Overclock.net › Forums › Intel › Intel CPUs › The Intel Devil's Canyon Owners Club
New Posts  All Forums:Forum Nav:

The Intel Devil's Canyon Owners Club - Page 1984

post #19831 of 22165
Quote:
Originally Posted by madboy128 View Post

VIO in HWmonitor and AIDA64 is with name VTT ?

On my board, an Asus Maximus VII Formula (So I guess it's the same for other Maximus VII boards), yes, Digital IO voltage is also labeled VTT.
post #19832 of 22165
Quote:
Originally Posted by new boy View Post

Probably needs more vcore.

My chip needs 1.26v for 4.8GHz, but aprox 1.33v for 4.9GHz.

The voltage requirements start getting pretty step towards the end.


Mine is 1.25 for 4.7 but needs 1.35 for 4.8. I guess 4.7 is the sweet spot for mine but I always need the max i can get and run it at 4.8 1.35 ..
post #19833 of 22165
With + 0.300 vccsa and VDio VDia i run IBT prioriiti high with 1.33vcore 4.9ghz

With +0.200 IBT give me error after 20 seconds, i understand why?!? ...

Asus z97pro wifi
4790k
Vcore 1.33(bios)
InputV 2.000v
Cache 4.0 with 1.15v
Dram 1.55v default 1600mhz

With vccsa And I/o i/a on auto i run prime95 1344-1344 and no crash
But craches in x264 in 7 loop and IBT... why?
post #19834 of 22165
Quote:
Originally Posted by madboy128 View Post

With + 0.300 vccsa and VDio VDia i run IBT prioriiti high with 1.33vcore 4.9ghz

With +0.200 IBT give me error after 20 seconds, i understand why?!? ...

Asus z97pro wifi
4790k
Vcore 1.33(bios)
InputV 2.000v
Cache 4.0 with 1.15v
Dram 1.55v default 1600mhz

With vccsa And I/o i/a on auto i run prime95 1344-1344 and no crash
But craches in x264 in 7 loop and IBT... why?

Dude, I'm going to say it once more. SA and IO voltages are relevant for MEMORY overclocking. Memory, not core. They help only if your core gets unstable because of your memory and shouldn't in any other case do much. And if they do, ther's probably something else wrong. I already told you to try setting your DRAM voltage to 1.65V. Set your input voltage a bit lower and see what happens. Set your cache voltage to 1.2V and see what happens. 1.3V VIO is too much and 1.1V VCCSA is pushing it. You wanna ruin your chip to run IBT? Be my guest.
Edited by tolis626 - 2/13/16 at 6:19am
post #19835 of 22165
Thank you man, tonight i test with 1.2vcache and 1.8, 1.9v input.
Im stable in prime95 but i cant find stable in x264 test.
What is your opinion, stop using IBT and use prime 28.7 1344fft and x264 bench?
post #19836 of 22165
Quote:
Originally Posted by madboy128 View Post

Thank you man, tonight i test with 1.2vcache and 1.8, 1.9v input.
Im stable in prime95 but i cant find stable in x264 test.
What is your opinion, stop using IBT and use prime 28.7 1344fft and x264 bench?

To be honest, if going over 1.3V I don't really use Prime95. I've also never used IBT. For my stability testing I will usually do a quick (like 5 min) run of Prime95 1344-1344 (Or maybe not even that) and then go to OCCT or RealBench or x264/x265. OCCT is almost as good at finding errors as Prime95 but isn't a resource hog and doesn't actually try to fry your CPU. Maybe boil it a bit, but certainly not fry. x264 is the best balance between finding instabilities and keeping the load and temps realistic, but it takes quite a while to find instabilities sometimes. Then there is my preferred stress test, RealBench. It's perhaps a worst case realistic load that tasks the CPU on multiple fronts and also uses the GPU and a huge chunk of memory. Bad thing is that it also takes a while to find instabilities sometimes and the PC is practically unusable while running RB, while with OCCT and x264 you can quite easily browse the web or something.

