Page 4 of 12

Re: As the Work Unit Crunches

Posted: Thu May 28, 2009 10:12 pm
by Ragnar Dan
I normally don't look for such cards, but scrolling through the deals it came up and I couldn't resist. It kept me up late considering, and then I did some calculations vs. the 8800GTS and decided it was probably a pretty decent deal if it works. But I really do have to move fast to verify that and get it back in time if it doesn't. I only remember buying one open box thing before, and it worked fine, so I don't think I have experience returning open box items.

That 295 of yours is a great deal. Don't worry too much about the MSI brand, I just see people on Newegg reviews and a few other places complaining and started to get the notion that their quality has declined. They used to make good stuff (at least one of their dual K7 boards was well regarded back in the day), so they know how to do things right. If you don't OC you'll probably be fine. That does seem like the smartest way to go, though. Whenever that thing comes I'm going to have to tear apart at least one machine and maybe 2 to get one it fits inside of while also holding too many hard drives. I have to get rid of some the ones I've got.

I should look for something to test the card with... I haven't run a game in a while, so I've got nothing except the possibility of downloading a demo, I suppose. And of course there's the killer of all hardware, folding (which I will do first). But pretty screenshots are more fun. :wink:

Re: As the Work Unit Crunches

Posted: Fri May 29, 2009 3:07 pm
by Ragnar Dan
Well, the vidcard's not supposed to get here until Tuesday, so I will have this weekend to measure the space I have, compare it to what I find others saying about their 260 sizes, and move drives around. I may have to remove my upper drive cage, and to do that means only 3 drives total (or 4 if I set one on its side at the case's bottom as I'm doing "temporarily"), which would mean getting rid of drives I've meant to get rid of already and consolidating them to larger ones. Hours of work, but hopefully it will be worthwhile. I should look for software to move installs in the registry, and other fun things like that.

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 7:35 pm
by Gerbil Jedidiah
I just lost 26% on another %&)!@#*$)#(%^)@&^$)&%)($#_@)$(%_)@^&#@$(&(!*&%#)($^@#)_$^&@#$(%&$!@#()$_!#$^&*!#_)$^&#@)!*%!)$^*@#!)$^&#!)$&^!)*(%!@#)$%(!)#$^&#!)$&^!)#@$(%!_$)!+_$^(!#$&%(#&$!*(&#!*(&#@(%$_)^%_^+@_#$%(@#)$%*($&%(*&!$(&#%$!)_%(+!#$^)@!#$+_/// WU. Why can't they make the SMP clients run on 8 cores? (waits on pins and needles for obligatory thread/core comment :wink: )

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 7:58 pm
by Flying Fox
Bite the bullet, do multiple VMs. The more rows you have on FahMon, the geekier you can feel prouder of yourself. ;)

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 8:02 pm
by Ragnar Dan
Were I less lazy and had a better memory this time of day, I probably wouldn't have to ask this, but, don't you know that it's 4 cores and 8 thre--- [ :P] you're running all of them from the same user account? If so, are you [telling shortcuts which cores to run on | adding affinity switches to shortcuts to launch them]?

Assuming I'm thinking within reason of reality, maybe you can run the clients from different user accounts and get some protection that way. I can't remember what the WinSMP version of the mpiexec is in the version you run (from the download page, it looks like it may be MPICH), but it may help protect them if each is on its own account. As long as each account has a different directory from which to launch, I don't see why it wouldn't work, unless the MPICH thing acts like a system service and doesn't care who owns the launching program. One easy way to find out is to test it. :)

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 8:28 pm
by Gerbil Jedidiah
No need to give me advice guys(for now). I'm just ranting about the sad state of the SMP client. I'm tired of babysitting affinities, priorities, VMs, etc... I just want the @#$@$#%^@ things to work and fully utilize my system. and make me pancakes. Yes, especially pancakes.

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 8:34 pm
by Ragnar Dan
My questions and suggestion assuming the answers to them remain unanswered.

Edit: BTW, I read the other day that apparently that 260-216 can use up to 282 watts all by itself. Sheesh.

