PDA

View Full Version : CPU temp the same as Core temp?


Vilandra
04-29-2009, 10:30 PM
Ive got a Boxx workstation (8-cores running at 4Ghz) and I wanted to know what the temp was for both CPU's.

I used Everest, and it says the following...

CPU 1: 44C
CPU 2: 43C

CPU1 / Core1 : 53C

And it lists all 8 cores. When im rendering, the highest temp ive seen for the cores, is pushing 80C.

Is the core temp important? It seems like its always 10-15 warmer than the CPU temp.

The reason why im asking is because lately the machine has been rebooting itself right in the middle of rendering. It doesnt do it unless im rendering. No errors, no codes, no blue screen. Just 3 seconds of freezing (the 8 renderlines will come to a stop) and then the system reboots. (Im using Cinema 4D)

I thought it may be rebooting from overheating, but I am unsure as to whether the temps mentioned above would cause it to reboot.

Adam? Got any ideas? I was hoping it would be a quick and easy fix, but I may need to call BOXX and talk with them.

Thanks
Shane

ThE_JacO
04-29-2009, 11:57 PM
Usually the core temp is a side-die sensor, so that's how hot your cpu is actually running, the cpu temp is normally provided by an on board probe, so that's usually (depending on mobo and probe placement) the temperature of the air inside the socket between cpu and board.

On my recent 920 i7, with a raven case, all silent and no exotic cooling (large downvolted fans is all I have) and full cpu load for 24 hours I never go past 55-58, and there are plenty people staying below that with air cooling alone if you don't mind 30db top mid frequencies noise.

80C for the core is entirely too high, what cpu though? Some models run hotter than others.
Said that, if you get powerdown kind of freezes and sudden reboots, 99.9% of the time it's one of two things, underpowered PSU or CPU running too hot. You definitely seem to suffer for the latter, you might have problems with the former too compounding it if the amperage coming from the psu doesn't have enough pressure or the output is too noisy.

Would definitely give boxx a call.

imashination
04-29-2009, 11:57 PM
It sounds like a typical heat problem, but 80c isnt really excessive enough to cause an instant reboot, usually it would throttle down first at least.

Another possibility is ram, as it renders it slowly uses more memory to store the output until it finishes and can dump it off to disk. If theres a bad patch then maybe hitting that would cause it, but again that would usually result in just c4d bombing rather than a full reboot.

A simple test, make a 1000x1000 plane and start copying and pasting with the task manager open, This wont stress the cpus as itll be limited to 1 core doing this, and even then its just small bursts. see if it dies doing this.

Vilandra
04-30-2009, 12:09 AM
@The Jaco, the CPU's are Core2 Quad Extreme QX9775's. Clock speed is 4Ghz. Power supply is a 1000w Silverstone. Cooling system is a liquid setup from Asetek.

I installed "speedfan" as a second source to see what it said, and it only list CPU temp, and not the core temp. Its reading the same as Everest.

@imashination, ill try the plane test and see what happens.

On a side note... I just got done running a RealFlow simulation, and had no problems with it. All 8 cores were in use, but they were not taxed at 100%. It was more like 45%.... and there was no rebooting.

Vilandra
04-30-2009, 12:15 AM
Ok, I did the little test as you suggested with the 1000x1000 plane. There was no crashing, rebooting or freezing, however when the task manager RAM usage hit 1.7GB, I got the "Not enough memory" error. However C4D did not crash.

DanielWray
04-30-2009, 12:32 AM
I'm sure there was another guy posted a few weeks back with the same issue, using cinema 4D.

Anyway, run memtest from the boot screen, or if your on vista use the built in memory tester. Also try running this to test out your CPU; http://users.bigpond.net.au/CPUburn/

If that program gives you an error or you machine restarts you know it's an overheating problem.

However, if your memory is clean, your CPU's arent overheating (Although 80c is very hot) and your not reaching the 2GB limit if your running a 32-bit OS then i'd say it could be a software conflict between the OS and 4D?

