know if the optimized app for v0.17 has been released? If so does anyone have the link, i'm getting a bit confused now about which app is which!  :roll:

The only source I know for opti apps for windows is the planet3d site I posted earlier, which is also the same place for the ATI app

[Ben]

The only source I know for opti apps for windows is the planet3d site I posted earlier, which is also the same place for the ATI app

Sorry to be a bit thick  :oops: But i can't seem to find the latest one? I'm sure i read somewhere about an updated optimized application being released for 0.17.. But now im not so sure.

What version are you guys/gals using? Im currently on: v0.11 Optimised

[Temujin]

What version are you guys/gals using? Im currently on: v0.11 Optimised

try the optimised v0.16 posted earlier

[Ben]

What version are you guys/gals using? Im currently on: v0.11 Optimised

try the optimised v0.16 posted earlier

Thanks, missed that one  :oops: D'oh!

[Ben]

In regards to my post on the previous page about running MW on an ATI GPU, i was interested to see what the latest card from ATI was like so i did a little digging. And here is what i found:


ATI Radeon™ HD 4870 X2
1600 shader processors (compared to my 112 on my 8800GT!)
2 GB of GDDR5 memory
Combined peak power of 2.4 teraFLOPS of graphics horse power

So that got me thinking about Crossfire (which is the equivalent of Nvidias SLI). If you where to string together 3 of these GPU's then thats:
4800 shader processors
6 GB of GDDR5 memory
Combined peak power of 7.2 teraFLOPS of graphics horse power in your PC.

If a project was to utilize multiple GPU's (such as PS3 Grid) then you would have a mini super computer!

URL ---> http://game.amd.com/us-en/unlock_radeon ... 2.aspx?p=1

[Temujin]

having seen a little video, I think Ben now wants an ATI GPU

and who can blame him, stunning work

[Timby]

having seen a little video, I think Ben now wants an ATI GPU

and who can blame him, stunning work

Mmmm interesting and all that for as little as £160 !!!!  If you shop around

[Ben]

having seen a little video, I think Ben now wants an ATI GPU

and who can blame him, stunning work

Yeahhh, saw that in the MW forum!  :D Fantastic bit of kit! And yeah, who wants to trade an ATI for an Nvidia? Any takers?  :lol:

[Ben]

Ok you lucky ATI people, i have found the link to the GPU application for milkyway!

From what i understand it will work on:
HD38x0
HD4670
HD48x0

http://www.file-upload.net/download-141 ... 4.zip.html

I so hope they bring out a CUDA version also!

[UBT - mickyb69]

Another 2 or 3 days and we can go up to 7th place  :lol:

[Temujin]

Ok you lucky ATI people, i have found the link to the GPU application for milkyway!

Be aware though that the ATI GPU MW app is still limited to 300 WUs a day and that it claims/receives about 1/100th the credit of the same WU done on a CPU

It's still very much only a testbed for GPU crunching

[melter65]

The proverbial 'swings and roundabouts'!!  :lol:

[Ben]

Ok you lucky ATI people, i have found the link to the GPU application for milkyway!

Be aware though that the ATI GPU MW app is still limited to 300 WUs a day and that it claims/receives about 1/100th the credit of the same WU done on a CPU

It's still very much only a testbed for GPU crunching

Only 300!? Ah well...  :evil:

[Temujin]

Ok you lucky ATI people, i have found the link to the GPU application for milkyway!

Be aware though that the ATI GPU MW app is still limited to 300 WUs a day and that it claims/receives about 1/100th the credit of the same WU done on a CPU
It's still very much only a testbed for GPU crunching

Only 300!? Ah well...  :evil:

MW limits to 300 WUs per core per day (and even 108 credits/hour/core)
so a fancy new i7 could get 2400 WUs but that ATI app will still rattle them off way before 24 hours and only claim a fraction of what a CPU would claim.

Give them time to integrate it properly into boinc and that'll be the way to crunch MW :thumbup:

[Ben]

Seems daft that they have only decided to do CUDA. I guess it could be that they want to do one at a time and iron out the issues... But from what i understand they don't seem all that keen, even though the ATI cards (the newer ones) can outdo even the GTX280 on double precision floating point numbers (by more then 2x!) which are useful to scientific projects like MW which need a high level of accuracy.

Let's hope it gets incorporated!

