Overclock.net › Forums › AMD › AMD Motherboards › ROG Crosshair VI overclocking thread
New Posts  All Forums:Forum Nav:

ROG Crosshair VI overclocking thread - Page 950

post #9491 of 17390
Quote:
Originally Posted by jsp001 View Post

Been doing a lot of reading here, and playing with overclocking for a couple of weeks off and on. I'm on 082 with 1800X and F4-3200C14Q-64GVR...

I finally have the ram at 3200, but cant get the timings down (18-16-16-36) is the best that will run. Maybe I should abandon 1T and try 2T?

CAS18 is the lowest you can go with anything higher density than 2x8 B-die. If you want lower CAS at 3200, you have to use 2666 or lower divider and bclk overclock. Maybe this will be fixed in the May update.
Quote:
Originally Posted by hotstocks View Post

It's important because every computer I have built and every laptop I have owned in the last 15 years passes Intel Burn Test on maximum.
If your rig doesn't, then it is not 100% stable. Sure I can run my Ryzen 1800X at 4-4.1 ghz passing all other tests or just using the
computer as you say, but it blackscreens/off within minutes of IBT maximum. I have found that to pass I need to drop to 3.9ghz, so to me that
is true stability.

Have you tried increasing the overcurrent threshold in bios?
Edited by infraredbg - 4/15/17 at 2:37pm
post #9492 of 17390
Quote:
Originally Posted by Spartoi View Post

I'm missing processor states and it seems this is the issue. Is there a way to overclock and still allow downclocking at idle? I thought using offset mode would fix it but the problem seems to come from using custom core ratio.

You've to use pstate overclocking and not Multiplier (also called "CPU Core Ratio"). Go to "Advanced/AMD CBS/Zen Common options/Custom pstates" in BIOS.

Set the FID of p0 to a hex number suitable for you, 98 is 3.8GHz for example. You can see the frequency on the right side. Change the applied voltage under "Extreme Tweaker/CPU vcore/offset" to something suitable for your chip (default for non-X is 1.18V (add +0.15V for 1.33V), for X it is 1.35V (add 0V for 1.35V)).
Edited by TwoBeers - 4/15/17 at 2:38pm
post #9493 of 17390
Quote:
Originally Posted by MNMadman View Post

Most likely no, and again most likely no.

The 3200C16 kits aren't the best at running at ratings on Ryzen. Highest chances of running at ratings are had with 3200C14 kits, specifically F4-3200C14D-16xxx kits (2x8GB).

There are a few that can run high speeds with four sticks. A very lucky few.

Hopefully all of this will change after the May update.

Thanks!
post #9494 of 17390
Copy that, Thanks!
post #9495 of 17390
Quote:
Originally Posted by Gadfly View Post

I did the same until I got the AMD edition mounting screws to use the factory backplate.

Did you got them from EK or from somewhere else?
post #9496 of 17390
Hi @elmor, during the oc of Pstate have had this error in sequence : ad 30 oc ​​02 69.
Safe Boot does not work, even the clr CMOS struggled. At reboot the bios result blurred, now work..it's normal??

Buona Pasqua / Happy Easter !!!

1700x
16 tridentz 3600c16
Bios 0081
post #9497 of 17390
@waltercaorle: Did you change FID/VID of p1 or p2? If you go too low it will result in a not working boot with probably reflash of BIOS necessary.
post #9498 of 17390
Quote:
Originally Posted by infraredbg View Post

CAS18 is the lowest you can go with anything higher density than 2x8 B-die. If you want lower CAS at 3200, you have to use 2666 or lower divider and bclk overclock. Maybe this will be fixed in the May update.
Have you tried increasing the overcurrent threshold in bios?

