F@H - New X6800 & X1950 XTX

KE1HA

KE1HA

KE1HA

Associate
Joined
3 Jul 2004
Posts
866
Location
Helena, Montana
Well ... I finially broke down down and went to the big boy's toy store today.

Loaded up two shopping trollys :D

CPU - X6800
RAM - 2GB of XMS2
MB - ASUS P5W DH Deluxe
GRAPHICS - ATI X1950XTX
SOUND - X-Fi Platinum
KB - Logitech G15
RAT - Razer Diamondback
SOUND MAKERS - Logitech Z-5300e
MONITOR - NEC 90GX2
2x DL DVD's
and all the other Jazz required to keep things ticking over.

I've fired up two CPU cores and the GPU. I hope i've got the config correct. I followed the instructions on the sticky, but have a couple fo quesiotns:

--> Do I need to leave one CPU core Idle ?
--> What about flags ? do we set them the same as we would for the CPU's?

I must say, this thing is seriously fast compared to anything i've ever had before. Iit blew right through the WinDoze installation & updates like it was Linux.

[EDIT]
Opps, Spoke to soon ...Getting errors on the ole GPU :mad:

What do you make of this ?
[08:52:36] Trying to unzip core FahCore_10.exe
[08:52:36] Decompressed FahCore_10.exe (2298880 bytes) successfully
[08:52:36] + Core successfully engaged
[08:52:41]
[08:52:41] + Processing work unit
[08:52:41] Core required: FahCore_10.exe
[08:52:41] Core found.
[08:52:41] Working on Unit 01 [October 22 08:52:41]
[08:52:41] + Working ...
[08:52:41] - Calling 'FahCore_10.exe -dir work/ -suffix 01 -priority 96 -checkpoint 15 -service -verbose -lifeline 2364 -version 590'

[08:52:45] CoreStatus = C0000135 (-1073741515)
[08:52:45] Client-core communications error: ERROR 0xc0000135
[08:52:45] Deleting current work unit & continuing...
[08:52:49] - Warning: Could not delete all work unit files (1): Core returned invalid code
[08:52:49] Trying to send all finished work units
[08:52:49] + No unsent completed units remaining.
[08:52:49] - Preparing to get new work unit...
[08:52:49] + Attempting to get work packet
[08:52:49] - Will indicate memory of 2047 MB
[08:52:49] - Connecting to assignment server
[08:52:49] Connecting to http://assign-GPU.stanford.edu:8080/
[08:52:49] Posted data.
[08:52:49] Initial: 41AB; - Successful: assigned to (171.65.103.159).
[08:52:49] + News From Folding@Home: GPU folding beta
[08:52:49] Loaded queue successfully.
[08:52:49] Connecting to http://171.65.103.159:8080/
[08:52:50] Posted data.
[08:52:50] Initial: 0000; - Receiving payload (expected size: 87593)
[08:52:50] Conversation time very short, giving reduced weight in bandwidth avg
[08:52:50] - Downloaded at ~171 kB/s
[08:52:50] - Averaged speed for that direction ~171 kB/s
[08:52:50] + Received work.
[08:52:50] + Closed connections
[08:52:55]
[08:52:55] + Processing work unit
[08:52:55] Core required: FahCore_10.exe
[08:52:55] Core found.
[08:52:55] Working on Unit 02 [October 22 08:52:55]
[08:52:55] + Working ...
[08:52:55] - Calling 'FahCore_10.exe -dir work/ -suffix 02 -priority 96 -checkpoint 15 -service -verbose -lifeline 2364 -version 590'

[08:52:59] CoreStatus = C0000135 (-1073741515)
[08:52:59] Client-core communications error: ERROR 0xc0000135
[08:52:59] Deleting current work unit & continuing...
[08:53:03] - Warning: Could not delete all work unit files (2): Core returned invalid code
Click to expand...
.
.
 
Last edited:
KE1HA said:
Well ... I finially broke down down and went to the big boy's toy store today.