By the way, what OS are you running and is it 32-bit or 64-bit and how much RAM?

olson
04-30-2009, 12:34 AM
Last time I looked its possible for permanent processor damage to occur at 80 degrees Celsius core temperature and above. That was from AMD but I can't imagine Intel processors being that much different. How accurate the temperature sensors are is another story. :shrug:

Vilandra
04-30-2009, 12:45 AM
@DanielWray: Im running XP 32bit. I also have XP 64bit on a second drive, but havent tested this issue yet using XP64. The system has 4 Gigs RAM.

I'll try the burn test to see how that goes.

About a month ago, I had a similar issue using VRay and FinalRender. Rendering large scenes would crash, so I called Boxx and went through the process of elliminating the possibility of faulty RAM. No matter how many RAM sticks I took out, or replaced, it still kept crashing. It came down to the software, or not enough RAM as I was pushing the 4 Gig limit in XP64.

Right now, modeling in C4D, the core temps are ranging from 42C - 56C. Its rather odd, because I can just do a quick test render in the viewport, and it will reboot with me. I dont even have to render long animations. Just a quick test render for a couple seconds will do it.

DanielWray
04-30-2009, 12:50 AM
This might not help, but could you try running 4D in software mode, no OpenGL etc. Then try rendering, or rendering in the viewport?

I'm just having a look around and a few people have had the issue where 4D would crash when using accelerated mode.

Vilandra
04-30-2009, 01:14 AM
So far, no rebooting in "software" mode. But ill keep at it to see what happens. It usually only happens every 20 or 30 preview renders.

Vilandra
04-30-2009, 01:31 AM
Ok, it just rebooted with me in "software" mode. Im going to run the burn in test now to see what that does.


Burn-in test didnt produce any rebooting. While running the test, and running RealFlow, the temps for the cores were ranging from 45C to 67C. The RAM was showing from 67C to 76C.

DanielWray
04-30-2009, 11:07 AM
That's pretty wierd then, although with it still rebooting in software mode i think you can rule out the graphic card and/ or driver issues.

Now with the burn in test i noticed that i didn't reach that high of a tempature, usually my CPU idles around 53c and goes upto into the 70c's when under high load, it's a quad core 2.4ghz, just for refrence.

I remember writing in my one of my posts about your CPU hitting 80c, i'm looking over the previous posts and i can't find any refrence to that, so i'm not sure if your CPU was hitting 80c with 4D or if i just imagined that =/ Anyway if that is true, it's wierd that the burn in test only reaches those tempatures, where the rendering reaches 80c. I'm not sure if you've already done this, but could you take your computer panel off the side and check for dust, on the memory, motherboard and espeically the CPU fan and inbetween the fan and the block.

Not so long ago my machine would reboot when i did intensive CPU work, i took the side off and it was pretty clean, however i found about an inch of dust between the CPU heatsink and the Fan, it was slowing the fan down alot and also trapping quite alot of heat. it's worth checking out.

imashination
04-30-2009, 11:28 AM
Hang on, 4GHz...

Step 1: turn off the overclocking and see if it stops crashing :P

dmeyer
04-30-2009, 11:34 AM
Is this an 8450 series?

DanielWray
04-30-2009, 11:34 AM
lol i just checked, imashination is right, you have it OC'd by 800Mhz

ThE_JacO
04-30-2009, 12:12 PM
Hang on, 4GHz...

Step 1: turn off the overclocking and see if it stops crashing :P
Well, that would explaining the water boiling point temperature even with liquid cooling :)
Maybe it was one of those boxx workstations that come pre-overclocked?

DieMachinist
04-30-2009, 12:20 PM
Well, that would explaining the water boiling point temperature even with liquid cooling :)
Maybe it was one of those boxx workstations that come pre-overclocked?

Water boiling point is 100C, I thought OP was getting 80C? Do liquid coolers use actual water? I thought that it was some special liquid with magical properties :beer:

DanielWray
04-30-2009, 12:22 PM
I think they use a thicker fluid, so if a leak were to occur it would slowly seap out, rather than squirt boiling water everywhere.

If you can, down clock your CPU, to it's standard setting, which is 3.20GHz then try running cinema 4Ds rendering on a demanding scene. If it still happens it could be that the CPU has become damage, or the connecter/ motherboard connection has been damaged by the excess heat, it's rare but the PCB and connecters can warp with various tempatures.