[NaRyan]

I started on MW again yesterday, Using the v0.16 opt app.
Just wondered if any of you peeps are getting the odd workunit that has stupid high completion estimates?

They all seem to be the ones that have _l40_ in their name.
They all seem to estimate a runtime of nearly 2 days, yet they complete in about 5 to 6 minutes.

Also is anyone having their Task duration correction factor jump up to higher levels?
One of my AMD systems has jumped from 0.989412 to 2.024829 since starting MW, I have already had to "fix" it once due to it getting to such a high number that MW refused to give me any work.

Using Boinc v6.5.0 (on the 2 systems running gpugrid) as v6.4.5 (running this on the non-gpugrid ones) would not play nice with gpugrid.

[Ben]

Not sure about that  :? I haven't checked, but would it be giving you those times because you have been doing PG for a while?

Looking at my results they take between 6 and 10mins each, although saying that there is one in the pending list that has an est time of 18 mins... Which sticks out like a sore thumb (nm_s79_l40).

[hgblade]

My systems are running lots of different WU types but I have nothing which has a time to completion that looks weird. I have I40 and I41s that have different times - but isn't it the 'S' number which gives an indication of type and hence runtime. I seem to remember something posted on MW's website that indicated that S21 wus would run longer than others - and that is the case on my systems.

The S21-I40 and S21-I41s are set to around 10 mins to complete, but the S79-I41 tasks, S86-I40 and S86-I41 tasks are just under 8 minutes.

[NaRyan]

Just noticed that all the ones that have silly high runtimes are all re-sends from peeps who never returned them.

And after looking in (the excuse called) their forum, it seems a few workunits were a bit bugged after they changed the Flops based estimate.
So they gave stupid high runtimes of several days, when in fact with the opt app they take between 6 to 30 mins (depending on what poota got it).

So meh just another mistake from the peeps at MonkeyWay....

P.S The Ironic thing is Ben that thanks to the optimized app, I'm doing more workunits per day on Milkyway compared to PrimeGrid's PPS LLR
Done 1400 MW ones in the last 24 hours compared to the 1100 to 1150 I done per day on Primegrid. Least I don't have a few 1000 pending workunits on MW like I do have on Primegrid

[Temujin]

The MW project has adjusted the estimated completion time of the WUs because a lot of perople were getting DCFs of less than 0.02 which then causes boinc to not request any work (apparently).
They initially went too far, hence the huge estimates but have since reduced it a bit.

Let them run and it will sort itself out

[UBT - Mikee]

2009-02-05: MilkyWay@home - Project News: WUs purged quicker
Part of the problem with our server being slow is we've been crunching quite a few workunits and it's getting bogged down. I've recompiled a modified version of the db_purge which will let me purge in increments of less than a day, so
WUs will be purged every 6 hours (instead of 24). Hopefully this should decrease the load on the database and speed things up. Once the purge catches up tonight I expect we'll see a bit of an improvement in server performance.

Hmm, not quite as planned! MW has been totally off line since 1pm today!

[melter65]

Still down, all the peeps with fifteen+ machines all crunching MW will be peeved!  :lol:

[hgblade]

Still down, all the peeps with fifteen+ machines all crunching MW will be peeved!  :lol:

and those of us with only 4 machines  

[NaRyan]

The way I keep Jinxing these high credit projects, I'm surprised the Kreddit Kops ain't begging me to be a member.....

[Woodles]

Still down now, good job I'm not a credit whore!  :D

[NaRyan]

It must be working at times, then going wonky again.
As looking at Boincview logs, it shows that I have done some MW work today.
The last one being at 08:05.

[Woodles]

Mine was working fine until 13:50 yesterday when I got "Server error: can't attach shared memory" then no communication until 15:54 when things started working again ... until 16:02 when I got "Server error: can't attach shared memory" again. Nothing since then.

Still doing work from the (nearly empty) cache but can't send any results.

[NaRyan]

Same here.
I have 70 workunits stuck at uploading.
Before I went to bed at 04:00 I had 58 of them stuck at uploading.

So I somehow got 12 more of them, but the server must be well pooched.
Guess the new DB purge script has broke it....

[Woodles]

From Travis as reported on BoincStats at 14:16:49 UTC:

I'm afraid it was serious. The database got corrupted and took down the whole server.
We are working to sift through all the data but we're not sure when it'll be back up. My deepest apologies.

It's not looking good.

[NaRyan]

Guess MW needs to hire more Monkeys....

[Ben]

I'm afraid it was serious. The database got corrupted and took down the whole server.
We are working to sift through all the data but we're not sure when it'll be back up. My deepest apologies.

D'oh!

[NaRyan]

As seen on the MW forum....

indeed the feeder is crashed. working to get it back on now. we are starting with fresh workunit tables so the work that was done when the server was offline will likely not get credit. we apologize for this.

*mutter.....

EDIT: It's working again, and yup those 70 workunits I had got 0 credit

EDIT2:

Server Issues
February 7, 2009
I'm not generating any new work tonight, because the server is being really loaded down with all the old workunits being reported. I'm hoping by morning things should die down and I'll be able to start up work generation again. Thanks again for bearing with us through the downtime. We're trying to fix all the problems and make sure it doesn't happen again.

Pfffffft n00bs

[UBT - Mikee]

Work Started
February 7, 2009
Work should be flowing now, let me know if you have any problems.

Move along - nothing to see here...

http://milkyway.cs.rpi.edu/milkyway/server_status.php

[Timby]

Work Started
February 7, 2009
Work should be flowing now, let me know if you have any problems.

Move along - nothing to see here...

http://milkyway.cs.rpi.edu/milkyway/server_status.php

Your not wrong there Mikee

More Problems
February 7, 2009
It looks like we're having some kind of problem with the transitioner. It's running very slowly due to the fact that something else is hammering our mysql server. I'm hoping this is just because of people returning results which aren't in the database. We're looking into the problem to try and figure out what's wrong, so things are going to be a bit slow until we figure it out.

And still no work !!!

[Ben]

Looks like we are back in business  8)

Come and get 'em!

[Timby]

Ah thats better, just as a point of interest has anyone noticed the first two or three wu returned are particularly low with the credits ? Or is it just me !!!

[Ben]

I'll have a look into that on mine, but it wouldn't surprise me if they have altered the credit whilst they were down. Sneaky!

***Edit***
Seems to be fine here:
CPU          Claimed    Granted
410.42   2.10   12.31

[UBT - The Prof....]

Looks like we are back in business  8)

Come and get 'em!

Taken a while, but I finally got some wu's. It's given me 24 all @ 1 go
Plodding through them as best as I can

[Ben]

Do you have the optimized application prof?

[UBT - The Prof....]

Do you have the optimized application prof?

No I don't
Just been looking back at the previous post's about this, but not sure which one to download nor how to install it if I work out the correct one to download.  :oops: If it helps anyone who would like to help me out, startup info on boinc is as follows:-

08/02/2009 19:11:34||Starting BOINC client version 6.4.5 for windows_intelx86
08/02/2009 19:11:34||log flags: task, file_xfer, sched_ops
08/02/2009 19:11:34||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
08/02/2009 19:11:34||Data directory: C:\Documents and Settings\All Users\Application Data\BOINC
08/02/2009 19:11:34||Running under account Bob
08/02/2009 19:11:36||Processor: 2 GenuineIntel               Intel(R) Pentium(R) 4 CPU 3.20GHz [x86 Family 15 Model 2 Stepping 9]
08/02/2009 19:11:36||Processor features: fpu tsc sse sse2 mmx
08/02/2009 19:11:36||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, v.5657, (05.01.2600.00)
08/02/2009 19:11:36||Memory: 1023.48 MB physical, 2.41 GB virtual
08/02/2009 19:11:36||Disk: 48.83 GB total, 28.15 GB free
08/02/2009 19:11:36||Local time is UTC +0 hours
08/02/2009 19:11:36||Not using a proxy
08/02/2009 19:11:36||System clock was turned backwards; clearing timeouts
08/02/2009 19:11:37||No CUDA devices found
08/02/2009 19:11:37||No coprocessors


I would appreciate some help. Thanks

[Ben]

Sent you a PM regarding it

[UBT - JohnR]

Another SSE3 application here http://www.planet3dnow.de/vbulletin/sho ... 16&page=34.

Babel Fish gives "Since MW runs now (nearly) again, can you times report, how the new GPU App makes itself, which can count also the Multi Stream WUs now. To the memory, the App gives it here, needs Win64, SSE3 and a ATI Radeon HD38x0 or HD48x0."  My german is non existent.

[Temujin]

Another SSE3 application here http://www.planet3dnow.de/vbulletin/sho ... 16&page=34.