Loaded up two shopping trollys :D

CPU - X6800
RAM - 2GB of XMS2
MB - ASUS P5W DH Deluxe
GRAPHICS - ATI X1950XTX
SOUND - X-Fi Platinum
KB - Logitech G15
RAT - Razer Diamondback
SOUND MAKERS - Logitech Z-5300e
MONITOR - NEC 90GX2
2x DL DVD's
and all the other Jazz required to keep things ticking over.

I've fired up two CPU cores and the GPU. I hope i've got the config correct. I followed the instructions on the sticky, but have a couple fo quesiotns:

--> Do I need to leave one CPU core Idle ?
--> What about flags ? do we set them the same as we would for the CPU's?

I must say, this thing is seriously fast compared to anything i've ever had before. Iit blew right through the WinDoze installation & updates like it was Linux.
.
Click to expand...

just imagine what linux would do!


iirc you need to leave one core free for the GPU to poll.
 
Possibly. If both the GPU client and CPU clients are set at the same Idle level, good chance the GPU may not be getting a look in!

I suggest getting rid of one of the CPU clients, then setting the GPU clients priority to LOW.

See how it goes. Oh, gorgeous rig btw :D Curious tho... why were you crazy enough to spend THAT much on the X6800...?

SiriusB
 
SiriusB said:
Possibly. If both the GPU client and CPU clients are set at the same Idle level, good chance the GPU may not be getting a look in!

I suggest getting rid of one of the CPU clients, then setting the GPU clients priority to LOW.

See how it goes. Oh, gorgeous rig btw :D Curious tho... why were you crazy enough to spend THAT much on the X6800...?

SiriusB
Click to expand...
I wanna know what kinda clocks he gets out of it!


its strange, both my E6300+ds3/4 PCs have reached 3.15 with ease, but will not clock any higher! the DS4 just sets it back to 3.15 if I up the FSB more!
 
Looks Like the Direct-X Driver was the problem, or at least that problem.
[18:21:21] Trying to unzip core FahCore_10.exe
[18:21:21] Decompressed FahCore_10.exe (2298880 bytes) successfully
[18:21:21] + Core successfully engaged
[18:21:26]
[18:21:26] + Processing work unit
[18:21:26] Core required: FahCore_10.exe
[18:21:26] Core found.
[18:21:26] Working on Unit 01 [October 22 18:21:26]
[18:21:26] + Working ...
[18:21:26] - Calling 'FahCore_10.exe -dir work/ -suffix 01 -priority 96 -checkpoint 30 -service -verbose -lifeline 2280 -version 590'

[18:21:26]
[18:21:26] *------------------------------*
[18:21:26] Folding@Home GPU Core - Beta
[18:21:26] Version 0.08 (Fri Oct 20 11:06:45 PDT 2006)
[18:21:26]
[18:21:26] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 13.10.3077 for 80x86
[18:21:26] Build host: CYGWIN_NT-5.1 vishal-gpu 1.5.19(0.150/4/2) 2006-01-20 13:28 i686 Cygwin
[18:21:26] Preparing to commence simulation
[18:21:26] - Looking at optimizations...
[18:21:26] - Created dyn
[18:21:26] - Files status OK
[18:21:27] - Expanded 86981 -> 443705 (decompressed 510.1 percent)
[18:21:27]
[18:21:27] Project: 2711 (Run 1, Clone 3, Gen 12)
[18:21:27]
[18:21:27] Assembly optimizations on if available.
[18:21:27] Entering M.D.
[18:21:33] Working on Protein
[18:22:08] Starting GUI Server
Click to expand...
.

I haven't OC'd either the GPU nor the processor, as I wanted a good benchmark on the stock setup, then I'll up the watts and see how it goes :D

The CPU core is doing a 2124 segment in about 17 minutes or so at 2.9Ghz.
.
 
I think i finially got it figured out.

