Overclocking non-K cpu.

[NicholasNicholasNicholas]

2 minutes ago, crzyces said:

Does it show that it's running at over 28?

yes, its now 30-34. sadly i cannot change boost

[SpaceGhostC2C]

6 minutes ago, Nickathom said:

I did flash the bios earlier with my portable harddrive and theres still no blck. but i did up the monimum clock up from 28, so its now 30-34. should i increase this more or leave it the same?

First things first: what did you get with that? Is it runnung at 3.0?

 

(and mind you, i have no experience with said board nor cpu. I'm purely speaking from logic - so probably not more helpful than you can be by just being methodical). 

[NicholasNicholasNicholas]

Just now, SpaceGhostC2C said:

First things first: what did you get with that? Is it runnung at 3.0?

 

(and mind you, i have no experience with said board nor cpu. I'm purely speaking from logic - so probably not more helpful than you can be by just being methodical). 

 

yes, it is now at 3.0 to 3.4

[SpaceGhostC2C]

Just now, Nickathom said:

yes, it is now at 3.0 to 3.4

Then the next step is to test that it is stable and doesn't overheat under load (while checking that it never goes below 3.0 during 4-core loads).

[crzyces]

1 minute ago, Nickathom said:

yes, its now 30-34. sadly i cannot change boost

Download http://download.cnet.com/Prime95-64-bit/3000-2053_4-75869492.html

 

install it. Start running a low stress test and monitor the temps. take notes, screenshot, or try to recall what you see to make sure it is running at the correct speeds and temps on each core.

[crzyces]

2 minutes ago, SpaceGhostC2C said:

Then the next step is to test that it is stable and doesn't overheat under load (while checking that it never goes below 3.0 during 4-core loads).

If he gets this to 3.2, even though it is a small jump, I'll be pretty darn happy as I was willing to give up and pushing him to start on easier chips/board combos to oc, as both the chip and board are made to not oc.

Edited April 10, 2017 by crzyces
If he gets it to 3.0 I'll be happy to be honest. I want him to confirm it in the software tests though. I'm at 4.2 but dxdiag and such shows 3.4, whereas cpu-z shows 1.6 until I start stressing. Normal properties always shows 3.4 though.

[NicholasNicholasNicholas]

9 minutes ago, crzyces said:

If he gets this to 3.2, even though it is a small jump, I'll be pretty darn happy as I was willing to give up and pushing him to start on easier chips/board combos to oc, as both the chip and board are made to not oc.

 

 

 

i ran the torture test on prime95 and was getting around 3190mhz with 60c average. i think i could push it up to maybe 3.3? also the cpu was on 100% always

[SpaceGhostC2C]

1 minute ago, Nickathom said:

i ran the torture test on prime95 and was getting around 3190mhz with 60c average. i think i could push it up to maybe 3.3?

You might, but you are already getting 3.2 on all cores under load, so it's not like upping the minimum multiplier had any effect so far. Except that you don't let it fall as much at idle  

If your CPU is happy to run Prime95 at 3.2, then I guess the only way to force it to go even faster would be setting the minimum above 3.2. But I would check again the clocks you actually get when doing whatever it is that you do with your PC. It may be the case that you are already running above 3.4 in some cases (both my office 4570 and my laptop 3720 do this -going over turbo clocks- on their own).

[NicholasNicholasNicholas]

Just now, SpaceGhostC2C said:

You might, but you are already getting 3.2 on all cores under load, so it's not like upping the minimum multiplier had any effect so far. Except that you don't let it fall as much at idle  

If your CPU is happy to run Prime95 at 3.2, then I guess the only way to force it to go even faster would be setting the minimum above 3.2. But I would check again the clocks you actually get when doing whatever it is that you do with your PC. It may be the case that you are already running above 3.4 in some cases (both my office 4570 and my laptop 3720 do this -going over turbo clocks- on their own).

 

oooh. now i get it. im going to try 3.4 on all around

[SpaceGhostC2C]

Just now, Nickathom said:

oooh. now i get it. im going to try 3.4 on all around

That's fine, but I do think you need a pre-OC baseline to know how your PC typically behaves before you change anything. Otherwise how can you know if you succeeded in making it go any faster in practice?

[crzyces]

1 minute ago, SpaceGhostC2C said:

That's fine, but I do think you need a pre-OC baseline to know how your PC typically behaves before you change anything. Otherwise how can you know if you succeeded in making it go any faster in practice?

He doesn't, as his max turbo is 3.4 which I just noticed.

[NicholasNicholasNicholas]

5 minutes ago, crzyces said:

He doesn't, as his max turbo is 3.4 which I just noticed.

is there a way to tell the cpu to ignore the temps? it should be running at 3.4 (i think) when its under load since its only at 60c. Or is it voltage thats the problem?

