Folding@home [V8]....the origami continues.

[krips]

How many PPD does a Q6600 get guys?

[powerarmour]

:lol:
How many PPD does a Q6600 get guys?

Stock or Overclocked..?

@ 3Ghz I get ~4300PPD running two SMP Linux clients.

http://foldingforum.org/viewtopic.php?f ... 24&start=0

[j3pflynn]

Well, I'm down a PC for at least a week; my son's PC bit the dust, burned up the 5V lines on the PSU. Funny, that's the 3rd time in as many years that an Antec PSU/Asus MB combination has done that to me. Got a replacement PSU, don't know yet if the MB is recoverable. I'll probably have to drive up to his school to pick it up this weekend, fix it and ship it back.

[Linky]

:lol:
How many PPD does a Q6600 get guys?

Stock or Overclocked..?

@ 3Ghz I get ~4300PPD running two SMP Linux clients.

http://foldingforum.org/viewtopic.php?f ... 24&start=0

Think this will be my project tonight. Please stand by for questions that are sure to follow.

[Linky]

Well, that was it, folks; UAC wasn't allowing things to kick off in the first lines of the install. So now it seems to be running on both cores; at least they are both mostly occupied(between 80-100%) with nothing else going on. Now if I could just get DIMES to register, it'd be all set!

I had the same problem as Mr. Flynn here. The ole' "OpenSCManager failed: Access is denied. (error 5)" Solved it as above but I have one simple question. Is one instance of folding all I want? CPU usage is at 100% but I used to have two instances running on my AMD dualie, so not sure.

Thanks in advance.

[powerarmour]

Well, that was it, folks; UAC wasn't allowing things to kick off in the first lines of the install. So now it seems to be running on both cores; at least they are both mostly occupied(between 80-100%) with nothing else going on. Now if I could just get DIMES to register, it'd be all set!

I had the same problem as Mr. Flynn here. The ole' "OpenSCManager failed: Access is denied. (error 5)" Solved it as above but I have one simple question. Is one instance of folding all I want? CPU usage is at 100% but I used to have two instances running on my AMD dualie, so not sure.

Thanks in advance.

If your running on Windows (I assume you are) then just run one instance of the SMP client on your Quad, otherwise the MPI gets even more unstable than it is already. The is a new 5.92 Windows SMP client with an MPI replacement, but that only works on 32bit OS's AFAIK.

If you running Linux then I say run two instances/clients per Quad for a decent PPD increase, though Stanford recommend that just one instance/client is run because it slows down the science. Though the PPD is due to be revised soon if WU's are returned faster, so it may be better all round in the future running a single client.

[Linky]

Yup. It's on my Vista x64 set-up, which requires 5.91. Am I high or did they take down the 5.92 version:

http://www.stanford.edu/group/pandegrou ... nload.html

It was up there last night and I actually fiddled around with it before I realized that it only worked on a 32-bit OS. It uses the Deino MPI doo. That seemed a bit more complicated than the MPI that comes with 5.91 (forgot what it's called).

Should I expect quicker turn around times on WUs with this or kind of a same turn around but more points per WU or some other experience? Just curious.

And thanks for your help. Cheers.

[powerarmour]

Yup. It's on my Vista x64 set-up, which requires 5.91. Am I high or did they take down the 5.92 version:

http://www.stanford.edu/group/pandegrou ... nload.html

It was up there last night and I actually fiddled around with it before I realized that it only worked on a 32-bit OS. It uses the Deino MPI doo. That seemed a bit more complicated than the MPI that comes with 5.91 (forgot what it's called).

Should I expect quicker turn around times on WUs with this or kind of a same turn around but more points per WU or some other experience? Just curious.

And thanks for your help. Cheers.

The 5.92 is still up on this url :-

http://folding.stanford.edu/English/Download

But yeah I wouldn't be surprised if it was taken down, at least they should mention that it's 32bit only. I would just like to see a v6 x64 SMP compatible Windows client already. In any case the 5.91 is pretty damn stable on Vista x64, I never have any problems with it.

Running a single SMP client on a quad results in rapid WU turn around times, points are the same as if you would get on a dualie, but the WU's are crunched nearly twice as fast...

[Linky]

Cool...but even twice as fast is not gonna get close to the numbers you posted above (~4300ppd or something with two Linux SMPs) PA.

One more quick thing. After I installed the MPI, I turned UAC back on. That's not gonna affect me being able to run the client is it? I know it's running right now but is it gonna run normally?

Danke.

[powerarmour]

Cool...but even twice as fast is not gonna get close to the numbers you posted above (~4300ppd or something with two Linux SMPs) PA.

The Linux client is just faster though, but it's only the port of MPI that holds the Windows client back really. If you run a single Linux client on a quad the CPU usage never hits 100%, so you can gain a bit of speed by running two instances instead, but the Windows SMP client is always 100% anyway, as that's more likely down to the fact that the MPI hogs more of the CPU time.

Maybe there is a bonus to running two instances of the Windows SMP client, but I've never found that to be the case, maybe some of our other wise members can enlighten...?

One more quick thing. After I installed the MPI, I turned UAC back on. That's not gonna affect me being able to run the client is it? I know it's running right now but is it gonna run normally?

Danke.

Don't think so, but then again I wouldn't really know as UAC is the first thing I disable on a Vista install...