High temp with low Voltage EVO 212 (i5 4690k)

Toggle sidebar Toggle sidebar
M

Markidjani

Reputable
Oct 5, 2014
20
0
4,510
Hello, community

The last couple of days I have read dozens of posts on this forum but I haven't find anyone with the same problem.

Configuration:

I5 4690k
Cooler Master EVO212
Gigabyte z97x gaming5

I am trying to OC my CPU. I am running at 4.2 ghz and 1.16v but using prime 95 my temps are jumping to 93 degrees! The problem is that lowering from 1.2v to 1.16v didn't lower temp at all. I am running small FFTs. I read about AVX2 might be a problem but wouldn't turning that off give false information about temperature under load. Basically whatever I do temp stays about the same. I've tried 4.4 ghz and.1.22 v and 4.2 ghz with 1.16v and temps are maxing 90degrees in both cases. When i am monitoring temps during Avida64 they are usually around 70 degrees. LinX alsow shows lower MAX temp then prime 95. I also notice that my fan rpm never goes above 1600rpm and max should be 2000rpm for EVO212?
I am OC mostly for the purpose of streaming my video games. Idle temp is below 40 degrees and gaming around 60 degrees.

Question.
1)Should I be focusing on those max temperature peaks when doing OC (yes I've read Temp guide on this forum - 80 degree as an upper limit). The problem is that max temp is going 90+ degrees while one monitored are usually around 70 during stress tests.
2)Amongst prime 95, avida64 stability test and Linx (since they provide different temp results) which are the temperature limits for each of this that shouldn't be exceeded. Link this to 1st question if I should watch Max temp or the current one manually observed through the most part of the stress test.
3)A good step by step guide how to increase fan speed to 2000 rpm when temp above 70 degree

More details about my hardware
https://imgur.com/a/jxR8EJM

Im doing this first time and not so familiar with electronics or programming.

About OCing I am aware of the process in the next way: increasing multiplier till the stability is compromised (stress test doesnt provoke blue screen or restart) with current voltage then increasing voltage that allows higher increase of multiplier but also increases temperature and that whats limits your OCing stable potential. Correct me if i got sth wrong.

Thanks in advance.
 
Solution
geofelt
1) I consider about 85c. to be the point where a stress test should be stopped. That is the default for OCCT.

2) If your test fails, you should either lower your multiplier or raise the voltage.
Monitor with cpu-Z. I would not go past 1.3v.

Your motherboard bios should have a fan control section.
That is where the cpu cooling fan control will be set. Sophistication will vary by motherboard.

AVX generates unusually high temperatures when used.
Keeping it means that you may have to lower your multiplier to accommodate the possibility of high AVX usage.


A couple of observations:

I see poor airflow setup in your case.
Remove the unnecessary drive caddies to let the front intake air get to your parts.
I assume you have a fan...
Sort by date Sort by votes
geofelt

geofelt

Titan
Oct 9, 2006
54,295
4,145
174,240
1. On max temperature.:
Do not worry too much. The cpu will slow down or shut off to protect itself if it senses a dangerous temperature.
That is around 100c.
A 85c. limit while stress testing is appropriate.
That is the default with OCCT.

2. The key metric to monitor is your vcore. It should not go past 1.3v.
Monitor this with cpu-Z.

Realize that how high you can oc a 4690K is primarily determined by your luck in getting a good chip.
Your process to overclocking is fine. Gradually raise the multiplier until you reach limits.

The cm hyper212 is a popular cheap cooler, but not the best.
What is your ambient temperature?
I would expect to see 10-15c. over ambient.
Any cooler needs a good source of fresh air to cool the cpu and the gpu.

If you take the case covers off and direct a house fan at the innards, does this help?
If so, look to case cooling.
 
Upvote 0 Downvote
M

Markidjani

Reputable
Oct 5, 2014
20
0
4,510


How can my Vcore fluctuate over upper limit Ive set in BIOS to 1.19v?