[crzyces]

5 minutes ago, Nickathom said:

is there a way to tell the cpu to ignore the temps? it should be running at 3.4 (i think) when its under load since its only at 60c. Or is it voltage thats the problem?

How long is the torture test running? I have my voltage set at 1.35.

[NicholasNicholasNicholas]

Just now, crzyces said:

How long is the torture test running? I have my voltage set at 1.35.

its been running since i started it and its still at 60c. i think my voltage is 1.0. should i try invreasing it to 1.2?

[NicholasNicholasNicholas]

30 minutes ago, crzyces said:

How long is the torture test running? I have my voltage set at 1.35.

Ohhhh I figured out the problem. It wont let me set the 4th core above 32, and the 3rd core above 33, but core 1 and 2 are now running at 3.4 and appear to be stable. Id like to thank you a bunch for your help. I never expected this kind of aid. Thanks a lot.

[crzyces]

Yes. Try 1.2, then 1.25 up to 1.35 (or whenever you reach the speed you want, as once you hit that speed, don't raise the voltage any more. Do NOT go past 1.5!

[crzyces]

Just now, Nickathom said:

Ohhhh I figured out the problem. It wont let me set the 4th core above 32, and the 3rd core above 33, but core 1 and 2 are now running at 3.4 and appear to be stable. Id like to thank you a bunch for your help. I never expected this kind of aid. Thanks a lot.

That's odd (the fourth core), but I'm glad we got you where you wanted to be. I am darn surprised too! I think we deserve some thumbs up on the bottom right :^) .

[NicholasNicholasNicholas]

Just now, crzyces said:

That's odd (the fourth core), but I'm glad we got you where you wanted to be. I am darn surprised too! I think we deserve some thumbs up on the bottom right :^) .

 

Yeah for some reason i had to make it so i control all of the cores individually, otherwise it wouldnt go over 32

[SpaceGhostC2C]

1 hour ago, Nickathom said:

is there a way to tell the cpu to ignore the temps? it should be running at 3.4 (i think) when its under load since its only at 60c. Or is it voltage thats the problem?

 

47 minutes ago, Nickathom said:

Yeah for some reason i had to make it so i control all of the cores individually, otherwise it wouldnt go over 32

 

The maximum turbo clock applies, in principle, to single-core loads. I have seen, however, CPUs go above the turbo clock, sometimes even on more than one core simultaneously, so I really don't know how Intel manages that and how it varies from model to model. Being a locked CPu in a supposedly locked motherboard, however, I wouldn't be surprised if you are constrained by the rated TDP, which could mean that 3.4 - 3.4 - 3.3 - 3.2 is the maximum you can get without drawing more power than intended for this CPU (when overclocking unlocked chips you typically drive them way beyond their TDP, hecne why motherboards with overkill power delivery are needed).

 

I'm glad you got to push it to its apparent limit. Now the only question is whether your computer wasn't doing this on its own under load  

[crzyces]

1 minute ago, SpaceGhostC2C said:

 

 

The maximum turbo clock applies, in principle, to single-core loads. I have seen, however, CPUs go above the turbo clock, sometimes even on more than one core simultaneously, so I really don't know how Intel manages that and how it varies from model to model. Being a locked CPu in a supposedly locked motherboard, however, I wouldn't be surprised if you are constrained by the rated TDP, which could mean that 3.4 - 3.4 - 3.3 - 3.2 is the maximum you can get without drawing more power than intended for this CPU (when overclocking unlocked chips you typically drive them way beyond their TDP, hecne why motherboards with overkill power delivery are needed).

 

I'm glad you got to push it to its apparent limit. Now the only question is whether your computer wasn't doing this on its own under load  

I really wish I had access to that board and cpu so I could really dig in, especially safety wise as he does not want 4 cores running over 3 when it's idle. I'd also like to see if the turbo clock is locking it or if it can go higher (it should go above the boost for a true oc). I don't want him to fiddle too much though, as with no one there, it is pretty easy to say "Eureka", then smell smoke. Hopefully he reads this and leaves core temp on to see what his base clocks and temps drop to.

[SpaceGhostC2C]

Just now, crzyces said:

I really wish I had access to that board and cpu so I could really dig in, especially safety wise as he does not want 4 cores running over 3 when it's idle. I'd also like to see if the turbo clock is locking it or if it can go higher (it should go above the boost for a true oc). I don't want him to fiddle too much though, as with no one there, it is pretty easy to say "Eureka", then smell smoke. Hopefully he reads this and leaves core temp on to see what his base clocks and temps drop to.

Yes, I mean, when fully idle the power states kick in (if he didn't disable them), and the CPU will go to ~800 MHz or something like that. But any tiny load will kick them up all the way to 3.2-3.4 instead of 2.8. If he didn't change any voltage (could he?), it shouldn't matter too much, though.