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 9:15 pm
by Gerbil Jedidiah
Ragnar Dan wrote:
My questions and suggestion assuming the answers to them remain unanswered.


I have no idea what you meant by this. Yes I'm slow. Maybe you mean rhetorical?


Edit: BTW, I read the other day that apparently that 260-216 can use up to 282 watts all by itself. Sheesh.


Not while folding though I imagine. I have 720 shaders total on a 750w PSU and system runs fine. I'm actually a bit amazed at the load this PSU is handling... GTX295, GTX280, and i7 920 @ 3.6GHZ all 100% utilized... Rig just purrs along, if only the SMP software would catch up...

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 9:39 pm
by Ragnar Dan
Gerbil Jedidiah wrote:
Ragnar Dan wrote:
My questions and suggestion assuming the answers to them remain unanswered.

I have no idea what you meant by this. Yes I'm slow. Maybe you mean rhetorical?

I asked if you had the SMP clients set up such that they were all launched by shortcuts specifying processor affinity. Then, assuming you had things that way, I suggested that instead of launching all of the clients from one user account, you launched each one from a different account. Of course when I wrote that I was forgetting about how Windows only lets desktop OSes have one user running at a time... but let's just forget that part for now. :oops:

There may be a way to get 2 four-[core/thread] SMP clients running from different accounts, though. At least that might protect things a bit.

Gerbil Jedidiah wrote:
Edit: BTW, I read the other day that apparently that 260-216 can use up to 282 watts all by itself. Sheesh.

Not while folding though I imagine. I have 720 shaders total on a 750w PSU and system runs fine. I'm actually a bit amazed at the load this PSU is handling... GTX295, GTX280, and i7 920 @ 3.6GHZ all 100% utilized... Rig just purrs along, if only the SMP software would catch up...

Well, the biggest PSU I have is a Silverstone 600W (ST60F), and it's ~3.5 years old. Hopefully it will run that card, but I might have to buy a power cable adapter for it for the other 6 pin connector it requires (I haven't opened the case yet for a particular reason). I'll definitely have to remove a drive cage to fit it if I want PCIe-16x for it, though.

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 9:59 pm
by Gerbil Jedidiah
Ragnar Dan wrote:
I asked if you had the SMP clients set up such that they were all launched by shortcuts specifying processor affinity.


How you do that?

Right now I just click on all 5 shortcuts and let them duke it out for supremacy. GPUs are at a slightly higher priority, but I haven't messed with affinities.

Re: As the Work Unit Crunches

Posted: Mon Jun 01, 2009 10:50 pm
by Ragnar Dan
I don't know what it is to get different threads (if that's even possible), but I think for cores it's "start /affinity N program.exe" where N is the core(s) you want it to use, in some form which I forget off hand.

It's probably either bit mask form or something similar, and assuming it's bit mask form, that would mean something like: 1 is for the first core (2^0), 2 for the second (2^1), 4 for the third (2^2), and 8 for the fourth (2^3). Don't take my word for it, though, since I haven't looked it up. But if it works that way, then 12 would be the fourth and third cores, 7 would be for third, second, and first, 3 for the first two cores, etc.

Re: As the Work Unit Crunches

Posted: Tue Jun 02, 2009 9:10 pm
by Ragnar Dan
The GTX260 came today, and it's will be a beast to fit it in the intended case, and I wonder if I can evacuate the excess heat, too. It will, oddly enough however, fit in the case I bought in 1998 for my Celeron 300A@450. I'd put it in there, too, but I want to see if I can fit it in my main system first for testing purposes. The 1998 case, an Inwin A500 or some such thing, is for the 8800GT-256 card.

One good thing is that it seems to have come with everything a new card should have, including DVI->VGA, DVI->HDMI, and the part I really need, the dual molex->6-pin-PCIe power connector. Maybe tomorrow I'll be able to install it. Drive rearrangement is time consuming.

