• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Mixing GPU Series?

nitrobass24

[H]ard|DCer of the Month - December 2009
2FA
Joined
Apr 7, 2006
Messages
10,477
Ok so i have heard that you can run G92 and GT200s in the same box without performance issues.

What I am thinking of doing is making all the boxes that I already run for other task double as folders.
So I am going to put a K9A2 in my HTPC and I have 3 GTX275s going into it. Can I put a 8800gs in there just for video output for the HTPC? or a Radeon 3450 for that matter.
It wouldnt be folding just for DXVA.
 
To date that has mixing a G92 (8xxx, 9xxx series, GTS 2xx) with a GT200 (GTX 2xx series) is impossible to make fold on the G92.

But I noticed a beta FaHCore_11.exe (version 1.31) that apperently Stanford is attempting to fix this. Additionally if contains fixes for improperly deleted check points from hosing you on a restart. No additional performance tho, I'm running it on 2 of 3 GPU clients here (all GTX 260s) and it seems fine, I haven't had the usual 1 in 5 failed WUs, so they might have fixed something

Read more here

I guess I should have made a thread about this, because proper mixing of G92 and GT200 would be a big deal for a lot of people.

 
Last edited:
Yea but I wouldnt be folding on the G92.

I would only fold on the 200's
 
Oh, if that's the case, been there done that, works just fine, no performance hit. GTX260 192 + 8800GS, I used it for physx in games as I could not make it fold. Ultimately upgraded to Dual GTX260s in my sig

Tho I'd be interested to know if that new core does fix the mixing problem. I'm sure others might like to know as well.
 
To date that has mixing a G92 (8xxx and 9xxx series) with a GT200 (GTX 2xx series) is impossible to make fold on the G92.

But I noticed a beta FaHCore_11.exe (version 1.31) that apperently Stanford is attempting to fix this. Additionally if contains fixes for improperly deleted check points from hosing you on a restart. No additional performance tho, I'm running it on 2 of 3 GPU clients here (all GTX 260s) and it seems fine, I haven't had the usual 1 in 5 failed WUs, so they might have fixed something

Read more here

I guess I should have made a thread about this, because proper mixing of G92 and GT200 would be a big deal for a lot of people.


This would be my problem with mixing my GTS250 and 8800GTS320 or 9600GT together.
 
The GTS 250 is a G92 core, not GT200 core like the GTX series. GTS != GTX.

Confused me for about a year after release, then somebody was nice enough to tell me.

GTS250 will be recognised as G92 based.
 
This would be my problem with mixing my GTS250 and 8800GTS320 or 9600GT together.
Nope, those cards are all compatible with each other. Any 8000 series, 9000 series, and GTS250 cards can be used together, and any GTX260, GTX275, GTX280, GTX285, or GTX295 cards can be used together, but the two groups cannot be mixed.
 
Nope, those cards are all compatible with each other. Any 8000 series, 9000 series, and GTS250 cards can be used together, and any GTX260, GTX275, GTX280, GTX285, or GTX295 cards can be used together, but the two groups cannot be mixed.

read what I linked to, it would seem that stanford is working on fixing that. I don't know if they have it fixed just yet but they claim that the core_11 beta adressess this incompatability

and for what it's worth, I did not stupidly confuse GTX (GT200) and GTS (G92), I just neglected to mention that particular rebadging in my post. I'll go ahead and go back and edit that now to aleviate your confusion.
 
read what I linked to, it would seem that stanford is working on fixing that. I don't know if they have it fixed just yet but they claim that the core_11 beta adressess this incompatability

and for what it's worth, I did not stupidly confuse GTX (GT200) and GTS (G92), I just neglected to mention that particular rebadging in my post. I'll go ahead and go back and edit that now to aleviate your confusion.

How mean of you!:D
Is the beta up of core_11 (1.31? I think) up yet? Or is it just going to be pushed automatically a few weeks from now?
 
My 2x GPU clients downloaded the beta core automatically.
 
Oh my! Well, guess that means it still might be my mobo then, haha. ;)
 
A link to the beta is in the article. look for the word "here" that is a slightly different shade of gray.

"s and can be downloaded from here if you wis" (it's in this sentence)
 
My 2x GPU clients downloaded the beta core automatically.

how'd that happen?

I manually installed the .exe between WUs (run -oneunit, change exe, and restart)

for the record, while I have not seen any performance increase, I have seen stability increase.
ever since the 185.85 drivers one of my 260's would EUE about one in five WU. So far after 36 hours of running the beta on that card no EUEs. I have even upped the shader clock to 1566 (the card would EUE every unit on drivers after 185.85, any higher and I get display crashes on the desktop). I wouldn't realistically consider this a performance increase. It works out to be 1 second per frame. or just over 1.5 minutes per work unit. So you'd have to fold for days just to squese in one extra WU.
 
what should I look for in FAHMon to see if I got the new core? Someone said they got it auto-DL right?

HMM - I'm still on 1.19 It says in Core version
 
Yea mine says "1.19 - beta"

Just delete the core and force it to download a new one.
 
but it will still download the most recent "released" core.

You need to manually download, and over write the 1.19 core with the 1.31 core.

FAHmon won't tell you the core you are using unless you use it read back in the logs to see which version was used when your client loaded the WU.

FAHspy will tell you which core you are using as one of the fields that it has when giving info about thw WUs you have.
 
FAHspy will tell you which core you are using as one of the fields that it has when giving info about thw WUs you have.
FahSpy will display the core type but it doesn't display core version. Unless it's changed in a newer version of the monitor. Speaking of which, what happened to the site?
 
I stand corrected. I was thinking of HFM.net

still better than drunken posting tho.
 
Back
Top