My ambient temp is 28 and idle is ranging 38 to 43 (43 on first core)
I tried OCCT and it shows temperature around 65 degrees which was constant
When i use Prime95 it goes over 90 degrees (small FFTs, max heat,FPU stress, data goes in L2 cache,RAM not tested much) constantly (both MAX and in time observation since it doesnt fluctuate) while LinX is going MAX temp of 85degrees and fluctuates btw 60 and 85.

It confuses me since those different programs provide different info.
I understand that prime95 is hard core temp test since it gives max load all the time while LinX is simulating real usage with fluctuations. On the other hand the OCCT provided stable 65degree which differs from both of those. Currently at 4.3 ghz and 1.19v. Max TEMP with prime 95= around 94 degrees, with LinX 85degree with OCCT 65 degree (tests were about 10 minutes)
Difference btw OCCT test and prime 95 is around 30 degrees!

Whats your opinion of setting up my volts and multiplier and can u explain me those differences.


 
Upvote 0 Downvote
geofelt

geofelt

Titan
Oct 9, 2006
54,295
4,145
174,240
Different stress testing apps use different instructions.
Prime95, for instance uses those which will calculate prime numbers best, but that is not what the usual app uses.
Intel burn test is designed to generate maximum cpu temperatures.
Again, not what the usual app uses.
Some tests use avx instructions which are not commonly used but which put an unusual load on the cpu.

You might find the CPU-Z option of stress testing to be as good as any.

Unless you are seeking your max overclock, just gradually raise the all core multiplier and see how you do.
You can probably leave the voltages to auto on most motherboards.
If you get a failure, back off a notch or two.

Monitor your test with cpu-Z and temperatures with HWmonitor.
Back off the overclock a notch if temps get to 85c. or if vcore approaches 1.3v.
I do not know if there is a stress tester that has a statistically correct mix of instructions for the typical gamer or app user.
 
Upvote 0 Downvote
M

Markidjani

Reputable
Oct 5, 2014
20
0
4,510
I understand but which stress test should i be using.
My pc will be used for playing World of Warcraft and streaming to Twitch/YT using OBS program.

I did 10hrs stability test with AIDA 64 stability test and it looked like this:

https://imgur.com/a/G0hT1SD

After that i did run prime95 and my temp were reaching MAX 88 degrees and then my PC went blue screen.

It was 4.2ghz and 1.13v. My voltage is fixed its not on auto.

Question:

1)What MAX temps are safe when using prime95 and should I focus on AIDA 64 that was stable for 10hrs running, or the Prime 95 that resulted with blue screen after 10 minutes of work

2) Should I lower my multiplier or raise the voltage at this point.

P.S. I dont think disabling AVX is good, as i know OBS is using it. Correct me if wrong.

Thanks for all the answers by now and I would like someone to answer why my fan is maxing 1600rpm when its 100% load should be 2000rpm

You can see my rig here and OCCT 10 min testing results
https://drive.google.com/open?id=1TDCk-YxYqLu7LgledJ1dpaoYC0opxnwc
 
Upvote 0 Downvote
geofelt

geofelt

Titan
Oct 9, 2006
54,295
4,145
174,240
1) I consider about 85c. to be the point where a stress test should be stopped. That is the default for OCCT.

2) If your test fails, you should either lower your multiplier or raise the voltage.
Monitor with cpu-Z. I would not go past 1.3v.

Your motherboard bios should have a fan control section.
That is where the cpu cooling fan control will be set. Sophistication will vary by motherboard.

AVX generates unusually high temperatures when used.
Keeping it means that you may have to lower your multiplier to accommodate the possibility of high AVX usage.


A couple of observations:

I see poor airflow setup in your case.
Remove the unnecessary drive caddies to let the front intake air get to your parts.
I assume you have a fan mounted in front as intake.
Install a exhaust fan at the rear.

Chieftec is considered to be a suspect quality psu.
The stability of voltages supplied may not be good.

 
Upvote 0 Downvote
Solution
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom