Welcome to MilkyWay@home

Posts by David Glogau*

21) Message boards : Number crunching : 2x 5970 CrossfireX Stuck wus (Message 39788)
Posted 17 May 2010 by ProfileDavid Glogau*
Post:
I run twin 5970's in crossfire mode, and it is impossible to run Collatz in this mode, except by running one WU at a time.

Take the bridge away, turn crossfire mode off.

Make sure you have a monitor, or dummy plug for each card.
(Mostly I use the VGA adapter into the second monitor port, the monitor does not need to be active for this to work).

Go into windows and set the desktop across all cards

Restart BOINC.
22) Message boards : Number crunching : 2x 5970 CrossfireX Stuck wus (Message 39643)
Posted 12 May 2010 by ProfileDavid Glogau*
Post:
I wish. I have problems every time. I normally, strip the current drivers out, then install the new ones. Then it doesn't work, so I rollback and redo several times tweeking as I go.

Except for the fact I can't run Collatz in this configuration (crossfire) I am pretty happy.

Cheers

Have you ever updated or reduced your driver without any problems? I'm thinking the problem might be an overheat on NB or a corrupted DLL.

23) Message boards : Number crunching : 2x 5970 CrossfireX Stuck wus (Message 39635)
Posted 12 May 2010 by ProfileDavid Glogau*
Post:
I run three boxes in 2 x 5970's crossfired mode.

BOINC 6.10.43
Windows 7 64 Bit.
17 870, 1156 socket.
2 with 10GB Ram. 1 with 8GB Ram.
CC 10.2 installed.
24) Message boards : Number crunching : Aaargh! Server out of new work! (Message 39191)
Posted 27 Apr 2010 by ProfileDavid Glogau*
Post:
Typical, just as I pass 200 million everything falls over!
25) Message boards : Number crunching : ATI graphics card maximum temps ? (Message 39190)
Posted 27 Apr 2010 by ProfileDavid Glogau*
Post:
Hi Ian

I try to keep mine about 85 to 90 degrees for the aircooled boxes.

Anything over 100 is bad for lifespan. Although I personally have not tested that theory for long periods.

Cheers
David
26) Message boards : Number crunching : Hardware Poll (Message 38571)
Posted 11 Apr 2010 by ProfileDavid Glogau*
Post:
Well, I spent quite a bit of time reading forum posts.

The problem is you cannot have two controllers, eg, Catalist and Nvidia. So you just load the specific card drivers, ONLY. So I run Catalist and the 260 drivers.

Reboot.

BOINC will show the various cards.

Seti is easy, it just looks for a NVidia card, finds it and starts crunching WU's.

Milkyway needs a project reset, at which point it will download two WU's ONE Ati application and ONE Nvidia application.

Select the Nvidia WU and abort it. Thereafter the Project only downloads ATI WU's.

At least that is how it works for me.
So I have twin crossfired 5970's crunching ATI 24/7, and the GTX 260 crunching Seti 24/7, well except for the times, like now, when the WU's are non existant haha.

One has a GTX260 for Seti as well.

Edit: P7P55D PRO

How did you manage to get ATI and NV running together? On my Vista64 - System the NV-Card is always disabled - Windows says, incompatible grphics card.
It would be grat if you could give me a hint!

Alexander

27) Message boards : Number crunching : Hardware Poll (Message 38454)
Posted 9 Apr 2010 by ProfileDavid Glogau*
Post:
Can't help, my three are all 1156 sockets. Each has dual 5970's. One has a GTX260 for Seti as well.

Edit: P7P55D PRO
28) Message boards : Number crunching : Kick the feeder pls ? (Message 38348)
Posted 8 Apr 2010 by ProfileDavid Glogau*
Post:
Bloody hell, Travis, you made me choke on my coffee with that statement, I am still laughing.

I KNOW that feeling well hahahahaha.



Basically the server likes to wait for me to be asleep or at a meeting and then crash. When I'm watching it, it plays nice.
29) Message boards : Number crunching : Kick the feeder pls ? (Message 38322)
Posted 7 Apr 2010 by ProfileDavid Glogau*
Post:
From the front page about the NEW Milkyway V3

This new application should really improve server performance as it doesn't use search parameter files to send out new workunits, it will take the parameters from the command line instead. Additionally, it will report the fitness in stderr, which gets moved into the result's xml.
What this means is the server will be creating 1 less file for every workunit made (it currently creates 1 file), and receiving 1 less file for every result sent (it currently receives 1 file). This means sending workunits out and getting results in won't be hammering the filesystem nearly as hard. Right now the recent server crashes have been happening because the file deletion daemon just can't handle the sheer mass of WU files that are being generated, and it brings the server to a screeching halt which then crashes the other daemons.


Skye Hunter, you will find this NEW version will stop all these problems.


For the third time today.

