Crash when rendering with Cycles.

I know similar thread already exists and for the most part the solution is to oversee the cooling system ect…
Here’s the strange part.

I’ve just started to use cycles because until recently I’ve been using a computer from the early 80’s.
I open my startup scene (that I got from an experienced blender user/friend) which contains a few “lightning-walls” with a basic setup for rendering decent images.
Sampling is set to “Render: 24 & Preview: 12”
Now if I turn on “render view” and it’s starts to sample, the CPU temp goes from 45C to 97-100C in a couple of seconds
and usually freezes the computer.
Also, the render time is incredible slow… about 2min to render all the “path tracing sample” to 144, and that’s with just a few planes in the entire scene.
I was running one of the official “getting started” tutorials from blender and the same thing happened there, except the rendering time was a bit faster (I assume because the default settings were lower), but eventually it froze/crashed.

I have a i7 4GHz Cpu, 4-cores, multi-threading with water cooling.
Radeon R9 290x tri-x oc, 4GB
32GB ram, 2400mhz
The case has 6 additional fans, placed for optimal airflow.
In other words, bad cooling or low specs are NOT the issue.

Any suggestions?
Thanks!

Now if I turn on “render view” and it’s starts to sample, the CPU temp goes from 45C to 97-100C in a couple of seconds

In other words, bad cooling or low specs are NOT the issue.
If your cpu temperature rises this amount in 2 seconds then what makes you think bad cooling is not the issue. You may have a good cooling system but is it working correctly.

If your cpu is working 100% (by any application) then your cooling system should be capable of controlling its temperature.

What happens with just the normal default scene with a lamp and basic material

Thanks for the quick response!

Just tried with the blender default startup scene, (box & light) no material and with Cycles as renderer.
It goes up to 100C in seconds…

The lights on the water cooling sys. is lit, so I assume it’s working.
Gonna open it up and make sure it has a good connection against the CPU this weekend, but I’ve been running the latest games and benchmarking on max settings for hours without any problem.
I realize this can’t really compare to a renderer that uses the cpu in real-time to its fullest, but still I think I’ve would have noticed something.

I was wondering if there might be some specific settings for different cpu’s required to optimize rendering,
but hopefully, as you suggested, it’s a bad connection with the cooler.

Hardware problem, you must check/fix CPU cooling. If you not believe, run CPUburn test for a while.

I removed the cooler and re-attached it again and now I’m getting a temp around 72-76C on each core.
It’s better than 100C, but with a water cooled system you’d expect it to be much cooler than that…

Anyone with a similar system, what is your CPU temp when rendering in cycles? (render view)
I’m not sure what is considered normal when using 100%.
I have:
“Intel Core i7-4790K” with a “Corsair H100i Hydro Series CPU cooler”

Yes no doubt. Could be a CPU thermal paste issue if everything else looks good…

Is the CPU overclocked? Core voltage not too high?

My i5 @4.5GHz 1.32v runs 75-80c during long-term 100% continuous load… air cooled in a 25c room.

Welcome to BlenderArtists.org

-LP

Well, in that case there’s definitely something wrong with my cooler.
No overclocking and from what I understand the cooling paste I used is among the best.
Gonna do a thorough checkup this weekend.

Thanks for your help!
All of you.

No overclocking and from what I understand the cooling paste I used is among the best.
As with the cooling system, something may be the best but if you apply it incorrectly it will be no more than useless

Oh the moderated posts thing got me again. I couldn’t see your last when I posted. My “no doubt” was yes you definitely had a cooling problem at 100c.

Your new temps do seem a bit high, like maybe 10 degrees but that’s just a guess based on what I remember seeing in discussions at overclock.net and other sites. Idle temps can be a good indicator of paste problem too. I seem to remember reading about issues with loose backplate on that cooler model too. You might want to do some research on that chip + cooler before digging into it again.

-LP

Alright, I read up on my cooler and apparently it was quite common with this model that the back plate was slightly loose (a bit tricky to get it fully secured).
So I managed to fix that.
Although even when the backplate was loose, once the cooler was fastened over the cpu it was tight and un-movable, so I don’t really see how it helped.

But anyway, now I have a idle temp on 30C and “hardcore rendering 100% on each core” temp on 65-70C.
Which seems reasonable to me.

Thank you all for your help and suggestions!