Restarts and temps

[gobeavers]

I just built a new computer recently with an Athlon 64 3200+ , Antec Super lanboy case, 6800 GT, and MSI K8N Neo platinum. I get program crashes (opera, AIM, Outlook Express), and the occasional random reboot. I don't know what is causing this. The nVidia system utility reports idle CPU temps at 48 C, case at 44 C, and my GPU at 58 C. My bios reports idle CPU temp at 57.

1. Which tempature is more likely to be correct? I have the 2 in/out fans on the lanboy running, also I have a pci slot cooler below the 6800 GT. All cooling is stock.

2. Could tempature be the cause of my problems, if not, what is?

Thanks for any help!

EDIT: I am running SP1 and 1 GB of RAM. After program crashes it comes up with an error reporting thing if that helps any...

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton<P ID="edit"><FONT SIZE=-1><EM>Edited by gobeavers on 08/28/04 08:02 PM.</EM></FONT></P>

 

[peteroy]

The temps look alright to me.

Maybe it not hardware related, make sure all drivers are WHQL, that you have latest DirectX 9.0c

<A HREF="http://www.download.com/Spybot-Search-Destroy/3000-8022_4-10289035.html?tag=pop" target="_new">Install Spybot Search and Destroy to remove any spywares.</A>

<A HREF="http://housecall.trendmicro.com" target="_new">Scan for viruses</A>

===========================
<A HREF="http://www.clancas.net" target="_new">clan CHAOS</A>

 

[Cybercraig]

Go fish around <A HREF="http://forum.msi.com.tw/board.php?boardid=28&sid=4ff8360ad19990bb6fe206936399bfc4" target="_new">here</A> for the latest BIOS revisions. It's hard to get that board to report temps correctly. By the way, how do you like the Lanboy case?

Abit IS7 - 2.8C @ 3.4ghz - Mushkin PC4000 (2 X 512) - Sapphire 9800Pro - TT 420 watt Pure Power
Samsung 120gb ATA-100 - Maxtor 40gb ATA - 100
Sony DRU-510A - Yellowtail Merlot

 

[gobeavers]

WHQL? I have the latest release drivers for my 6800 GT and motherboard. Those are the only drivers I need IIRC...

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[gobeavers]

I like the super lanboy case a lot. 120mm fans so its not to loud, and there are hard drive dampners and such.

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[gobeavers]

I just had another restart..for a second it displayed a blue screen (not a BSOD) and a memory error. I had backed off the timings before so I don't think timings are a problem.

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[Johanthegnarler]

I'd do a memtest86 make sure your ram is ok. If it's not heat related most likely its the memory.
Don't forget the a64 although not picky on the speed of the memory the board is picky on what brand of memory.


<A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=277124623" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=277124623</A>
46,510 , movin on up. 48k new goal. Maybe not.. :/<P ID="edit"><FONT SIZE=-1><EM>Edited by Johanthegnarler on 08/29/04 03:06 PM.</EM></FONT></P>

 

[gobeavers]

Did that and it came up with like 26k errors
Taking it back tomarrow. Thanks for all your help!

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[addiarmadar]

Make sure you are using the latest chipset drivers for the mobo. Also make sure to get the latest bios version for the MOBO for 57c in the bios, supposedly no load, and getting 48c in windows nvidia utility, with little more load is raising some flags with me.

<i><font color=red>Only an overclocker can make a computer into a convectional oven.</i></font color=red>

 

[endyen]

I got the Asus K8N-E board on Saturday. It started my pc3200 at pc1600 speeds. I was pissssed. I played around for a while, and got it up to pc2700 speeds. Still choked, I swapped for some pc3200, I was using in another system. Same thing. Ran some benches, and a slight OC, and went into bios. Now all the ram works at pc3200.
Moral is, the ondie controller takes time to get up to speed. Now even my timings are down to 2.5, 3, 3, 6. when at pc1600 they were 3, 4, 3, 8.
Try setting your mem to 100 fsb, with relaxed timings, then up them slowly.

