Welcome to MilkyWay@home

Posts by Brickhead

61) Message boards : Number crunching : GPU units only start if I suspends projects (Message 33153)
Posted 8 Nov 2009 by Brickhead
Post:
Boinc 6.6.38

And therein lies your problem, methinks. Pre-6.10 versions of BOINC don't recognise CAL devices (ATI GPUs) and see MW as just another CPU project - subject to debt competition on equal terms with any other projects.

Upgrade to BOINC 6.10.17, restore the COPROC section(s) from CP's original app_info.xml, and you should be laughing.
62) Message boards : Application Code Discussion : GPU app teaser (Message 32777)
Posted 26 Oct 2009 by Brickhead
Post:
I've got a 512 MB 4850, running 1 unit at a time, catalyst 9.9, Windows 7 x64, app 20b, seems to be running fine, no lockups that I've seen.
Is this card capable of running 2 WU's at a time without incident?

Yes, mine do.
63) Message boards : Number crunching : No work?! (Message 32271)
Posted 11 Oct 2009 by Brickhead
Post:
Got a few 'ubnable to contact the database', but now it seems that things are working and the server is slowly catching up.
64) Message boards : Number crunching : BOINC 6.10.12 fixes Ati issues? (6.10.13 has been released as well) (Message 32185)
Posted 9 Oct 2009 by Brickhead
Post:
Question in general about the MW & Collatz apps. Since v2 Collatz has been more or less stable on my machines (both HD 4770 / WinXP32 / v9.8 no CCC / BOINC v6.10.13). MW runs fine when the machines are left as dedicated crunchers, but if I try to do anything else when MW is running (like open Windows explorer or running notepad, sometimes even bringing BOINC manager into focus) I'll get an immediate video crash. Given buggy ATI drivers, what could be the difference that allows Collatz to run with other apps and causes MW to crash? Possibly that MW is using double precision?

I have really no idea. Both applications access the GPU in exactly the same way. The only difference is what gets calculated. And I don't have those problems when using Catalyst 8.12. ATI apparently changed something in the drivers, what has broken them for MW on XP. Vista and Win7 drivers appear not to be affected and strangely it doesn't happen with Collatz either. I have a WinXP machine running Collatz with Catalyst 9.9 (and a 790GX chipset IGP needing just 11,000 seconds on average ;) without any problems.

I'm not entirely convinced that the CAL bugs introduced in ATI drivers earlier this year are fully exterminated. After I switched from BOINC 6.4.7 to 6.10.13, I tried both Cat 8.12 and 9.9 for a while on a HD4870, everything else unchanged.

Cat 8.12 rock solid up to 790MHz.
Cat 9.9 barely stable at stock, VPU crash guaranteed anywhere beyond 750Mhz.
65) Message boards : Number crunching : 3rd.in - optimized apps (Message 31492)
Posted 25 Sep 2009 by Brickhead
Post:
Guess you have Crossfire disabled. With the newer drivers it's generally okay to run it with activated crossfire so both GPUs can be used.

What, they finally got around that issue? And your 3870x2 (which I seem to remember has both connectors on one GPU) is using both fully?

If so, that's great news for those who haven't just sold their x2 cards (d'oh!)
66) Message boards : Number crunching : Cluster Physik and others ... Yea! (Message 30609)
Posted 12 Sep 2009 by Brickhead
Post:
Hear, hear!
67) Message boards : Number crunching : No work?! (Message 30077)
Posted 1 Sep 2009 by Brickhead
Post:
Got a Windows machine sitting somewhere? Then it *might* be possible have one of these alert you via email or sms:

http://www.ks-soft.net/hostmon.eng/
http://www.whatsupgold.com/
68) Message boards : Number crunching : Website slow for anyone else? (Message 29401)
Posted 15 Aug 2009 by Brickhead
Post:
Not just the website, it seems :(
69) Message boards : Number crunching : Donating to Milkyway@Home (Message 29371)
Posted 15 Aug 2009 by Brickhead
Post:
I believe the saying in the USofA is it doesn't matter the amount, but that you did donate. I'm sure it's all appreciated, right people?

I believe you can safely broaden the geographical basis for that sentiment.
70) Message boards : Number crunching : Donating to Milkyway@Home (Message 29361)
Posted 14 Aug 2009 by Brickhead
Post:
2. We need the billing address so we can send out thank you letters with tax deduction information.