Also, you seem to be ignoring my suggestion to run your RAM at 1.65V. It's not going to mess it up, 1.65V is within spec. I'd try that first if I were you.
post #19837 of 22165
Why should he run 1600 MHz RAM at 1.65v when he is not really overclocking RAM? Same logic you applied for VCCSA..
post #19838 of 22165
Quote:
Originally Posted by GeneO View Post

Why should he run 1600 MHz RAM at 1.65v when he is not really overclocking RAM? Same logic you applied for VCCSA..

Because raising VCCSA and VIO to very high values helps stability, which points to memory being an issue. Maybe upping VDIMM to 1.65V, that's well within the safe zone unlike 1.1+V VCCSA and 1.3V VIO, will help without long term risk. I would even suggest giving the chip a little more VCORE if it helps keep the other voltages in check. Maybe his IMC is particularly bad, maybe his RAM sticks are utter crap, maybe 4 DIMMs is pushing his chip. There's many maybe's to check before concluding that he needs extreme IO voltages. smile.gif

PS : @madboy128 What memory timings are you using? Also, what exact model DIMMs do you have? Have your tried stock VCCSA and VIO? Sometimes too much can also cause trouble. Also, try something like +0.15V VCCSA and +0.075V VIOD/A. On my system it helps if I give it a bit more VCCSA than VIO. Try stuff before going all out.
Edited by tolis626 - 2/13/16 at 10:11am
post #19839 of 22165
Quote:
Originally Posted by tolis626 View Post

Because raising VCCSA and VIO to very high values helps stability, which points to memory being an issue. Maybe upping VDIMM to 1.65V, that's well within the safe zone unlike 1.1+V VCCSA and 1.3V VIO, will help without long term risk. I would even suggest giving the chip a little more VCORE if it helps keep the other voltages in check. Maybe his IMC is particularly bad, maybe his RAM sticks are utter crap, maybe 4 DIMMs is pushing his chip. There's many maybe's to check before concluding that he needs extreme IO voltages. smile.gif

I think it is not always that simple, since all of this stuff runs of an internal voltage rail. I had a 4770k that I couldn't overclock at all without bumping up VCCSA with 1866 XMP RAM. My 4790K doesn't need any with the same RAM and MB @ 4.7 and 2200 MHz @ 1.6v. I think overclocking the CPU and changing CPU voltages can affect the IMC even if you do not have RAM overclocked (and yes I know anything over 1333 is technically an overclock).

I agree I would look elsewhere first, but I try to keep my RAM under 1.6v myself for 24x7.

.
Edited by GeneO - 2/13/16 at 10:18am
post #19840 of 22165
Quote:
Originally Posted by GeneO View Post

I think it is not always that simple, since all of this stuff runs of an internal voltage rail. I had a 4770k that I couldn't overclock at all without bumping up VCCSA with 1866 XMP RAM. My 4790K doesn't need any with the same RAM and MB @ 4.7 and 2200 MHz @ 1.6v. I think overclocking the CPU and changing CPU voltages can affect the IMC even if you do not have RAM overclocked (and yes I know anything over 1333 is technically an overclock).

I agree I would look elsewhere first, but I try to keep my RAM under 1.6v myself for 24x7.

.

I get what you're saying. And what you described is similar to my case where 2400MHz RAM wouldn't even boot at 1.75V VDIMM without touching VCCSA and VIO. But all I needed was moderate increases in secondary voltages to be able to run it stably at 2400MHz 1.6V. VCCSA needs to be in the 0.9-0.95V range for total stability and VIO doesn't need to go over 1.1V. Adding +0.3V to both seems too excessive. So I've reached the same conclusion as you, the IMC is having trouble keeping up with that overclock. Since there are alternatives, he should look into them.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › The Intel Devil's Canyon Owners Club