Babel Fish gives "Since MW runs now (nearly) again, can you times report, how the new GPU App makes itself, which can count also the Multi Stream WUs now. To the memory, the App gives it here, needs Win64, SSE3 and a ATI Radeon HD38x0 or HD48x0."  My german is non existent.

Caution required!!
Rather than an SSE3 CPU app, that looks to be their ATI GPU app

[NaRyan]

Geez just noticed that the duration correction factor on Mr temperamental poota is up to 41.575338 now

Boinc 6.4.5 go loco on win2k....
Heh 69 (behave Bob!!! )days to do a workunit... Did a CPDN workunit sneak into Milkyway or what

[UBT - Mikee]

2009-02-13: MilkyWay@home - Project News: Application Code v0.18b Released
I've upated the alpha release of the v0.18 application. It should now compile correctly. It has some changes for a small checkpointing bug and more performance improvements. I'm going to let other people try it out for a bit before I update the stock application. So if all goes well stock should upgrade to 0.18 in a couple days. After we've updated the stock application to v0.18 and we see that it's working correctly and we have optimized versions of it available, I'm going to stop awarding credit to pre 0.18 compiled applications, because we don't have any other way to get people to upgrade their optimized applications. I'll post a news item as a warning and provide links to appropriate v0.18 applications so that everyone has time to upgrade their applications before I make the switch.
add comment

[UBT - The Prof....]

Looks like server problems again

feeder milkyway Not Running
transitioner milkyway Not Running
db_purge milkyway Not Running
file_deleter milkyway Not Running
milkyway_av milkyway Not Running

[Ben]

Needs a good kicking again

***Edit***
NA shot it a few times with rubber pellets and it's back in action.

[UBT - Mikee]

Windows to v0.19
February 16, 2009
Dave didn't update the version number in the code, so the windows apps are now at v0.19... Please direct all flames his way
Updated Windows Stock Application

February 16, 2009
The stock windows application has been updated to v0.18 as well.
Updated Stock Application

February 15, 2009
I've updated the stock application for linux and os x to v0.18. The code is in the code release directory as mw_v0.18d. This has a couple new performance improvements (thanks to our users) which has it running 30% faster, or more. This should bring our stock application (and some of the other optimized versions of it) in line with the fastest optimized apps we're seeing out there. Because of this, in the next day or so we'll have another (much loved) credit modification. I'll most likely be dropping the credit rate by about 30-50% but at the same time removing the credit limit. Due to the nature of the performance improvements, older hosts should see better performance improvements than newer ones, so for most people the credit/hr should be the same, while faster hosts might lose some credit or gain some credit depending on where they are compared to the speed limit.

[NaRyan]

I'll most likely be dropping the credit rate by about 30-50%

As times I wonder if MW are trying to loose users.
High paying credit projects always attract more users.
To get the credit they have to do the science, more users = more science done, more science done = faster results for whatever science you are trying to prove...

Still think one of the project Admins I know had the best idea when D.A was harassing him about the "higher than seti credits".
He Told D.A that unless he butts out he would increase it so peeps got 100x the credit they were getting...
D.A kept his nose out (shame, as over 3,000 credits for 18 mins work would have been nice )

I swear if I ran a Boinc project, the credit would be that good, D.A's nose would bleed, oh and I would ban the Kredit Kops from the project
he he he he

[Ben]

*shout, cheer, applaud* na for the job!

[hgblade]

I swear if I ran a Boinc project, the credit would be that good, D.A's nose would bleed, oh and I would ban the Kredit Kops from the project
he he he he

Only trouble is  it'd be permanently jinxed and off-line  :lol:  :lol:  :lol:

[Ben]

I swear if I ran a Boinc project, the credit would be that good, D.A's nose would bleed, oh and I would ban the Kredit Kops from the project
he he he he

Only trouble is  it'd be permanently jinxed and off-line  :lol:  :lol:  :lol:

Good point there Keith  :shock: I take that back, no hobnobs for the n00b    :lol: Come get some shrapnel ya zombie!  :P hahaha

[NaRyan]

I swear if I ran a Boinc project, the credit would be that good, D.A's nose would bleed, oh and I would ban the Kredit Kops from the project
he he he he

Only trouble is  it'd be permanently jinxed and off-line  :lol:  :lol:  :lol:

Well it would be called Jinxed@home.
You don't do any work, but ya get credit for jinxing other projects.
And if ya Jinx seti, ya get a millionbillionquadzillion.12 credits

[Ben]

Well it would be called Jinxed@home.
You don't do any work, but ya get credit for jinxing other projects.
And if ya Jinx seti, ya get a millionbillionquadzillion.12 credits

This is starting to sound rather good  :lol:

[hgblade]

It's time to get all the older applications out there updated. As of February 17, midnight EST, we won't be awarding credit for applications flagged as v0.15 or earlier. These applications have a checkpointing problem and can frequently return invalid results to the server. So nows a good time to update to the newest optimized applications if you're using an older one. There are links to them in the number crunching and code discussion forums.

[NaRyan]

You can get the new 0.18d (linux)/0.19 (win) opt apps from here

Also they have adjusted the credits AGAIN!!!
My slower pootas now get less credit per workunit, yet my faster ones get more credit per workunit.....
The 108 credits per core is now gone... so yay.
However I now get less credits per hour
not down by that much (about 100 credit per hour), and I'm not using the 0.19 opt app yet, letting the cache run dry before I change...

So will see how things go after the app change.......

[Ben]

I'm now using the MW 0.19 app and it seems OK, although it does take slightly longer then the 0.16 app. Mine are touching about 9-13mins per WU.

[melter65]

A real n00b question here!

Do we have to delete the old app before installing the new app? Or will the new app install over the top of the old app?

I thank you!

[Temujin]

Or will the new app install over the top of the old app?

Yes
just copy the new files into the milkyway folder and boinc sorts what it needs when it reads app_info.xml

[melter65]

Thanks, John

[NaRyan]

The new app is not slower, it is some of the workunits that are longer.
I had a few 8 to 10 min ones using the v0.16 app
They all seem to be the ps_s20/ps_s21 range.

Looking at the Boincview logs at 0.16 vs 0.19 results in the same workunit number, the times are identical.

[UBT - The Prof....]

Running v19, no problems. Runs a bit faster on this machine, but getting less credit/wu than before. First one returned only gave 12.23, whereas before was getting up to 30
Ah, well. Good while it lasted

[melter65]

I'm not using the 0.19 opt app yet, letting the cache run dry before I change...

So will see how things go after the app change.......

I'll take my lead from you, NaRyan, and led my MW cache run down before installing the new app (whuich is d/loaded and ready to go)

[UBT - Halifax-lad]

I've just over written mine, only just put the other older version on yesterday

[Ben]

Same, otherwise come midnight (if you forget) you'll be doing WU's for free  :shock:

[melter65]

I think there will be a lot of aborted wu's come midnight!

[NaRyan]

From the MuppetWay home page.

What I'm thinking about dropping credits to around 4-6 credit per 79, 82 and 86 WU, and 6-9 credit per 20 and 21 WU

What the heck.....
If that happens they may as well shut MW down.

Those 20/21 workunits take between 30 mins to 45 mins on my slower pootas.... and get between 6-9 credits.

That happens then MW will join the rank of Seti for me.
As in not a *ahem* chance I will do any more of their workunits...

ALL ABOARD THE FAILBOAT!!!!

[Timby]

Same, otherwise come midnight (if you forget) you'll be doing WU's for free  :shock:

Sorry have I missed something? As I read it if you are using V0.16 you should be OK?

February 17, 2009
It's time to get all the older applications out there updated. As of February 17, midnight EST, we won't be awarding credit for applications flagged as v0.15 or earlier. These applications have a checkpointing problem and can frequently return invalid results to the server. So nows a good time to update to the newest optimized applications if you're using an older one. There are links to them in the number crunching and code discussion forums.

OK its time to get on my bike and ride, credits have now dropped so its goodnight from him etc and offsky back to ABC for me......

Mark

[NaRyan]

yup credits are now way down.
6.xx credits for the short ones, 9.xx for the long one...

The failboat has sailed and sunk....
bye bye me off to new project.

[Woodles]

Weird, at 08:42 I was still getting 12.28 for the quick ones and 18.52 for the slower ones, think I'll keep an eye on them for a while.

[Woodles]

Typical, as soon as I post, the next one gets 6.14  :(

Oh well, it was good while it lasted. (Still turns in a handy 1200+ credits per core per day though )

[Temujin]

The party's over
back to seti for me

[NaRyan]

The party's over
back to seti for me

Glutton for punishment ain't ya

[UBT - mickyb69]

Back to good old ABC with the nice 6 hr WU's .

I understand that theres a need for some kind of credit limits but why do the project admin think that we are loyal to 1 and whatever they do we will stay .
36 credits an hour on 2 cores ...nah

[Temujin]

The party's over
back to seti for me

Glutton for punishment ain't ya

Well Seti has always been my main project.
MW was only ever a distraction that paid rather well and now they've reduced its credits to the same (or lower) level as opti seti, I'll be heading back.

If/when they get MW running on cuda, then I could well move some of my cards from ps3grid to MW, especially if MW cuda runs cooler than ps3grid cuda as I think my home computer room will get a tad hot in the summer running ps3grid.

[UBT - mickyb69]

 :lol:  :lol: Travis is talking about increasing credits from 6 to 8..
Wow i wish he would make his mind up  :lol: When its back at 18 for a 10 min wu i might go back  

[Woodles]

 :lol:  :lol: Travis is talking about increasing credits from 6 to 8..
Wow i wish he would make his mind up  :lol: When its back at 18 for a 10 min wu i might go back  

8.14 for a 'short' WU from about 2pm, I've not reported any long ones yet.

[UBT - The Prof....]

Last batch I sent back a few mins ago averaged over 12.66 each, for a 15 min wu. Not too bad

:-EDIT-:

Last two wu's returned only gave 8.44 credits, so looks like I joined the low score group as well

[Woodles]

From Travis on the message boards:-

I'm going to move the credits up to 8 for 79, 82, 86 and 12 for 20, 21 and let things settle a bit.

There's two different types of WUs, one completes in about 2/3 of the time of the other and gets about 2/3 of the credit.

Currently giving 8.14 and 12.26 for me.

[hgblade]

Although my credits have dropped it doesn't seem too significant. I was getting around 25-26,000 a day, exclusively MW, until yesterday. Today I'm running at around 17,200 MW, but also I have 1 core finishing off some Cosmology and my laptop's 2 cores are off air for some repair work.

[Woodles]

Mine have actually gone up since yesterday, ~1200 per core per day yesterday, ~1600 per core per day today

... Of course, on Tuesday I was getting ~~2600 per core per day

And it still beats Seti at ~1000 per core per day

[NaRyan]

Looking at the MW credit it seems like it's sown 50% for me.
I was getting about 24,000 per day from it, now it looks like it's going to be 12,000 per day for it.

We need Riesel Sieve back with it's silly credits per day
Used to get 16,000 per day on that

[Temujin]

Thank goodness I have a mate at work who is an ultra early adopter :thumbup:

Earlier today, he parted company with an ATI HD4870 and I parted company with £120

In the last 10 minutes I've got the MilkyWay GPU app working on it and it's crunching 4 WUs at a time and completing all 4 in 50 seconds

Assuming I can keep the thing fed with WUs I reckon it'll give over 69,000 credits a day, FROM ONE GPU

If you can spare the cash for an HD3800 or HD4800 I highly recommend one :thumbup: :thumbup:
Sapphire HD4830 at overclockers, £86.24 inc VAT

[Ben]

Thank goodness I have a mate at work who is an ultra early adopter :thumbup:

Earlier today, he parted company with an ATI HD4870 and I parted company with £120

In the last 10 minutes I've got the MilkyWay GPU app working on it and it's crunching 4 WUs at a time and completing all 4 in 50 seconds

Assuming I can keep the thing fed with WUs I reckon it'll give over 69,000 credits a day, FROM ONE GPU

If you can spare the cash for an HD3800 or HD4800 I highly recommend one :thumbup: :thumbup:
Sapphire HD4830 at overclockers, £86.24 inc VAT

Holy mother of moo moos...  :shock: Nice on John, if i said i'm not jealous then i am a liar  :lol:

How many credit per WU is it? Just spent too much on a new Quad rig  :oops: But for £86.24... I'm half tempted to spend this months wages  :P

[Temujin]

How many credit per WU is it? Just spent too much on a new Quad rig  :oops: But for £86.24... I'm half tempted to spend this months wages  :P

this is the chap
So far its got 12.34 credits/WU mostly ps* WUs
This last hour it turned in 3479 credits
if you refresh the page every 15 seconds you can watch its credit grow