Re: As the Work Unit Crunches

Posted: Tue Jun 02, 2009 9:40 pm
by Gerbil Jedidiah
Ugh, I installed VMPlayer today, and now my PC won't reboot. PPD is gonna be down. :-? :-? :-?

Re: As the Work Unit Crunches

Posted: Tue Jun 02, 2009 10:56 pm
by Gerbil Jedidiah
OK I got two GPUs back up. and the SMP clients... I'll try to get the last GPU working tomorrow.

Re: As the Work Unit Crunches

Posted: Tue Jun 02, 2009 11:26 pm
by Flying Fox
GJ, your experience pretty much sums up why the TR community is not as involved as before and the result that we are dropping fast in the rankings. It's the sad state of affairs of these new fangled clients that even experienced folders finding the problems a little too troublesome for our taste. Sure we are more technically minded and are not afraid of changing settings here and there, but there is a limit on the amount of trouble we are willing to tolerate just to keep the clients up. Until Stanford addresses some of these issues I don't think our participation can improve a whole lot, although the huge points advantage of the GPU client can offset the loss in points from farms whenever we get people with single machines with single GPU to join up.

Why the **** do we have to babysit so much? I just got a couple of 2665's EUE'ed on me too. Seems like this box can only do 2653's but it never bothered to try the A2 WUs. Stupid assignment server. :evil:

/rant

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 9:10 am
by Gerbil Jedidiah
Flying Fox wrote:
GJ, your experience pretty much sums up why the TR community is not as involved as before and the result that we are dropping fast in the rankings. It's the sad state of affairs of these new fangled clients that even experienced folders finding the problems a little too troublesome for our taste. Sure we are more technically minded and are not afraid of changing settings here and there, but there is a limit on the amount of trouble we are willing to tolerate just to keep the clients up. Until Stanford addresses some of these issues I don't think our participation can improve a whole lot, although the huge points advantage of the GPU client can offset the loss in points from farms whenever we get people with single machines with single GPU to join up.

Why the **** do we have to babysit so much? I just got a couple of 2665's EUE'ed on me too. Seems like this box can only do 2653's but it never bothered to try the A2 WUs. Stupid assignment server. :evil:

/rant


I wonder how feasible it would be to make the SMP WUs smaller. Then an EUE wouldn't matter as much. If you error out at 99% it's much more of a PITA if it took a day to get there than it would be if it only took 3 or 4 hours.

Yeah TR is doing respectably in the folding ranks, but we could be so much better. I've been assisting noob system builders like crazy in the hopes of bringing in some new folders. I think I've had minor limited success. I think it's a better use of our time to get more people folding on a single GPU client than it is for a few of us to see how much hardare we can turn into a folding farm (although I like this and will keep doing it).

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 12:16 pm
by Ragnar Dan
FF: I agree that's a problem, but why does it seemingly affect TR's team more than others?

We've been left in the dust by 4 teams in the last year or so. Do you think that it's a simple matter of others recruiting a lot more than we are and even though they're having a similar rate of teammates quitting it doesn't affect them as much?

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 1:06 pm
by Gerbil Jedidiah
Can't wait to find this puppy in an "open box" deal on Newegg: http://www.nordichardware.com/Reviews/? ... 554&page=2

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 5:43 pm
by farmpuma
Gerbil Jedidiah wrote:
~Closing one Win SMP client crashes the other~

IIRC, the SMP client spreads out over as many processors as it can find, so when running two SMP clients you are sharing already shared (HT) cores. I imagine Windoz gets confused and just shuts them all down causing the second client to crash.

How many PPD do you lose when running only one Win SMP client instead of two? A single SMP WU crunching on eight processors should crunch fairly fast and it might increase your GPU PPD.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 6:00 pm
by Flying Fox
farmpuma wrote:
Gerbil Jedidiah wrote:
~Closing one Win SMP client crashes the other~

IIRC, the SMP client spreads out over as many processors as it can find, so when running two SMP clients you are sharing already shared (HT) cores. I imagine Windoz gets confused and just shuts them all down causing the second client to crash.

How many PPD do you lose when running only one Win SMP client instead of two? A single SMP WU crunching on eight processors should crunch fairly fast and it might increase your GPU PPD.

