Search results

AB, you need a PSU with 2 EPS12V connectors. You're not using a proprietary factory PSU so you don't need to worry about the extra power board.

Amen! But what about power consumption? It may be tricky balancing the overall cost when you consider what power hogs some of those GPUs are.

I can't quite parse that sentence structure, but I hope you're not wasting any researcher's time on a question like that. http://foldingforum.org/viewtopic.php?f=16&t=19042&p=190618&hilit=+GPU+QRB#p190618 July 2, 2011: We have been considering QRB for GPUs, which should finish the...

Install i7z from google code to get a somewhat realtime display of the actual frequencies.

I'm just not worthy to be seen with someone with such leet spelling skills. ;)

Finally my evil plan has been exposed!

1) see http://foldingforum.org/viewtopic.php?f=16&t=8

I ordered the "socket 2011" clamp for the older 212s but it did not fit the spacing on my SM board. Tech support claimed the spacing is different for "consumer" vs "server". I sent them back and just modified the stock brackets with a Dremel, which I should have done in the first place. All...

I have found several times now that when I see what looks like a "slow" 8101, if I reboot, the time per frame drops 1-2 minutes. Restarting without a reboot did not help. This is on a dual E5 box running Ubuntu 12.04 LTS with thekraken. I have not seen the same type of behavior on any of the...

You can download non-installer "packages" for 10.10 if you drill down from https://fah-web.stanford.edu/file-releases/beta/release/ For example, the client is at https://fah-web.stanford.edu/file-releases/beta/release/fahclient/ubuntu-10.10-64bit/v7.1/latest.deb (or latest.tar.bz2) and the...

The ifconfig output looks fine. You would see errors in the eth2 stats (right after RX and TX packets) if there were hardware issues in your NIC or cabling.

Have you looked at the output of the "ifconfig" command to see if there are errors being reported there?

I didn't mention any teams in that thread. Skillz is right, don't worry about it if you're not cheating.

I'd guess rather than any nefarious reasons that the assignment logic is either flawed or too limited to handle every case correctly. There have been similar issues in the past - for example at one point if you were using the beta flag and there was no beta work then you got no work at all. In...

Actually if you look at one particular OCN user you will see more than half of the source of that drop (hint: HPCS) http://folding.extremeoverclocking.com/user_summary.php?s=&u=554874

If you mean the SLES11 issue, yes, it's in the thread that's pointed to in the Support section of your README. ;) http://www.amdzone.com/phpbb3/viewtopic.php?f=521&t=138463&p=218502#p218502 Looks like the OS is misbehaving, not the Kraken.

Apparently theKraken isn't working correctly on SLES11, if you look at the thread at amdzone, though that's not the issue here. There was a request on ff.org in February 2010 to be able to force DLB on. Maybe in a few more years it will be added to the client...

There is no way to automatically force DLB via Stanford's software. However, the latest version(s?) of theKraken takes advantage of the behavior you observed by automatically restarting a new WU shortly after a checkpoint (IF you enable the feature). You should download the latest version...

What are the memory speeds on the two boxes? Given your data plus the dual Sandybridge numbers, it may be that 8101 is more memory bandwidth intensive than the 69xxs.

On February 10, I posted in the beta forum that even a true 24-core dual 6164HE (1.7GHz) wouldn't make the deadlines with up/download times. Dr. Kasson did mention that they would consider the deadlines independently of the points values, but they don't seem to have changed from that original...

What is this "core hack or other work around" that some or all of your team is using?

I'm sure tear could whip up a little shell script that has the same effect as theKraken but is run manually once per WU. Maybe that wouldn't be as scary.

Looks like Kendrak has been out testing them. Pulled this off his facebook page:

You must be looking in the wrong place. There has been mention of this. It is natural that as the points structure has shifted to give more emphasis to larger SMP systems that EVGA doesn't even sell that we will fall behind - after all, much of our points still come from GPUs and boards that...

Did you notice in the board's specs: "Intel® Xeon® processor E5-2600/E5-1600 processors (up to 8 cores and up to 130W)". The 2687W is a 150W part, so maybe it is being ignored in order to prevent damage to the VRs or other parts of the board. It's discouraging though that there is no response...

Matches in what way? That sounds similar, but not the same, as the warning you get if you change the number of threads mid-WU - something about the simulation results not being binary identical - but that message also says the results will still be valid. If they didn't want that to happen...

I had the chance to try F@H on ScaleMP a while ago. I'll post some data when I find it so you don't have to waste any money trying. I was running multiple nodes and was not impressed. I do remember that the resultant theoretical PPD was lower than the sum of the PPD from the individual...

I have an X9DR<something> and it works fine both with 1333 registered ECC and 1333 unbuffered non-ECC. The support matrix for the board doesn't include the unbuffered stuff.

It's a Dell product (R720).

But then Interlagos has AVX. If Gromacs ever adds AVX support that could pull the balance back to Interlagos' favor.

Okay, I was assuming that voltage increases were included along with frequency increases in order to maintain screen colors other than blue. If you can keep a system stable with any significant frequency increase and no voltage increase, great! I guess it depends how much margin AMD builds in...

As do watts. Has anyone been measuring power as they increase the clock to see if PPD/watt is increasing, decreasing or holding steady?

As usual tear gets right to the heart of the matter => A2 processes with their own memory spaces vs A3/A5 threads with shared memory. Back in the good old A2 days when we were playing with recompiled mpiexec's, there actually was very good scaling on a transparent clustering solution I was...

I see some good points in the points references above from Amaruk - but there is one inconsistency. If 6900/1 are to be phased out, why bother increasing their points? It may be a red herring, but it certainly gives the appearance that points will be increasing. And if 6900/1 were increased...

Yes, that's exactly what I said (not, for the sarcasm impaired). There has never been anything but hand-waving justification of the QRB. If their project is all about science, they should be able to prove that the QRB represents the scientific value correctly.

They are increasing the gap between the "haves" and the "have-nots", without any true justification. I think in their rush to advance the science of distributed computing (which is their real contribution, not the medical research - see a post in this forum), they are forgetting their roots in...

I did not fabricate any data in the beta forum thread - it was all measured. I think what you meant was "if went off of 7im's and Punchy's opinions", but I never expressed any opinion about 8101 points, only that 6903/4 are overvalued. If they want to discourage GPU, uniprocessor and SMP...

As Nathan said, it was mighty close on my X5670's - 2.11 days not including up/download times. Any power or network glitches and it might not have made it.

There is a thread at EVGA with several different fixes listed. The basic problem seems to be that the BOINC client is looking for 64-bit "cores" which don't exist. Since you already have the client installed, I'd suggest the fix in the first post. http://www.evga.com/forums/tm.aspx?m=1330298

I see that's the understanding on your team. And you didn't look hard enough if you don't think there were complaints about beta leaks before the beta forum became read only. Regardless, you missed the point. As I said on ff.org, the eventual points when the project becomes public are an...