Installed 6.10 Drivers
Installed Latest DirextX Drivers
Downloaded the Client

When I ran it the first time, I instaled it as a service. Now Im running the standard client, seems to be working.
[04:00:41] + Processing work unit
[04:00:41] Core required: FahCore_10.exe
[04:00:41] Core found.
[04:00:41] Working on Unit 01 [October 23 04:00:41]
[04:00:41] + Working ...
[04:00:41]
[04:00:41] *------------------------------*
[04:00:41] Folding@Home GPU Core - Beta
[04:00:41] Version 0.08 (Fri Oct 20 11:06:45 PDT 2006)
[04:00:41]
[04:00:41] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 13.10.3077 for 80x86
[04:00:41] Build host: CYGWIN_NT-5.1 vishal-gpu 1.5.19(0.150/4/2) 2006-01-20 13:28 i686 Cygwin
[04:00:41] Preparing to commence simulation
[04:00:41] - Looking at optimizations...
[04:00:41] - Files status OK
[04:00:41] - Expanded 86994 -> 443705 (decompressed 510.0 percent)
[04:00:41]
[04:00:41] Project: 2711 (Run 1, Clone 0, Gen 4)
[04:00:41]
[04:00:41] Assembly optimizations on if available.
[04:00:41] Entering M.D.
[04:00:47] Will resume from checkpoint file
[04:00:47] Working on Protein
[04:00:48] Starting GUI Server
[04:00:52] Resuming from checkpoint
[04:00:52] Verified work/wudata_01.log
[04:00:52] Verified work/wudata_01.edr
[04:00:52] Verified work/wudata_01.trr
[04:00:52] Verified work/wudata_01.xtc
[04:00:52] Completed 1
[04:08:45] Completed 2
[04:16:40] Completed 3
[04:24:32] Completed 4
[04:32:22] Completed 5
[04:40:20] Completed 6
Click to expand...
For some reasone, using ATI Tools and ATI Overclocking, I cant seem to get it to clock up, but will keep working on that one.

Have the X6800 clocked up to:
Core Speed: 3.4
Multiplier: x11.0
Bus Speed: 306.6
Rated FSB: 1226

Dont know how that compares elsewhere, but it's a start.
.
 
Last edited:
I got my E6600 to run 1MB superpi at 3.6 (400x9) but it wasn't stable enough to fold with. I did run it for a while at 3.5 but the temps were a bit high - it's currently running at 3.15 with lower volts to keep the temperature down. I'm going to swap it over with the E6300 into the PC62 case as the cooling is much better and probably swap the Ninja which is cooling it for the Arctic Pro 7 from the E6300 and see how it fares with that.

I would imagine you could get the E6800 a bit higher than that but it all depends on the chip and the setup you have.

Stan :)
 
Bigstan said:
I got my E6600 to run 1MB superpi at 3.6 (400x9) but it wasn't stable enough to fold with. I did run it for a while at 3.5 but the temps were a bit high - it's currently running at 3.15 with lower volts to keep the temperature down. I'm going to swap it over with the E6300 into the PC62 case as the cooling is much better and probably swap the Ninja which is cooling it for the Arctic Pro 7 from the E6300 and see how it fares with that.

I would imagine you could get the E6800 a bit higher than that but it all depends on the chip and the setup you have.

Stan :)
Click to expand...

I've broken something :(

Swapped everything over as quoted (except the coolers - just reseated the Ninja). E6600 is working like a dream (couldn't resist making it crossfire - now I'm a GPU thread down :o ).
Put the E6300 back togeher in the other case (TJ07), switched on PSU (fan turning), switch on pooter and nothing, absolutely SFA. I have checked the obvious to no avail. CBA to faff about with it tonight so I'll get back to it tomorrow.

:mad:

Stan :)
 
Double check everything is seated correctly! I know you said you have checked the obvious but this can be a blind spot for some people :p

If you still cant get it to work, better donate it all to me :D

SiriusB
 
Took it apart, tried it nekkid - no problem.

Put it back together, tried again - smoke :eek:

Switched off quickly and took apart again.
There seems to be something about that case it doesn't like (worked fine in the other case) :confused:

I may have to run it nekkid for now until I get another case :(

Stan :)
 
You must log in or register to reply here.
Back
Top Bottom