I just got mine. Our counterpart to the IRS would of course just laugh at me, but the 'thank you' bit was nice.
71) Message boards : Number crunching : Terminology (Message 29337)
Posted 14 Aug 2009 by Brickhead
Post:
It's in cobblestones.
72) Message boards : Number crunching : Donating to Milkyway@Home (Message 29257)
Posted 12 Aug 2009 by Brickhead
Post:
Donation made.
73) Message boards : Number crunching : Small WUs is doing it again! (Message 29185)
Posted 10 Aug 2009 by Brickhead
Post:
Hello again,

The server should be running nicer now (let me know if it isn't), we went from having 4 runs [2.5deg 2s (2): 2.5deg 3s (2)] to 8 [2.5deg 3s (4): 5deg 2s (2): 5deg 3s (2)]. So there are WUs in total, more 3 stream runs, and more 5 degree stripes.

On a thread-related note-- yes, we do go out of the way to make you guys happy with the server speed. But I think this is part of our relationship-- we keep you happy by giving you work and credits, you keep us happy by providing the computing power we need. If keeping your boxes busy is the price to keep you crunching for us; well, that's a small price for 430 TFlops :)

Cheers,
John Vickers

Nicely done - 'ready to send' is again a 3-digit figure.
74) Message boards : Number crunching : deja vu! (Message 29086)
Posted 9 Aug 2009 by Brickhead
Post:
To me, this looks like someone had a discussion similar to this:

1. "Work seems to flow smoothly after we switched to longer WUs - which was not the case with the old shorter WUs - and all is well."

2. "Let's go back to shorter WUs."
75) Message boards : Number crunching : MilkyWay_GPU - Almost There! (Message 20331)
Posted 24 Apr 2009 by Brickhead
Post:
Well folks. Before you can run the script to make the project, you have to build the boinc system. Which, to say the least, is not pleasant on FreeBSD systems. Anywho, to update on the status, Travis is working on the GPU app. He has been working on it for the past 2 weeks. He has to translate the entire code from C into CUDA, which obviously isn't trivial. He told me the other day that he's got some functionality out of it but it's not yet complete.

And the ATI version? Last word I got was that our expert programmer was waiting for the (un-translated) source code.
76) Message boards : Number crunching : Catalyst 9.4 question (Message 18832)
Posted 15 Apr 2009 by Brickhead
Post:
Er... I think you can add 4850/1G to the list of cards with Cat 9.4 problems.

I noticed another case of frozen MW WUs, and decided to investigate. Disabled the "problem camouflage" called VPU recovery and restarted BOINC: OS freeze, no less! Hard reboot, same story within few minutes.

Replaced 9.4 with 8.12 and all is well (also on the 4870 cards still using 9.4).
77) Message boards : Number crunching : Catalyst 9.4 question (Message 18822)
Posted 15 Apr 2009 by Brickhead
Post:
That would mean only the 4890 cards are affected :(

I chose my wording carefully: "I haven't noticed". I'm not always watching, and a VPU recovery might have gone unnoticed. But nothing of the severity PB reports, no.
78) Message boards : Number crunching : Catalyst 9.4 question (Message 18818)
Posted 15 Apr 2009 by Brickhead
Post:
4870X2, 4870/1G, 4850/1G. All with Windows XP x64, Cat 9.4 and MW 0.19e.

I've noticed no difference from previous Cat versions at their best, apart from a few cases of all MW WUs freezing until BOINC restart or even reboot. But for all I know, that might not be driver-related.
79) Message boards : Number crunching : 3rd.in - optimized apps (Message 18517)
Posted 13 Apr 2009 by Brickhead
Post:
Please note a 'thank you' from me as well, Ice. Many of us could host the apps for public download, for free, as a favour to all, but you actually did it.
80) Message boards : Application Code Discussion : GPU/CUDA Milkyway (Message 18420)
Posted 12 Apr 2009 by Brickhead
Post:
Surprise, surprise.
Cuda capabale cards are listed here.

http://www.nvidia.com/object/cuda_learn_products.html

As long as double precision capability is required, mere CUDA capability is irrelevant. And resistance is futile. (Sorry, couldn't resist:)


Previous 20 · Next 20

©2024 Astroinformatics Group