Well if you really want the air cooled ones, those are $1430. I don't know if I would want to convert a $700+ card to water cooling myself yet. The water blocks don't seem out of line with DIY either.
Dan

That's a good bit cheaper. Putting water block on gpus......the hardest part about most of them is figuring out how to take them apart. But since the 200 series it's just a few screws and some plastic clips and everything comes right off. It's very DIY, simple to do now.

---

Traveling through space at ~67,000mph!

It is half a dozen one way and 6 the other. The Koolance water blocks are $150 each so $300 for the pair, or exactly what EVGA wants for 2 water blocks and the ones from EVGA are already installed. Guess it just depends on if you think the water block design is better in the aftermarket or if EVGA got it right.
Dan

---

It is half a dozen one way and 6 the other. The Koolance water blocks are $150 each so $300 for the pair, or exactly what EVGA wants for 2 water blocks and the ones from EVGA are already installed. Guess it just depends on if you think the water block design is better in the aftermarket or if EVGA got it right.
Dan

For Me It would be easier to install two Hydro equipped cards into a PC, Instead of installing two blocks on two cards and then installing the two cards in a PC.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

You guys have read about all the issues with these cards, right?

---

You guys have read about all the issues with these cards, right?

I'm not worrying about overclocking, If that's all You're talking about, As I don't do that now anyway. Are there any other issues?

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

It is half a dozen one way and 6 the other. The Koolance water blocks are $150 each so $300 for the pair, or exactly what EVGA wants for 2 water blocks and the ones from EVGA are already installed. Guess it just depends on if you think the water block design is better in the aftermarket or if EVGA got it right.
Dan

That's why you get a better block at a cheaper price and get the $120 DangerDen blocks. The stock solution is just that, stock. After market coolers on average have a much better flow rate or better cooling characteristics than the EVGA stuff.

$865 - From EVGA water cooled.
$715 + $120 - DangerDen cooled

Save yourself $30 per card, $60 total that you can spend on upgrading something else. Of course this isn't including shipping for those of you keeping count. Either way you will come out cheaper as long as you don't leave in the boonies.

You guys have read about all the issues with these cards, right?

Unless something new has happened that has since been fixed in drivers, by limiting the voltage like they should have been.(?) Personally I still wouldn't buy them, I just haven't seen any dual card solutions that are worth the money. If they would just figure out a way of putting two top end cards on one PCB and keep it cool enough it would be a go, this whole we put two lower end cards on one PCB just doesn't do it for me. Yeah it's faster....unless someone has two top end cards in there. Speaking gpu chip for chip not card for card of course.

---

Traveling through space at ~67,000mph!

It is half a dozen one way and 6 the other. The Koolance water blocks are $150 each so $300 for the pair, or exactly what EVGA wants for 2 water blocks and the ones from EVGA are already installed. Guess it just depends on if you think the water block design is better in the aftermarket or if EVGA got it right.
Dan

That's why you get a better block at a cheaper price and get the $120 DangerDen blocks. The stock solution is just that, stock. After market coolers on average have a much better flow rate or better cooling characteristics than the EVGA stuff.

$865 - From EVGA water cooled.
$715 + $120 - DangerDen cooled

Save yourself $30 per card, $60 total that you can spend on upgrading something else. Of course this isn't including shipping for those of you keeping count. Either way you will come out cheaper as long as you don't leave in the boonies.

You guys have read about all the issues with these cards, right?

Unless something new has happened that has since been fixed in drivers, by limiting the voltage like they should have been.(?) Personally I still wouldn't buy them, I just haven't seen any dual card solutions that are worth the money. If they would just figure out a way of putting two top end cards on one PCB and keep it cool enough it would be a go, this whole we put two lower end cards on one PCB just doesn't do it for me. Yeah it's faster....unless someone has two top end cards in there. Speaking gpu chip for chip not card for card of course.

I looked at the Evga 590 Hydro cards, Shipping is for the pair at least(I didn't pay attention to the others): FREE...