Comparing this HD4870 to the HD4830
HD4870=1200GFlops
HD4830=736 GFlops

So, if that 3479 credits/hour is typical and the cards performance scales with GFlops, the HD4870 should manage ~83K and the HD4830 ~51k
None too shabby for £86.24 inc VAT  :D


edit
A word of warning
There is no way you could use this machine for anything other than dedicated MW GPU crunching, screen response is appalling
luckily for me, its a spare machine sat in an aircon'd computer room

[Ben]

How many credit per WU is it? Just spent too much on a new Quad rig  :oops: But for £86.24... I'm half tempted to spend this months wages  :P

this is the chap
So far its got 12.34 credits/WU mostly ps* WUs
This last hour it turned in 3479 credits
if you refresh the page every 15 seconds you can watch its credit grow

Comparing this HD4870 to the HD4830
HD4870=1200GFlops
HD4830=736 GFlops

So, if that 3479 credits/hour is typical and the cards performance scales with GFlops, the HD4870 should manage ~83K and the HD4830 ~51k
None too shabby for £86.24 inc VAT  :D


edit
A word of warning
There is no way you could use this machine for anything other than dedicated MW GPU crunching, screen response is appalling
luckily for me, its a spare machine sat in an aircon'd computer room

Mini super computer there John! Your P4 is outdoing the very best quads and no doubt i7's on a per hour basis!  :twisted:

[NaRyan]

Holy mother of moo moos indeed! (err Ben where you in that L4D game when I made that comment about the dude t-bagging the dead cows?)

Mind you by the time I can afford £86.24, TRAVISty & the Kreddit Kops (sounds like a punk rock band ) will have sorted out that nice credit....

[NaRyan]

Jeez just realised it takes my 8800GT 19 hours to get 3,848 credits on gpugrid.

On MW with a HD4870 in 19 hours, you would get 63,612 (going by Temujin's 24 hour average)
So with a HD4830 you would somewhere near the 40,000 credit mark....

 :shock:  :shock:

So shall we all go moan at GPUGRID for giving us poo credits

*Rattles begging tin... spare some change to help a poor credit starving chap??????

[Timby]

Well guys, if your interested I found this comment posted by Travis on 20th Feb

There isn't a CUDA application for milkyway yet. We just got an undergrad to start working on it, so hopefully there'll be a version out in the next couple weeks.

 :lol:  :lol:  Lets see how things develop shal we, oh and nice going Temujin with the ATi card, like you need the extra credits I suppose  

Mark

[NaRyan]

Well guys, if your interested I found this comment posted by Travis on 20th Feb

There isn't a CUDA application for milkyway yet. We just got an undergrad to start working on it, so hopefully there'll be a version out in the next couple weeks.

 :lol:  :lol:  Lets see how things develop shal we, oh and nice going Temujin with the ATi card, like you need the extra credits I suppose  

Mark

Or in MW terms, "There isn't a CUDA application (that works, as we can't figure out how to program it. 2+2=1 right??) .
We just got an underpaid Monkey to start working on it, so hopefully there'll be a version out in the next couple weeks, So we can then say credits are too high, and rip you off again...."


Think Temujin is trying to get to 1 Billion credits.
As he heard that once you get to 1 Billion Credits you can trade them in for a Pork Pie....

[Temujin]

So shall we all go moan at GPUGRID for giving us poo credits

I've asked the question over at GPUGrid

1 response so far, looks like ATIs are just better at MW than NVs are at GPUGrid

[Ben]

Think Temujin is trying to get to 1 Billion credits.
As he heard that once you get to 1 Billion Credits you can trade them in for a Pork Pie....

Or a hobnob? And woe betide anyone who can trade that in for a pizza  :lol:

I may have just got enough to buy an atom!

[melter65]

luckily for me, its a spare machine sat in an aircon'd computer room

That's Ben's room out of the reckoning then!

[Ben]

luckily for me, its a spare machine sat in an aircon'd computer room

That's Ben's room out of the reckoning then!

Look what happened when i installed the ATI 4870x2  :shock:

[melter65]

Melted cheese, Mmmmmmmmmmm!!  :lol:

[Ben]

Melted cheese, Mmmmmmmmmmm!!  :lol:

So that's what plastic cheese is??  :lol:

[MAVRIK]

Would this be any good

http://www.cclonline.com/product-info.a ... ehd48701gb