The default (-smp 2) setting spawns 4 instances of FahCore_A?.exe. In theory you should do (-smp [cores]) to maximize CPU usage. The WUs should finish faster but they may not be as much points as running multiple SMP clients, similar to the situation of HT'ed P4's with 2 single core clients.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 6:18 pm
by Gerbil Jedidiah
Flying Fox wrote:
farmpuma wrote:
Gerbil Jedidiah wrote:
~Closing one Win SMP client crashes the other~

IIRC, the SMP client spreads out over as many processors as it can find, so when running two SMP clients you are sharing already shared (HT) cores. I imagine Windoz gets confused and just shuts them all down causing the second client to crash.

How many PPD do you lose when running only one Win SMP client instead of two? A single SMP WU crunching on eight processors should crunch fairly fast and it might increase your GPU PPD.

The default (-smp 2) setting spawns 4 instances of FahCore_A?.exe. In theory you should do (-smp [cores]) to maximize CPU usage. The WUs should finish faster but they may not be as much points as running multiple SMP clients, similar to the situation of HT'ed P4's with 2 single core clients.


Here is my client config line. How should I change it?

[settings]
username=UnitedGerbilNation
team=2630
passkey=
asknet=no
machineid=3bigpackets=big
extra_parms=-smp
local=16

[http]
active=no
host=localhost
port=8080
usereg=no

[clienttype]
type=3

[core]
addr=

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 7:11 pm
by Flying Fox
Gerbil Jedidiah wrote:
Flying Fox wrote:
The default (-smp 2) setting spawns 4 instances of FahCore_A?.exe. In theory you should do (-smp [cores]) to maximize CPU usage. The WUs should finish faster but they may not be as much points as running multiple SMP clients, similar to the situation of HT'ed P4's with 2 single core clients.


Here is my client config line. How should I change it?

[settings]
extra_parms=-smp

It should be "-smp [x]", where x > 2. You should try 4 first and then 3, see how many instances of FahCore_A?.exe is spawned in the process. I suppose you have the latest 6.23 beta?

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 9:19 pm
by Ragnar Dan
That was fun...


I had a devil of a time smashing a straight SATA connector under that dang card. It really doesn't fit, but I got it under there. I decided to do a new install of XP Pro, which won't activate, tonight.

Meanwhile, the vid card is not clocked as the specs on Newegg list them, and it only has 192 shaders. It's at 576-1242-999, which I assume is approximately standard for such cards.

I began folding on it about 3 hours ago, and it's not that impressive. It was faster than my other cards on the 768 pointer it finished up, but not incredibly much faster.

Then it downloaded a 353 pointer, and it was a bit faster but noticeably affected by browsing. So I began overclocking it. It sped up to a decent speed, ~6500 PPD when nothing else is moving on the screen.

Then I sped it up some more and it crashed the WU (even though it was running no higher than 76C), and then downloaded a 511 pointer. I figured it would crunch through that better than my other cards ever could. Instead, it's the slowest of any card I've got. It's running it at 2:56 for each %, or about 2508 PPD. My 9600GSO will do over 3044 PPD when I leave it untouched, and the 8800GT-256 does ~3679 PPD. I don't know what the problem may be, but it's not impressing me. Since it doesn't meet the specs Newegg lists for the card, they should be embarrassed about it and refund my return shipping cost. But they won't. I'm fairly pissed they did this.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 9:37 pm
by Gerbil Jedidiah
Ragnar Dan wrote:
Since it doesn't meet the specs Newegg lists for the card, they should be embarrassed about it and refund my return shipping cost. But they won't. I'm fairly pissed they did this.


That's BS. They should refund shipping if they send you the wrong product. You must have spoken to someone in a bad mood.

They let me return my ASUS Striker mobo, open box version, after 2 months. So sometimes they are good.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 9:42 pm
by Ragnar Dan
Well I haven't talked to them yet. I'm just assuming, given the pissant warranty paragraph they show about open box products, that they'll do as little as possible, but hopefully you're right. I'll call them tomorrow, or maybe email them in the next hour or so...