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

All just a matter of preference on either buying the EVGA hydros or buying the air cooled and adding a block. All I am saying is that right now, I don't feel comfortable with ripping into an $730 card and installing a water block and void the warranty in the process. If you are comfortable doing that and want to, fine by me.
I understand the design of the Danger Den block may provide slightly improved cooling, just haven't seen any comparisons between the two. I would hope anyone using these cards water cooled, whether purchased from EVGA or hand modified, would have a very large radiator with lots of fans moving air through it and high flow pump just to make sure things don't heat up.
Dan

---

All just a matter of preference on either buying the EVGA hydros or buying the air cooled and adding a block. All I am saying is that right now, I don't feel comfortable with ripping into an $730 card and installing a water block and void the warranty in the process. If you are comfortable doing that and want to, fine by me.
I understand the design of the Danger Den block may provide slightly improved cooling, just haven't seen any comparisons between the two. I would hope anyone using these cards water cooled, whether purchased from EVGA or hand modified, would have a very large radiator with lots of fans moving air through it and high flow pump just to make sure things don't heat up.
Dan

Changing heat sinks don't void the warranty on an EVGA card. You merely have to reinstall the stock sink should you ever have to RMA the card. It's totally preferance, however I am not talking about you directly? So don't install them, I'm merely talking about the issue being raised by it being more expensive to buy them already installed or not and how good the stock sinks are on them.

As far as radiators are concerned, the only put out marginally more heat than a 480 etc. so a single 360mm rad per card would be the order of business to have a 'performance' delta of no more than 10c over room temp. You could probably get by with 2x on a 480mm rad if you are only looking for cool and quiet and not performance. I cool a GTX 480 and a q9650 on a single 360mm rad and they said it wouldn't perform good and I stay within 15-18c of room temp until it starts getting near 30c in the room, which is almost never. So performance is highly subjective, some people are happy with their air temps.

Now when it comes to 'slightly' better cooling that's an understatement. There was shown to be a 6-8c different from previous hydro series cards and after market higher end heatsinks. In reference to EK, DangerDen, Koolance etc. It has nothing to do with 'voodoo' in the blocks to make them better it's all about restriction. Also the reason you don't put ram, hdd, or chipset water cooling blocks on the same loops as your cpu and gpu if you can help it. The added restriction kills the loop. Will they be good enough? Sure, will they be optimal especially for people who are water cooling for performance? Probably not. I mean after all if you can manage to install water into your tower and run it around your $700 card what's removing 17 screws and applying thermal paste? I mean really....especially when you come out better on the money. Just my $.02.

---

Traveling through space at ~67,000mph!

Ok, i have now a few Values more:

WUs   GPU Load   Mem Contr Load  Mem Used   WU Time

4       89%          43%          675 MB     880 Sec.
6       94%          43%          904 MB    1230 Sec.
8       95%          45%         1216 MB    1610 Sec.

...

A few values with overclocked GTX590 (i do 8 at a time)
normal values for Shader/Mem are 607/854:

WUs   GPU Load   Mem Contr Load  Mem Used   WU Time    Overclocked 
                                                       (Shader/Mem)
8       98%          48%         1216 MB    1495 Sec.    650/875
8       98%          50%         1216 MB    1435 Sec.    668/908

Helli

Outstanding! It appears that card is going to be a monster.

---

Traveling through space at ~67,000mph!

I would say, one GTX590 can do the work of two GTX295 with 1.5x times the Power Draw of a GTX295. :-)
The GTX590 are very hot, so i'm not sure if it's possible to run more than one Card with no room
between each other. Maybe only watercooled...

Helli

I tried that, with 2x 480 & 1x 590.
Got way toooooo hot. > 90C and plenty of noise. 3 fans at 100% is .....

Moved the 480 over to another system with an older card.

Now have 1x 480 and 1x 590, with one card (a lot) of space between then.