Vilandra
04-30-2009, 03:47 PM
@DanielWray: I'll clarify about the 80C.... I dont always have everest running to show temps. In the previous posts I made, at the time I was writing them, I never saw the temp hit 80C. The highest at the time was in the mid 70's. Previous to that, I had seen the core temp hit 78 and 79, when at the same time, the "CPU" temp was still in the mid 40's.

Ive checked inside and it is dust free. Ive had the machine for about 6 months and have been using it almost every day. The dust filter on the front gets full occasionally and ill clean it, but the inside looks as clean as the day I first got it.

@imashination: Yep, 4Ghz. However ive never done any overclocking before, or the opposite, and I would really hate to mess up my machine. I think ill need to call Boxx on that one.

@dmeyer: Yes.. 8450SE. But it came OC'd from BOXX.


Last night I ran some tests, which included a very large scene in After Effects, and rendered it, and didnt have any problems, and if I understand AE and the way it renders, it makes 3 copies of itself (the .exe process) to take advantage of extra cores. So I know I had at least 3 or 4 cores going with no problems.

dmeyer
04-30-2009, 04:12 PM
I wonder if you got a bad CPU. Does it crash when running Cinebench?

Open the case, point a big box fan right at it and crank it up. Run your tests and see if it crashes.

Anyway, I'd call BOXX. The support is what you are paying for.

Vilandra
04-30-2009, 04:26 PM
Ive never had any crashing or rebooting with Cinebench.

Vilandra
04-30-2009, 05:40 PM
Is the core temp suppose to fluctuate?

Im currently rendering an animation and I have everest running to monitor the temp for all 8 cores. They seem to be fluctuating from the mid 40's, to the mid 70's within seconds. Its been rendering for the past 10 minutes with no problems, and the highest temp has reached 77C.

Its rather odd to see a core go from 75C to 48C within 3 seconds.

Srek
04-30-2009, 06:24 PM
Sounds normal to me. You have to keep in mind that the mass we are talking here is in the few gramms range. A small mass like that is easily and fast heated up and cooled down again.
Cheers
Bjrn

Vilandra
04-30-2009, 07:06 PM
Sounds normal to me. You have to keep in mind that the mass we are talking here is in the few gramms range. A small mass like that is easily and fast heated up and cooled down again.
Cheers
Bjrn


Now that you've said that, it makes sense. :lightbulb

Vilandra
04-30-2009, 09:09 PM
Ok, I decided to run R11 in XP64 to see if I could replicate the rebooting...

After several minutes, it did the same thing as before, but this time under XP64, it threw up a blue screen and all I could read was something about a physical memory dump error... or something like that. It was rather quick and I couldnt make out anything else.

Ive finally got free time now to give Boxx a call.

DanielWray
04-30-2009, 09:29 PM
If you wan't to get the bluescreen under XP64 again, but disable automatic reboot on bluescreen, by following these instructions

http://techqa.wordpress.com/2008/08/17/disable-automatic-restart-on-bsod-blue-screen-of-death/

Then if you write down all the error, it should give us and more importantly boxx a clue on what is happening. Usually when a bluescreen happens it'll dump out the memory to the HDD, thats what the part of the message that you saw was, but i don't know where windows dumps it too.

Vilandra
04-30-2009, 10:13 PM
I talked with Boxx and im currently going through each RAM stick (one by one, four 1-gig sticks) to see if I can get it to reboot.

ThE_JacO
04-30-2009, 11:38 PM
Water boiling point is 100C, I thought OP was getting 80C? Do liquid coolers use actual water? I thought that it was some special liquid with magical properties :beer:
I was saying that for scenic effect :)
The boiling point of water temperature is 100, yes, that's what the celsius scale was originally determined with. Water as a landmark substance, 0 for freezing, 100 for boiling, and then figure out the intervals.
It's 100 for demineralized water for the record, tap water with high contents of minerals and chlorine can go as high as 120.

As for the liquid used for liquid cooling, you can use anything that transmits heat and is fluid enough to not bust the pump. Normally demineralized water is ok, doesn't leave residues in the circuit and still has very good heat transport capability. Of course if you go subzero (like before things have warmed up enough in a Peltier cell cooled system) you'll need something else to lower considerably the freezing point, but still inert and non flammable and not damaging to the tubing.