 

[gobeavers]

Well, since I found all of those RAM errors, I think that is the problem, unless there is supposed to be a certain number of them...hopefully I can return that RAM tongight and get some new stuff.

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[gobeavers]

Chances are, the RAM is the problem, right? I don't want to get my hopes up...

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[endyen]

Your problem is in the memory cct. Most of the time, it is the ram, but it could also be from too little voltage ( up the v-dimm to 2.7) or the ondie controller (like I said try it at 100 fsb, with latencies of 3 ,4, 3, 8), but it also could be a bad trace on the mobo.

 

[gobeavers]

Ok. The saga continues. I tested to see which stick was the problem causer, so I tested.
Tested one dimm alone. Clean.
Tested the other dimm alone. Nothing.....huh?
Tested both in dimms 1+2. Errors...hmmmm.
Tested in both in dimms 1+3. No boot.
Continued to test, and now I have them in 2+3, and mem-test detects no errors, but I will need a week of having no problems before I am convinced.

Is this typical RAM behavior?

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[endyen]

No, but it is not unusual for socket 754. If the traces are not the same length, between chip and dimm, the ODMC will give errors. With use, the ODMC gets a little more relaxed. Play doom3 or do something intensive, like run prime95 for a couple of hours, as a burn-in.
BTW the old bios for your mobo reads the sensor wrong. It reports the chip temps about 10c too high. It's not a big deal because temp readings are never right, but if you want, you can update the bios and get a more accurate reading.

 

[gobeavers]

Forgive my ignorance, but with the traces, would those change between each individual board? I haven't heard of any other problems with the MSI K9N neo platinum. I don't really care if the bios reads high, as long as the nvidia utility is somewhere near accurate.
Now, I didn't have these problems for the 1st couple of days with my computer, does that explain anything? It is working now, but I will wait a week before I am confident in it...

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[endyen]

Actually, you have heard of this problem with all socket 754 boards. That's why all the bs about memory compatability. Yes two identicle traces could be different to the ODMC. Even a very slight flaw in the copper could cause problems, when working with tight memory. The "problem" will go away with useage however.
The dimm socket furthest from the chip is usually the most problem, but that can vary. Better OEMs will do a burn-in to get rid of the problem.
You should be fine from here on, though a little heavy use is a good idea.
In other words, go play, enjoy your new system.
By the way, many of the 865/875 setups, that use LL memory have the same problem.

 

[gobeavers]

Ok, thanks endyen, you have explained and helped a lot.

One last thing
Been running prime 95 on "blend" mode. It went on a streak of 100% CPU usage, where the nVIDIA tool showed it starting to really slow down the temp climb at 62 (C). It only ran probably between 5-10 mins, but it was climbing pretty slow near the end, so i would guess only another degree or so if it ran longer. Is this okay for a CPU temp? I run stock everything, a 120mm intake and outake, and then a pci slot cooler.

"Go forward until the last round is fired and the last drop of gas is expended...then go forward on foot!" -Patton

 

[endyen]

Leave it longer, and see if it stops on errors.
Dont worry, prim95 will stop long before it damages your chip. I would bet that your diode is still reading 10c too high, even in the nvidia prog.

 

[Johanthegnarler]

That's a little odd if you ask me, but i'm going to trust you.
I have only put one 754 system together for a friend..
Vnf3 chaintech with 512x2mushkin basic (75bucks a pop)and it had no memory errors with memtest86, and i never had a problem with any BSOD or any type of errors in general.

I did however burn in the memory, and cpu right after i did memtest out of the box. I usually do a memtest right after OS install.

<A HREF="http://arc.aquamark3.com/arc/arc_view.php?run=277124623" target="_new">http://arc.aquamark3.com/arc/arc_view.php?run=277124623</A>
46,510 , movin on up. 48k new goal. Maybe not.. :/