67C (590) 71C 77C with 25C ambient, may casing is open at one side.
Fan 73% and 85% ,so some room left.

Had to remove the SLI from the 590. When I did that the errors I got where gone.

---

TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking.

I tried that, with 2x 480 & 1x 590.
Got way toooooo hot. > 90C and plenty of noise. 3 fans at 100% is .....

Moved the 480 over to another system with an older card.

Now have 1x 480 and 1x 590, with one card (a lot) of space between then.

67C (590) 71C 77C with 25C ambient, may casing is open at one side.
Fan 73% and 85% ,so some room left.

Had to remove the SLI from the 590. When I did that the errors I got where gone.

If I had a setup like that, I would spend the extra money and water cool it. I know it's a broken record around here, but, that's a lot of heat to be sticking in ANY case without some pretty extreme cooling. If it was water cooled properly none of those cards would ever break 45-50c with them all in together. Granted it would cost about what a 590 costs to cool it that way which is not a draw.

As far as the power draw goes on the card, keep in mind this card is aimed at those people wanting SLi without needing to buy two cards. And this thing uses less power than 2x 580's which I believe is what it's meant for, I would buy a pair of them if I had the cash to throw at it.

Hummm dual xeon with 2x 590's would be a crunching machine for real.

---

Traveling through space at ~67,000mph!

Bit off topic but you guys are mentioning running more than one WU per GPU. Is that possible on my GTX 465? If so how do I go about setting that up?

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

Bit off topic but you guys are mentioning running more than one WU per GPU. Is that possible on my GTX 465? If so how do I go about setting that up?

In app_info change all instances of

<type>CUDA</type>
<count>1</count>

to

<type>CUDA</type>
<count>0.33</count>

Then you'll be running 3 WUs which I think would be best for your card. If you only want 2 WUs, just change it to 0.5 :)

Bit off topic but you guys are mentioning running more than one WU per GPU. Is that possible on my GTX 465? If so how do I go about setting that up?

In app_info change all instances of

<type>CUDA</type>
<count>1</count>

to

<type>CUDA</type>
<count>0.33</count>

Then you'll be running 3 WUs which I think would be best for your card. If you only want 2 WUs, just change it to 0.5 :)

That was a bit over my head. Could someone toss me a private message (so I can stop cluttering this thread) with an idiots howto version? It'd be much appreciated.

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

Thanks for asking the question Slavac as I was wondering the same thing.

I have hunted all over for the "app_info" file and can not find it. I am using Bionic version 6.10.60.

I am running two GTX-580's in SLI. Once I figure out how to change the setting, any one have a recomendation on how many WU I should run per card?

Thanks!!

Thanks for asking the question Slavac as I was wondering the same thing.

I have hunted all over for the "app_info" file and can not find it. I am using Bionic version 6.10.60.

I am running two GTX-580's in SLI. Once I figure out how to change the setting, any one have a recomendation on how many WU I should run per card?

Thanks!!

You won't have an app_info because you're running the Stock apps.

Claggy

2+ WUs/Fermi (GTX4xx-5xx) GPU:
NC subforum : 'Running SETI@home on an nVidia Fermi GPU - Message 1029675'


You can let run stock (which I do) also optimized CUDA app with app_info.xml file.

For newbies, installation with Lunatics Installer and you have the app_info.xml file in the project folder (check my profile for 'quick instruction').


- Best regards! - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC. - SETI@home needs your help. -

---

Hi Slavic,
In your app_info file you will find entries like this...
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<platform>windows_intelx86</platform>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>.5</count>
</coproc>

Where it says <count>.5</count> that means I'm running two work units at a time. If you want to run three at a time change that to .33. I am running a 64 bit system so I had to change six entries. If you are running a 32 bit you will only find three.

Also, anytime you make changes in your app_info be sure to use notepad and make sure it is saved as a .xml file.

edit note, I made the count bold.

---

PROUD MEMBER OF Team Starfire World BOINC