I found my reboot sped up the 511-point WU, though, so there's one bright spot. :wink: It went from 2:56 to ~1:45 per % (4204 PPD). I think it must have been the installation of VMware or whatever the other thing I can't recall that was slowing it down for whatever reason.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 10:11 pm
by Gerbil Jedidiah
Ragnar Dan wrote:
Well I haven't talked to them yet. I'm just assuming, given the pissant warranty paragraph they show about open box products, that they'll do as little as possible, but hopefully you're right. I'll call them tomorrow, or maybe email them in the next hour or so...

I found my reboot sped up the 511-point WU, though, so there's one bright spot. :wink: It went from 2:56 to ~1:45 per % (4204 PPD). I think it must have been the installation of VMware or whatever the other thing I can't recall that was slowing it down for whatever reason.


Well whatever I did the other day when I installed VMware has really f'd up my computer. I'm having fits trying to get it to work... It hangs, BSODs while running stock, internet doesn't work, list goes on and on... I now am back to where I started. Two SMP clients and 2 GPU clients running... I will try again tomorrow to get the third card working. I suspect there is a screwed up PCI-E driver keeping the GTX280 from being recognized.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 10:36 pm
by Flying Fox
Ragnar Dan wrote:
FF: I agree that's a problem, but why does it seemingly affect TR's team more than others?

We've been left in the dust by 4 teams in the last year or so. Do you think that it's a simple matter of others recruiting a lot more than we are and even though they're having a similar rate of teammates quitting it doesn't affect them as much?

There are lots of factors, I have a few points that hopefully shed some light in the picture:
  • The TR community itself is smaller than the others. This is a fundamental issue that is not going to change any time soon. Compare to other sites like AnandTech and [H] we simply don't have as big communities as theirs. To solve this we need to dramatically step up our recruitment and retaining efforts. Problem is some of us old-timers are somewhat elitistknowledgeable and massive recruitment may mean dumbing down the audience. Notice the relatively "meh" reaction to TR joining the new fancy stuff like Facebook and you can get a glimpse of what kind of people usually hang around here. /no-offense-just-an-observation
  • Now, we may be small, but the fact that we have been hanging on for so long is a testament to something. I did a quick Excel exercise and it showed. I took the top 25 teams from EOC and calculated the "total points per active user". TR is at #5 at 2.8 million points per active user. If you look at the 5 teams above us they are near the bottom actually. So we just have a relatively small group but they are already pretty productive.
  • The other teams that are surpassing us most likely have a lot of members folding on one or 2 machines. We seem to have lots of people with a farm of some sort, thus jacking up the per user production numbers. This can be a focus of our next recruitment drive in which we ask people to fold on one GPU on one computer. Hopefully we can get more people this time.
  • The TR community also prides itself in living in the "bang for buck" realm. However, this hurts us in our folding efforts because most of us are more cost conscious and in general are just more mature. So the hotter weather, a/c costs, etc. all lead to people not folding as much as they can. Just look at how UGN and other big-time folders downsizing their farms and we can see how much that hurts us.
  • Some of our past leaders have farms of CPU nodes, but the SMP client is just in such a bad state that they have quit altogether (drfish and leor are the prime examples). For the corporate folders GPU folding is usually not accessible to them.
  • When we lost drfish we lost our de facto captain so to speak. UGN, FBW, FH, and PinTO have been carrying the torch but with such a disparity of points among them, it does not capture the imagination. We need more mid-range gauntlets!
  • Our other "drumming up support" efforts fizzled most of the time. We need to improve on that.

Re: As the Work Unit Crunches

Posted: Wed Jun 03, 2009 11:16 pm
by Gerbil Jedidiah
Flying Fox wrote:
Ragnar Dan wrote:
FF: I agree that's a problem, but why does it seemingly affect TR's team more than others?

We've been left in the dust by 4 teams in the last year or so. Do you think that it's a simple matter of others recruiting a lot more than we are and even though they're having a similar rate of teammates quitting it doesn't affect them as much?

