computer instantly crashes during a render

[idmatic]

Hello all!

I have an interesting problem, one that someone might be able to explain.

I have a 3dBoxxSE4800 with 8GB Ram, clocked to 4GHz. Single HD..no raid.
XP64

The computer has been working fine up until 3 weeks ago. I thought something might be up with the vidcard (FX4600) so I swapped it out for another one of a dif brand and spec...when I went to render from 3DSMax\Vue Ex7...it doesnt even give a BSD...it just blackscreens and reboots.

Any info would be appreciated.

[SandmanNinja]

Have you considered undoing the overclocking (or does Boxx ship at that clock speed)?

If you've changed the video card when the OS/hardware was unstable, it may not be able to display any video output - good or bad.

Sounds like you need to boot into Windows Safe Mode and see if that helps any. Google your specific version of windows and Safe Mode - it will tell you the keys to press when it's booting.

[idmatic]

Thanks for the info,

Boxx ships the unit clocked at that speed.

As for the video card, it started doing this hard-reset with the original grfxcard..and then I swapped it for another (installing the drivers or course) with the same result.

Im wondering if there is some kind of motherboard fault or overheating that occurs when rendering starts since it seems to suck on the juice pretty heavily when this happens.

I'll def look into safeboot.

Again, thanx for the help. Its always appreciated.

[Cad64]

Looks like you've already identified a possibility, (overheating), that I was going to suggest. Make sure all your fans are working properly. What type of cooling system do you have? I've got a pretty massive cooling unit in my machine and my processor is only overclocked to 3.20 Ghz, so I would imagine you would need something pretty substantial for 4.0 Ghz.

[idmatic]

Yeah...im thinking it might be this and really dreading that thats the case...
Call me crazy...but i find it ultimately stupid to have a water block cooler being cooled by fans and not a semiconductor cooler ontop of the main fluid resevoir. The machine has to large fans in front of the tank and then one after it with a cheap plastic wind tunnel to direct "cooled" air over the ram modules...there is the addition of a mini fan on one of the chips via heatsink...but nothing directly cooling the ram...or exhausting the inner chamber center to that of the MOBO. I added 2 80mm fans on the backside and at least am getting some better cooling. i Just feel that its still running too hot...

I have these stats from the bios and wonder if these are normal op values:
Proc 0 thermal margin 30c
Proc 1 thermal margin 48c
Proc 0 area temp 56c
Proc 1 area temp 48c
MCH temp 70c
DIMM temp 46c

My first concern is the value for MCH temp..but maybe this is normal?!?

"Looks like you've already identified a possibility, (overheating), that I was going to suggest. Make sure all your fans are working properly. What type of cooling system do you have? I've got a pretty massive cooling unit in my machine and my processor is only overclocked to 3.20 Ghz, so I would imagine you would need something pretty substantial for 4.0 Ghz."

[SandmanNinja]

What's MCH? Machine?

I run some hardware monitoring program and my core CPUs sit at 35-37 degrees and jumps up to 70 degrees when doing renders, but they cool down really quick when the render stops (seconds).

[idmatic]

MCH is apparently: Memory Controller Hub (MCH)
Methinks there be some problems with the cooling of this bugger.

Does rendering (using an FX4600) occur on board only of the GFXcard or on the MOBO..or is it shared? I ask this question since my computer knowledge is dated and at one point it was the thing to do to offload gfx processing onto the vid card...if not...then its probably a heat problem associated with the memory chips...since as the data transfer increases significantly using RAM in the case of MOBO\ram process as opposed to on board GFX proc.

Ill add...since i read these values from bios, it would mean that there was minimal proc going on...yet they were normal ambient temp (even if they are to a degree misreported) since the machine was on for a while.

Would this seem likely?

[SandmanNinja]

I heard the same about the off-loading of some of the grunt work to the GPU, but cannot confirm that. My hardware monitoring program (HWMonitor.exe) monitors my GPU temp as well as hard drives, etc. I guess it's possible but the GPU temp does not go up when I render.

There is a special Max video driver for 3ds and Quadra cards, but I don't know if that will help offload grunt work or not.


Intel Core i7 920 hardware monitor
Temperature sensor 0 37°C (98°F) [0x3F] (Core #0)
Temperature sensor 1 36°C (96°F) [0x40] (Core #1)
Temperature sensor 2 38°C (100°F) [0x3E] (Core #2)
Temperature sensor 3 34°C (93°F) [0x42] (Core #3)

GeForce 9800 GT hardware monitor
Temperature sensor 0 35°C (95°F) [0x23] (GPU Core)

[F J]

do WinKey+Pause/Break > Advanced > Startup n Recovery > uncheck Auto-restart so that u can see the error when it crashes..

[BOXXLABS]

Hi John-Michael,

Sorry to hear you're having issues - I trust you've contacted our technical support team about this? I know they are always eager to help resolve issues.

We have hundreds of these systems being "banged on" by customers every day with only very few reports of failures or instability.

I know our engineering team spent a TON of time tuning/optimizing the system design from a thermal management standpoint. Even under full processing load, the thermal margins should be safely within spec as long as everything is working as designed. Further modification of the system should not be needed under normal operation.

In any case, I know the guys would like to know about any issues you are having with the machine with the aim of getting you back to work.

877-877-BOXX(2699)