Project 5801 issues. [Should be Offline]

[Xilikon]

Yes, I agree that the PG should get the 8600GT and the 9800GTX+ card in the labs and work to get the core stable with those cards. I bet that when they succeed at this, everything else should work fine.

It is not about the cards mind you.

What the exact changes were that went into 1.15 is not known to us. However, I think Prof. Pande pointed out that it is all about getting larger proteins to fold.

If it turns out that some older cards cannot reliably fold larger proteins, and over several hours, then the consequence has to be to stop folding on these cards.

I think this needs to be discussed because should it happen will many people be disappointed and the Pande Group cannot be expected to keep folding only small proteins.

I understand what you are telling and honestly, I don't think this is what they want. Doing that will shut them off from the big bunch of the GPU fleet and anger people. I don't expect the PG to push small proteins indefinitely and that's why they are pushing hard to have a stable core for large proteins.

[toTOW]

I know some guys with GTX2xx boards ... I'll ask them how they are doing with the latest cores.

[shatteredsilicon]

If it turns out that some older cards cannot reliably fold larger proteins, and over several hours, then the consequence has to be to stop folding on these cards.

I think this needs to be discussed because should it happen will many people be disappointed and the Pande Group cannot be expected to keep folding only small proteins.

It's not about capabilities of the cards. They are all capable of it. This is a clearly a big, be it in the core code or in the compiler.

[Xilikon]

I know some guys with GTX2xx boards ... I'll ask them how they are doing with the latest cores.

I have a GTX 260 myself and it works flawlessly beside 1-2 eue each few days due to a bad WU. Right now, it's out for Step-Up to a GTX 260 core 216 that I should receive next week for further testing.

[Drugless]

I know some guys with GTX2xx boards ... I'll ask them how they are doing with the latest cores.

I have 4 x GTX280's in one machine and was not experiencing too many eue's (1-2 per day across ALL my cards[9800gx2/gtx280]) I noticed the numerous posts re folders not able to fold at all and just thought I was being lucky!
Bear in mind I am referring to the period before the 5801 chaos started.Right now my stats indicate I getting back to normal so may still be able to provide more helpfull feedback later.

I will also see if I have a previous log for the GTX280's during the 5801 chaos and check if ANY managed to complete any part of a 5801 WU. I say this since if I understand Vijay's comment the GTX280's were used to regression test that WU (renamed / re-compiled and tested on a previous core?) without problems. If I'm wrong here please correct me. I suppose Stanford could check the stats their side and see if anyone returned a completed 5801 and check sender of that WU. That sender could then check which User ID and Machine ID to identify which type of card in USER world managed to complete anything.(unless this info is sent back on completed WU already) Also Stanford could have a record of WU's processed by UserID+MachineID and could establish if any particular GPU model has NOT returned a EUE for past 2/3 weeks.

Not sure if I am losing myself in my own discussion! (It's been a long 28 hr day!) Mostly just thinking aloud! I'm a Business Analyst by day and spend 90% of my time "Regression Testing" software so I am aware of the weird stuff that can appear in live but NOT QA! VERY FRUSTRATING and IMPOSSIBLE to explain to my boss!
Shut me up if applicable!

I am outa town at the moment so will post more details when I am back in 'folding land' if this helps as I don't have access to my machines remotely yet.

Sorry I did not post my 'non-issues' before but have been focussed on points and keeping system running rather than science lately!
YES I ADMIT IT! Points are important to me and if Stanford makes progress on their side because of my contribution then I can feel 'less' guilty.

[toTOW]

Well I have some reports of successes with GTX2xx and 1.15 core from team mates too ... that's a weird issue

[harlam357]

It's not about capabilities of the cards. They are all capable of it. This is a clearly a big, be it in the core code or in the compiler.

Agreed... if the cards are CUDA enabled, they should all be capable of running CUDA apps... no questions. If that is not the case with FAH, then something is wrong with the implementation talking to CUDA.

[jevans64]

Well I have some reports of successes with GTX2xx and 1.15 core from team mates too ... that's a weird issue

My GTX 280 w nV 178.24 and v1.15 didn't pass any of the p5801.

[toTOW]

We're not talking about p5801 anymore ... we know that nobody was able to fold them

That's why I'll close this thread, we've been too far off topic. To continue the talk about v1.18 core and it's successes on GTX2xx, let's move to : viewtopic.php?f=52&t=6552