There are lots of factors, I have a few points that hopefully shed some light in the picture:
  • The TR community itself is smaller than the others. This is a fundamental issue that is not going to change any time soon. Compare to other sites like AnandTech and [H] we simply don't have as big communities as theirs. To solve this we need to dramatically step up our recruitment and retaining efforts. Problem is some of us old-timers are somewhat elitistknowledgeable and massive recruitment may mean dumbing down the audience. Notice the relatively "meh" reaction to TR joining the new fancy stuff like Facebook and you can get a glimpse of what kind of people usually hang around here. /no-offense-just-an-observation
  • Now, we may be small, but the fact that we have been hanging on for so long is a testament to something. I did a quick Excel exercise and it showed. I took the top 25 teams from EOC and calculated the "total points per active user". TR is at #5 at 2.8 million points per active user. If you look at the 5 teams above us they are near the bottom actually. So we just have a relatively small group but they are already pretty productive.
  • The other teams that are surpassing us most likely have a lot of members folding on one or 2 machines. We seem to have lots of people with a farm of some sort, thus jacking up the per user production numbers. This can be a focus of our next recruitment drive in which we ask people to fold on one GPU on one computer. Hopefully we can get more people this time.
  • The TR community also prides itself in living in the "bang for buck" realm. However, this hurts us in our folding efforts because most of us are more cost conscious and in general are just more mature. So the hotter weather, a/c costs, etc. all lead to people not folding as much as they can. Just look at how UGN and other big-time folders downsizing their farms and we can see how much that hurts us.
  • Some of our past leaders have farms of CPU nodes, but the SMP client is just in such a bad state that they have quit altogether (drfish and leor are the prime examples). For the corporate folders GPU folding is usually not accessible to them.
  • When we lost drfish we lost our de facto captain so to speak. UGN, FBW, FH, and PinTO have been carrying the torch but with such a disparity of points among them, it does not capture the imagination. We need more mid-range gauntlets!
  • Our other "drumming up support" efforts fizzled most of the time. We need to improve on that.


I mostly agree, but the team's PPD is almost double what it was in a few months. We are having some wins.

What if we could start a competition where we try and take each other down... keep a ladderboard and competitions could be marketed on the front page... Could be neat.

Re: As the Work Unit Crunches

Posted: Thu Jun 04, 2009 1:31 am
by farmpuma
Flying Fox wrote:
The default (-smp 2) setting spawns 4 instances of FahCore_A?.exe. ...

Thanks for the correction, Flying Fox. I did not recall correctly and I need to stay a bit more current with the SMP FAQ.

F@H SMP FAQ wrote:
The SMP client was originally intended for multi-core CPUs, which generally do not support HT. For machines with 2 physical CPUs, we do recommend enabling HT for the SMP client as this presents the operating system with what looks like 4 logical processors (and our SMP client is intended for 4 processors). If you have 4 physical CPUs, we recommend against using HT, as this presents the operating system with 8 logical processors, which will make the SMP client run inefficiently (especially since the logical processors coming from HT run much slower than the normal ones).

Re: As the Work Unit Crunches

Posted: Thu Jun 04, 2009 7:46 pm
by Ragnar Dan
Ragnar Dan wrote:
Well I haven't talked to [Newegg] yet. I'm just assuming, given the pissant warranty paragraph they show about open box products, that they'll do as little as possible, but hopefully you're right. I'll call them tomorrow, or maybe email them in the next hour or so...

I emailed them (or what passes for email with Newegg) and they sent me a link to a prepaid shipping label but haven't said anything about the cost of shipping to me, though there's some mealy-mouthed stuff about open box items: "In this case, this item might be packed and labeled incorrectly when it was returned. For open box items, we are unable to issue replacement. Instead, we can issue a refund RMA for this item without the restocking fee and a prepaid shipping label." I doubt they're going to refund my cost.

Maybe I can sue them for not paying me a minimum due wage for verifying their item listings for them. :evil:

I'm still P.O.'ed.