*** F@H Tech Assistance Thread ***

[sneekypeet]

System is completely stock.

I use the high performance client with the purty display page.
Should I try changing to another client, and if so which one?

log file:


Top is where it was failing on the unit I posted above....has since kicked into action with a few restarts, on the WU in the bottom half( currently working 240/8000).

Thanks Bog, Id really like to add the 280's PPD.

 

[bogmali]

Yeah you might want to switch to the console client version, link HERE

LMK if you need more help

 

[sneekypeet]

1) vista 64 ready correct? currently running it must be.

2) it just runs, there is no install?

3) I seem to be stalled at: Starting GUI server. Working fine at 2% currently.

4) if it doesnt install, how do I track it via Fahmon? Sorted

EDIT: had to add this, as I found it funny. Same WU and the 275 @ FTW clocks is hammering my 280 @ stock clocks

 

[bogmali]

EDIT: had to add this, as I found it funny. Same WU and the 275 @ FTW clocks is hammering my 280 @ stock clocks

View attachment 28474

Better put some OC on that 280 I will boost mine's voltage tonight and OC it gradually

Glad everything is working out for you

 

[theonedub]

im getting EUE's again...

Its a cool day and the GTX is barely OCed. This sucks.

Rebooted, deleted work files, and disabled the 50mhz shader clock and all is normal. Im thinking this was a server side issue again. Im going to restore the OC and see what happens.

 

[BUCK NASTY]

im getting EUE's again...

Its a cool day and the GTX is barely OCed. This sucks.

Rebooted, deleted work files, and disabled the 50mhz shader clock and all is normal. Im thinking this was a server side issue again. Im going to restore the OC and see what happens.

Did it pause for 24hr?... Some EUE's are normal when running a healthy OC. I get 6-8 per day scattered across my farm. As long as the client re-starts on it own and continues, you should be ok. If you get "pause for 24hr", then you must lower your clocks.

 

[theonedub]

Did it pause for 24hr?... Some EUE's are normal when running a healthy OC. I get 6-8 per day scattered across my farm. As long as the client re-starts on it own and continues, you should be ok. If you get "pause for 24hr", then you must lower your clocks.

It did say pause for 24hrs. Can it really be the clock though? It ran for nearly 2 weeks straight in 90-104F temps without issue and its only a +50mhz boost. It would be a shame to run this thing at stock clocks. What else can I look for?

 

[hertz9753]

It did say pause for 24hrs. Can it really be the clock though? It ran for nearly 2 weeks straight in 90-104F temps without issue and its only a +50mhz boost. It would be a shame to run this thing at stock clocks. What else can I look for?

I was getting "unstable machine" and "pausing for 24 hours" after I had a stick of ram in my computer give out.

 

[theonedub]

How often were you getting the errors? Was it hourly, daily, weekly, monthly, etc? I hope thats not the case either. Gskill ram is usually pretty reliable, but if it is the ram hopefully its one of the 1gb sticks not the 2.

 

[hertz9753]

It was daily, but it was happening when I was at work. I would come and find those messages, restart the computer and it would fold fine the rest of the evening. that lasted for about two days. On the third day I got a BSOD while I was sitting in front of it. Turned out to be a memory error.

 

[bogmali]

Here is one minor issue that I've yet to solve; I have 2X GTS250 and a 8800GT in a 780i board. The GTS250's occupy GPU slots 1 and 3 while the 8800GT is sandwiched in between.

As you can see on the screenie, the GT has more PPD's than the 250's
Anybody know why? Does this have something to do with the board having a 8X8X8 PCI-E footprint (or is it 16X4X4) when all three slots are filled?

EDIT: Just checked the board specs and it says true 3 X16 PCI-E lanes so WTF?

 

[theonedub]

So they are all working on the same WU in that screenshot? I also thought the rule was only group together GPUs with the same amount of shaders or risk lower performance? Just looking at the simple reasons first

 

[bogmali]

So they are all working on the same WU in that screenshot?

Yup, 353 pointers.

I also thought the rule was only group together GPUs with the same amount of shaders or risk lower performance? Just looking at the simple reasons first

Well, 8800GT is 112 shader and the GTS250 is 128. In essence if you look at it 8800GT=9800GT (same die) and evolved to 9800GTX/GTX+ (with die shrink) = GTS250. I have a K9A2 with 2 GTS250 and 2 8800GT also and it is not having this issue

 

[theonedub]

Thats def strange. I know the G92 core is shared between them all- was thinking more about the shader count. Hopefully this gets answered, Id like to know what caused this too.

 

[mike047]

Yup, 353 pointers.

Well, 8800GT is 112 shader and the GTS250 is 128. In essence if you look at it 8800GT=9800GT (same die) and evolved to 9800GTX/GTX+ (with die shrink) = GTS250. I have a K9A2 with 2 GTS250 and 2 8800GT also and it is not having this issue

Are the 250's dropping to 2d clocks??

That is an issue with the 250, occasionally. A reboot usually cures the condition.

 

[bogmali]

Are the 250's dropping to 2d clocks??

That is an issue with the 250, occasionally. A reboot usually cures the condition.

Nope, rebooted and I thought I saw FAHMON showing and almost even PPD across the board and then it went back to the same thing I will probably switch GPU's around to isolate the issue.

 

[thebluebumblebee]

Bog, I think you have a "golden" 8800gt, because I only get 6354PPD with my 8800GTS/512 on the same WU! 4236 with the 9600GSO

 

[El Fiendo]

Have you removed the 8800GT to see if the PPD goes back to normal for the GTS 250s? And also, that 8800GT is pulling near the same PPD as one of my GTX 260s on a 353. And its overlocked.

 

[bogmali]

Have you removed the 8800GT to see if the PPD goes back to normal for the GTS 250s? And also, that 8800GT is pulling near the same PPD as one of my GTX 260s on a 353. And its overlocked.

Reshuffling the cards right now to see if this boards PCI-E lanes is the issue. The GT is stock
600/1500/900 so just like bumblebee said, I might have a "Golden" 8800GT

 

[BUCK NASTY]

Reshuffling the cards right now to see if this boards PCI-E lanes is the issue. The GT is stock
600/1500/900 so just like bumblebee said, I might have a "Golden" 8800GT

780i lanes are not designated the same as standard boards. Normally GPU-0 is closest to the CPU. On a 780i with 3 slots filled, GPU-0 refer's to the center slot, with GPU-1 being closest to the CPU and GPU-2 being the farthest. I had to find out the hard way with my 780i. I will admit, hands down, this is the best Mobo I have ever owned(aside from the Nb heat).

 

[bogmali]

OK after removing the 8800GT, this is what it looks like:

I think I'm going back to my original plan of slapping all 4 GTS250's on one of the K9A2 board while the 3 8800GT's will occupy the other K9A2

 

[bogmali]

This is more like it Slapped all 4 GTS250's in a K9A2 mobo:

I will post screenies of the 3X8800GT's on the 780i later

 

[JayliN]

http://forums.techpowerup.com/attac...28515&stc=1&d=1252897322[/QUOTE] :respect:

 

[theonedub]

Glad to see you got that worked out Bog. Very impressive output- im jealous! lol

Quick PSU question for you guys though. Will my Rosewill 450w (12v1@14A 12v2@16A) be enough to power my computer (WCG1 in specs) with a GTX260 216 in it instead of the 9800GT? Trying to lock down an upgrade

 

[El Fiendo]

Mmm Bogi, you're turning me on with pictures like that.