Nope. Where is it in the bios? And what is your recommended and safe value?
post #9499 of 17390
Quote:
Originally Posted by hotstocks View Post

It's important because every computer I have built and every laptop I have owned in the last 15 years passes Intel Burn Test on maximum.
If your rig doesn't, then it is not 100% stable. Sure I can run my Ryzen 1800X at 4-4.1 ghz passing all other tests or just using the
computer as you say, but it blackscreens/off within minutes of IBT maximum. I have found that to pass I need to drop to 3.9ghz, so to me that
is true stability. All other stress tests are a joke compared to IBT max. The heat and stress on the processor are far greater than any other
test. Sure, maybe I can get 4ghz and IBT passing at like 1.5v, but I would rather run 3.9ghz at 1.39v and pass. I may be losing 100 mhz for
everyday stuff, but the savings in not burning up your chip and hearing screaming case fans is worth it. I have tried a 1700 and 1800X and
basically all these chips hit a wall at 3.85-3.95 ghz depending on the chip. Approaching that wall with a normal 1.39v is great, going beyond
that wall even a tiny bit takes a ridiculous amount of added voltage and heat. It is just not worth it. From what I can tell so far, the only reason
to get a 1800X over a 1700 is they will both do 3.9ghz, but the 1800X will only need like 1.39v whereas the 1700 will need 1.425v AND LLC 3
instead of auto or 1. You basically are paying up to get a cooler running chip that needs less voltage to be 100% stable at 3.9 ghz or thereabouts.
I'm coming from a i5-2500k that was running 6 years 100% stable at 4.7ghz passing IBT, sure I could have run it at 4.8ghz and fail IBT, but why?
Don't get me wrong, I am NOT one of those guys who says you need to run Prime for 24 hrs and every other test to be 100% stable. IBT at maximum
takes about 2 hours to do its 10 runs, and if it passes those 2 hours, you are good to go, it will pass anything. So my point is Silicon Lottery's chips
will not pass IBT maximum at the speeds they are advertising. When you reach a Ryzens wall (and they are all about the same) and put the hurt on
them, you will see they all crap out similiarly and are not worth paying an extra $200 to Silicon Lottery to get a chip that is probably no better than one
from Amazon or Newegg. Now if Silicon Lottery changed their stress testing and included IBT maximum and LLC 1, well that would be a different story.
My point is their chips are not as stable as they are claiming and not really much different than any others.


Well... Candidly.. no... To all of it.

Your are overclocking to find stabilty to a specific tool... Which is fine, that is what you find to be fun, but it is a synthetic stress test that is not applicable to real use, nor is it a gauge of stability. It is a combination of operations speficly designed to generate heat that is only found in that tool.

Further all ryzens hit a wall, In fact they hit 3 as you scale up from 3ghz to 4.3ghz. The first at 3.3, the second at 3.8/9 abd the third at 4.2/3.

That said they are anything but equal. I have tested 6 cpu's personally, I own two 1800x, and there are huge diffrences in quality out there; in terms of clock speed, voltage, and memory controllers. More so than I have seen in any other cpu in easily 15 years. One of my 1800x will cold boot ddr4 3600, the other will only boot 3200; another I tested would only boot 2933, all with the exact same memory.


As for silicon lottery.... So far, every cpu I have purchased from them has done exactly what they say it will do. My 4.1 I got from them runs realbench @ 4.1 for 2 hours at 1.4v.

It will run realbench for at least an hour at 4.175 at 1.44v...

My other 1800x will run 4.0 at 1.4v for two hours and 4.1 @ 1.44v.

(all at llc 2).

Given that only about 25% will run 4.1, I think the extra $149 was well worth it. I got a great chip with a great imc.

When I get my machine back up in it's case (has been on a test bench) I will try the intel burner And let you know how it goes.
post #9500 of 17390
Quote:
Originally Posted by TwoBeers View Post

@waltercaorle: Did you change FID/VID of p1 or p2? If you go too low it will result in a not working boot with probably reflash of BIOS necessary.

P0 only. I was working with the blk frequency * FID. VID does not make me change and so I changed the offset vcore
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD Motherboards
Overclock.net › Forums › AMD › AMD Motherboards › ROG Crosshair VI overclocking thread