30) Message boards : News : stock ATI 58x0 apps updated (Message 38305)
Posted 7 Apr 2010 by ProfileDavid Glogau*
Post:
A few early hicups, reset all four boxes, and all is well so far.

Thanks for all your work CP, et. al.

Kind Regards
31) Message boards : News : validator strictness (Message 38231)
Posted 6 Apr 2010 by ProfileDavid Glogau*
Post:
So should I take my 5970's down for a rest, or has this problem been fixed yet?


I was browsing the ATI/AMD Developer KB and ran across this, might it have something to do with the apparent problems with 58xx series cards?

Nice catch!
This could be really the culprit. I hope I can build a modified version later today (should be easy, without the need to search for other issues).

32) Message boards : Number crunching : Question (Message 38225)
Posted 6 Apr 2010 by ProfileDavid Glogau*
Post:
Hi Nitro, the CPU's are good, the GPU is no good to this project. Try Collatz for the GPU, because they only use Single Point (SP) precision which is what your GPU is.

Cheers

Check out the GPU sticky for more info.
33) Message boards : Number crunching : Disecting the new validator. (Message 38084)
Posted 5 Apr 2010 by ProfileDavid Glogau*
Post:
34) Message boards : Number crunching : New WUs - Complete, can't validate? (Message 38083)
Posted 5 Apr 2010 by ProfileDavid Glogau*
Post:
Just checked my results. I have over 1000+ which have not validated.

Seems like about a 20% failure rate over the completed ones.
Also got heaps as inconclusive, so things might change as we go.

Travis is on to it, so things will work out in the end.

Regards
35) Message boards : Number crunching : Any 2/3 length ATI cards available for double precision? (Message 37969)
Posted 2 Apr 2010 by ProfileDavid Glogau*
Post:
Forget the 5970, 5870 NO way.

The 4770 is a good price/size fit for that board IMHO.
36) Message boards : Number crunching : It is fermi? How do you think? (Message 37968)
Posted 2 Apr 2010 by ProfileDavid Glogau*
Post:

David---I'm curious if you read what I posted directly before this post? It is being stopped (hopefully) at the server level. I don't make the changes...Travis does, and I am in contact with him. We are working on the issue.


Hi Blurf

Certainly I read your post, I even quoted it. I appoligise for the haste, of my last post, and problems you may have had understanding it.

My point is: While Travis is, correctly, focused on the science, is is my experience that most crunchers are more focused on credits.

Given that focus, anything that does not involve a reduction in credits is a bit of a waste of time if the end result is to discourage crunchers from trying to fiddle the system.

I like my 10% penalty idea, because it is easily understood, and easy to implement. It also think it would be a effective deterrence.

Regards
37) Message boards : Number crunching : It is fermi? How do you think? (Message 37946)
Posted 2 Apr 2010 by ProfileDavid Glogau*
Post:
Just received an email from Travis:

I'm actually in the middle of updating the validator; which should really stop the problem.



I can assure you Burf, If this user is allowed to get away with this, the message being sent is "Try and rort the system, IF you get caught, nothing much will happen".

There needs be a consistent point deduction for this type of behaviour, otherwise you will continue to play "wack-a-mole".

I suggest a 10% deduction of total points, plus, email. A further 10%, of the remaining balance, should be deducted each week until the behavior stops.


If you want another NZD1,300.00 donation this year, please give this request serious consideration.

Kind Regards
David Glogau
38) Message boards : Number crunching : Longer WU's for same Amount of Credit ??? (Message 37924)
Posted 2 Apr 2010 by ProfileDavid Glogau*
Post:
Bye, have fun, see ya later.

NO wait, I am already have a presence there hahaha.

Don't worry team, I am here for the science, so I will stay!
Must admit, haven't seen the old nothing then 140% then count backwards routine for a few weeks now. ~2:37 on the 4 x aircooled 5970's and ~2:20 on the 2 x water cooled ones.
39) Message boards : Number crunching : It is fermi? How do you think? (Message 37914)
Posted 2 Apr 2010 by ProfileDavid Glogau*
Post:
Good, As a major Donor to the project, this deliberate sabotage of the project makes me VERY upset. The credits for this user need to be zeroed forthwith. Any further tampering may result in his team being punished as well(Need to be careful on this one).

But as someone with a weekly $100 power bill for this project, compiling unauthorised versions of the code, to increase ones credit, needs to treated very seriously.

Actually, I am a BIT angry as well, A final warning leading to banishment could well be appropriate as well.

Kind Regards
40) Message boards : Number crunching : 5970 - only one GPU available in BOINC (Message 37584)
Posted 20 Mar 2010 by ProfileDavid Glogau*
Post:
I think my problem is that this mobo is Nvidia 790i :-(


Mine are ASUS P7P55D.

Do you use the xfire bridge, or software-only?


They are physically connected.


Previous 20 · Next 20

©2020 Astroinformatics Group