There are cooling specific mixes on the market which tend to be used for commercial assemblies where you don't want users to run maintenance on the circuit, even the trivial one like vehicle replacement/refill, but they're obviously more expensive and not used in the geekfest enthusiast systems where users want to play around with things.

aglick
05-01-2009, 03:55 PM
The Liquid Thermal Conditioning system that BOXX uses is a "closed loop" system that comes "factory sealed" and contains a "glycolic" mixture not unlike the antifreeze in your car.

It has a MUCH higher boiling point and much lower thermal expansion than water.

Vilandra
05-02-2009, 12:04 AM
Just in case any of you are curious of the outcome, im still working with BOXX on a solution.

I disabled the automatic reboot, so now I was able to see the error on XP64. It had to do with the nv4_disp.dll file, which is related to the graphics driver.

I uninstalled it, and then installed a MAXON certified driver, (on XP 32, havent gotten to 64 yet) but the rebooting problem is still happening.

Ive currently got the 2nd and 4th RAM stick out, testing to see if it will reboot. As of now, it hasnt rebooted yet, and ive been rendering a few animations. If this doesnt resolve the problem (being a faulty memory issue) the next step will be bringing the CPU speed down.

DanielWray
05-02-2009, 12:31 AM
With the NV display issue it could even be a faulty GPU or the graphic card memory.

Do you have access to a cheap PCIE graphic card or PCI graphic card that you just put in and test?

Although usually when you BSOD with that issue it's down to the drivers having conflict with the operating system.

Anyway, i hope you get this fixed sood, i bet your about ready to throw it out the window!

Vilandra
05-02-2009, 12:45 AM
I do have an old PCI card in my old P4 machine, but I dont think this mainboard has a PCI slot. I'll have to check.

Im sure ill get it fixed. Im just in the middle of a job for the studio I work for, so im having to put this on hold until I can finish the job.

ArsMod
05-02-2009, 12:49 AM
Haha, pop in that old PCI Geforce3 card and see how it rolls.

In regards to the temps, 80 C is a tad high. You mentioned that at idle your CPU was still at 54 C which I would argue is still high. My Q6700 idles less than 30C and never goes above the 50-55C range on air.

You may want to look into your cooling system because if you keep running hot like that you may end up doing some damage later.

Vilandra
05-02-2009, 04:34 PM
Ok, heres what occured yesterday. I started testing the memory as the Boxx tech suggested.

I tested them in pairs. (removed the 2nd and 4th sticks, leaving the 1st and 3rd). Theres only 4 slots, and each one has a 1 gig stick in it.

With just the 1st and 3rd sticks in, I had no problems. I rendered out several animations, stills, and viewport previews with no problems.

I then took out those 2 sticks and replaced them with the other two. Still, no problems. I rendered out the same stills, animations and previews with no problems.

So I put the 2 sticks back in (now all 4 are in) and tried to rendering again. Within minutes, the machine rebooted.

Seems to only do it when all four memory sticks are in. I sent this info to Boxx, but I think their closed on the weekends.

dmeyer
05-03-2009, 04:49 PM
Bad mobo? Curious to hear how this ends up.

Vilandra
05-03-2009, 05:46 PM
I hope not, but I wont hear back from Boxx until tomorrow (Monday). Either way, its under warranty.

ThE_JacO
05-04-2009, 01:06 AM
The Liquid Thermal Conditioning system that BOXX uses is a "closed loop" system that comes "factory sealed" and contains a "glycolic" mixture not unlike the antifreeze in your car.

It has a MUCH higher boiling point and much lower thermal expansion than water.
Unless you run an evaporative tower in place of, or in addition to, radiator cooling of the vehicle, pretty much every non industrial liquid cooling system in use is closed loop :)

aglick
05-04-2009, 02:01 AM
indeed you are correct - ;)

I just mean that there's no mucking about with needing to worry about filling/refilling - it's factory sealed and should never need servicing of any kind...

that's all...

Vilandra - I am going to call over to the tech support group tomorrow morning and get the scoop on this. You shouldn't have to do a bunch of troubleshooting.... I'm going to request that they get that system in for repair or replacement.

Adam

Vilandra
05-04-2009, 02:32 AM
Thanks Adam. I appreciate the courtesy and help.

Vilandra
05-04-2009, 03:56 PM
Heres is what Boxx suggested I do next....

Ive currently got all 4 RAM sticks back in, and ive changed a few CPU settings in the BIOS.

Multiplier is now set to 9. (originally set to 10)
half-ratio option is enabled. (originally diabled)

Ive been rendering a few animations this morning without any problems. This seems to be working, however im no longer running at 4Ghz.... now its 3.79Ghz.

I thought this "skulltrail" system could handle 4Ghz plus..... :shrug:

Im waiting for them to get back with me on this.....

DanielWray
05-04-2009, 04:08 PM
Disclaimer: I don't know what i'm on about :D

Maybe it's just the CPU that you got, some hardware can't handle the excess, i don't mean boxx bought some shady CPU's i mean even CPU's from the same batch can't handle OC'ing. Who knows though?

If boxx are going to replace or repair it then that's good :)

Vilandra
05-04-2009, 04:28 PM
As of right now, they said to continue to test it (rendering) for the next day or so and see if there is any problems. If not, then they will talk about taking the performance back up.

I personally dont know what would cause a reboot at 4Ghz with all the RAM, yet no reboot with only 2 sticks. Then, with a slower speed, theres no rebooting with all the RAM back in.

I have some knowledge of this stuff, but im stumped on this one. Boxx even has their Core i7 machine at 4ghz, yet im sure it shouldnt have this problem. If you're going to have rebooting issues with RAM, then why even have a 4Ghz machine?

Im going to continue to render with it for the next day or so, and then ill get back with them. I'll post back once I know more.

Vilandra
05-06-2009, 03:21 PM
Boxx has been doing some testing on their end to confirm what is thought to be a memory bus voltage issue.

They are thinking that a memory module is not receiving enough power.

hys17
05-10-2009, 09:58 AM
80 C is way below what I had one day.

My Q9550 OC to 3.4G and I dont have any exotic cooler. So when the day turns into summer,plus my pc was doing heavy rendering 24h a day. The core temp turned up to 100C(guess u dont believe~but it gets better:)
Then I noticed the Fre downed to 2.4 and never came back.So I searched for answers.Turns out Intel has some Auto Down Fre Set to prevent to be too hot.
Then the really silly part came,I turned that off.I found the Fre wasn't turning back(happying).But few minutes rendering later,my pc shutted up.I guess that's what they called:Temp Protection...

To prevent that,I went to buy a cooler has heat pipe.
Now,no matter what I do,the Core Temp could stay at 70C.Take this if there's any helps~

Vilandra
05-12-2009, 11:45 PM
After making some adjustments to the DDR and AMB voltage, I still had the rebooting problem, so Boxx sent me 4 new RAM modules to replace the ones I had, and now everything seems to be working just fine.

Im back up to 4Ghz, and theres been no rebooting while rendering. I guess it was bad RAM.

Vilandra
05-15-2009, 03:50 PM
If anyone is still interested in the outcome, apparently the new RAM didnt fix it afterall. Im waiting to hear back from Boxx.

imashination
05-15-2009, 04:04 PM
Youve made a blood sacrifice to the motherboard gods right? No computer will work correctly until youve grazed your knuckles installing a cpu, or caught your finger clipping in the ram, or at least sliced yourself a little on a sharp metal case edge. Id have thought boxx would have done this before it left the factory.

dmeyer
05-15-2009, 04:05 PM
I would be very surprised if its not a motherboard issue.

aglick
05-15-2009, 04:09 PM
really? well damn.

Please let me know if our support team can't get your issues resolved in good time.

Adam
BOXXlabs

Vilandra
05-15-2009, 05:15 PM
The mobo gods comment was funny.... however there are no razor sharp areas inside the case to cut myself... although I did scarp my fingernail on the RAM release thing.

Boxx got back with me on this and they are sending me two 2gb dimms to replace the four 1bg dimms seeing as how the machine works just fine with only 2 dimms in place.

If this doesnt work, then they said ill need to send it in for repair. They are thinking its either a bad mobo or CPU.

The problem with this, is that I dont have a backup machine. The only thing I have sitting around at the moment is a very old P4 system. (P4 non-HT 1.4Ghz, 768RAM, GeForce MX4000 128MB, IDE drives - no SATA)

olson
05-15-2009, 05:22 PM
The mobo gods comment was funny.... however there are no razor sharp areas inside the case to cut myself... although I did scarp my fingernail on the RAM release thing.

Boxx got back with me on this and they are sending me two 2gb dimms to replace the four 1bg dimms seeing as how the machine works just fine with only 2 dimms in place.

If this doesnt work, then they said ill need to send it in for repair. They are thinking its either a bad mobo or CPU.

The problem with this, is that I dont have a backup machine. The only thing I have sitting around at the moment is a very old P4 system. (P4 non-HT 1.4Ghz, 768RAM, GeForce MX4000 128MB, IDE drives - no SATA)

That's not a solution, that's a bandage. Then when you upgrade memory later to 4 modules or more the shit hits the fan again and probably out of warranty. Get it fixed right and don't even waste time with the new 2GB modules.

Vilandra
05-15-2009, 05:26 PM
That's not a solution, that's a bandage. Then when you upgrade memory later to 4 modules or more the shit hits the fan again and probably out of warranty. Get it fixed right and don't even waste time with the new 2GB modules.

And I completely agree, but I only have one old machine (as mentioned in the previous post), and it cannot read from the 2 SATA drives I have in the Boxx right now. I would have to reinstall all the apps I use on it (which work, but they are slow) and im currently working on a job for a client and cannot just stop half way to change machines.

olson
05-15-2009, 05:39 PM
If you just need to get to the data get a SATA controller card for the other system. Or maybe since Boxx has been sending the memory, maybe they can just send the board and/or CPU if you're comfortable replacing those items. Otherwise you'll just have to deal with what you get on their terms and time, one of the disadvantages of purchasing a system from a builder. Cheers!

Vilandra
05-17-2009, 11:47 PM
Im planning on installed the 2gb DIMMs when they arrive Monday, and ill use them until I can finish the job im working on. Hopefully ill have a small break this week so i can box this machine up and send it back to Boxx so they can fix it. Or at least I hope they can fix it. The machine in general has been awesome, and other than this rebooting issue while rendering, I still enjoy using it.

The last e-mail I got from tech support said they believe it is probably a faulty mainboard or CPU. From what I can tell, the overclock is not stable beyond 3.8Ghz, and I know these skulltrail systems can handle 4+Ghz, so this has to be a hardware issue.

Ill reply back once I get it back from them, or at least know what the problem was. This thread has been going for 4 pages, so I might as well keep it updated in case anyone is interested in the outcome.

Vilandra
05-18-2009, 09:16 PM
Got the 2gb DIMMS today. No matter how I installed them, the machine would not boot, so I called Boxx and talked to tech support. After going through a few procedures, the tech said I could have a dead core in one of the CPU's, or just a faulty mainboard.

So I boxed it up and sent it off for repair. Now ill just wait a few days to hear back from them.

DanielWray
05-18-2009, 09:18 PM
That sucks man, i hope it's not long until you get your machine back.

If it were a dead core it wouldn't show up in the task manager though i'm guessing, sounds like your Mobo is faulted.

Vilandra
05-18-2009, 09:33 PM
Tech support said they will have it no longer than 5 days. They will ship it back "overnight" which is nice. I shouldnt be no longer than 2 weeks without it. (3 days to get there, 5 days to fix assuming the worse, then 1 day to get back... 9 days excluding weekends)

Im currently running on my old P4 machine. Its slow and could probably be displayed in a museum with the dinosaurs, but at least it runs C4D.

Vilandra
05-26-2009, 03:35 PM
I just got a call from Boxx this morning. They said the memory was fine. No errors, so they checked the CPU's and found that both of them had a dead/bad core. So now they got the new CPU's in and testing them (burn in) and he would call back to let me know how that went.

Its a good thing ive gone with Boxx on this one. Not only for tech support, but if I would have built this myself, I would have to fork out $3k for both of those CPU's. :eek:

CGTalk Moderation
05-26-2009, 03:35 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.