Overclock.net - An Overclocking Community - Reply to Topic
Thread: ROG Crosshair VII overclocking thread Reply to Thread
Title:
Message:

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in


  Additional Options
Miscellaneous Options

  Topic Review (Newest First)
07-04-2020 12:51 PM
neikosr0x
Quote: Originally Posted by xeizo View Post
Judging by that score, I should not change to a XT, I get massive better multicore with CCX OC on my regular X. And same single core.

Did a baseline to compare:

https://browser.geekbench.com/v5/cpu...seline=2802539
yea they are just a refresh so makes sense AMD they want to keep prices up a bit to sell the ZEN3 at a good price. But i do expect this round of zen2 cpu to be even more stable and responsive.
07-04-2020 10:34 AM
xeizo
Quote: Originally Posted by nick name View Post
Yeah that isn't a good example for a score. There are better scores I've seen.

There is a 1423 Single Core score on a CH8, but none of the multi-core scores are great. And that makes me believe that the 3900XT is gonna stick with the one good and one bad CCD. Perhaps, though, the bad CCD is going to be an old good CCD from a 3900X and the new good CCD will be the better silicon.
I think you hit the head of the nail there! Of course, why waste TWO perfectly good dies on mere consumers. Yes, I believe that too. XT is for clearing stock I recon, there is both good and bad stock so while CCX0 is actually better than ever before CCX1 may even be worse than what we are used too. Remember, Intel has nothing to compete in multi anyway. Even with a boat anchor for CCX1 it will beat 10900K in multi.

A better CCX0 is a welcome addition though. Discovered yesterday that my 3900X CCX0 ain't that bad, when I tried CCX OC on my 3700X. It is way worse than my 3900X and could hardly beat out it's PBO scores stable. And it took much too high vcore to get anywhere above what PBO performs.
07-04-2020 10:10 AM
nick name
Quote: Originally Posted by xeizo View Post
Judging by that score, I should not change to a XT, I get massive better multicore with CCX OC on my regular X. And same single core.

Did a baseline to compare:

https://browser.geekbench.com/v5/cpu...seline=2802539
Yeah that isn't a good example for a score. There are better scores I've seen.

There is a 1423 Single Core score on a CH8, but none of the multi-core scores are great. And that makes me believe that the 3900XT is gonna stick with the one good and one bad CCD. Perhaps, though, the bad CCD is going to be an old good CCD from a 3900X and the new good CCD will be the better silicon.
07-04-2020 09:49 AM
xeizo
Quote: Originally Posted by nick name View Post
Apparently you don't need the latest BIOS to run the new 3900XT. In this Geekbench 5 run they're using BIOS 3004.

https://browser.geekbench.com/v5/cpu/2802539
https://browser.geekbench.com/v5/cpu/2802539.gb5
Judging by that score, I should not change to a XT, I get massive better multicore with CCX OC on my regular X. And same single core.

Did a baseline to compare:

https://browser.geekbench.com/v5/cpu...seline=2802539
07-04-2020 09:44 AM
nick name Apparently you don't need the latest BIOS to run the new 3900XT. In this Geekbench 5 run they're using BIOS 3004.

https://browser.geekbench.com/v5/cpu/2802539
https://browser.geekbench.com/v5/cpu/2802539.gb5
07-04-2020 02:07 AM
garf333 I set it to 1.39V, same as the DRAM Voltage.

Hitting the power button turns it on, it turns off once, then it will turn on for good.

Most of my timings are manually dialed in except for a few (TRFC2, some other minor timings)

Any way to speed this up or is this the fastest it will go?

Edit: Uploading my RAM timings.

SOC 1.06 I think.
VDDG/VDDP are on auto.
07-03-2020 07:25 PM
tcclaviger Did you leave ram boot voltage at auto? Changing it will trigger double posts every time it starts up, once to train timing at 1.2, shutdown, then boots with trained timings at bios set boot voltage.
07-03-2020 06:53 PM
garf333
Quote: Originally Posted by xeizo View Post
Hynix AFR on 3600? That is not everyday, as they usually clock like ****. Could explain some of the BSODs, just amazing they even boot
Yeah, it's interesting I got these to 3600 at all. First time to "win" the silicon lottery.

I have these Vengeance LPX sticks on 3600mhz at 1.39V @ 16-19-19-20-38-58
TWR - 18
TFAW - 32
TRFC - 510

AIDA giving me 69.x - 70ns runs.

Running a 3700X with just PBO on a minor negative voltage offset.

Several 4 pass memtest86 runs all came back clean but was running into idle BSODs with the old BIOS, but "stable" while the computer is actually in use.

Noticed reboots are faster with this BIOS too, but cold boot/double pumps still take as long as ever.

These sticks would actually do 3733 @ 16-21-21-21 and it was memtest stable but every 2nd cold boot would fail. I don't know which voltages to tweak for that so I went back to 3600...
07-03-2020 06:48 PM
tcclaviger Still fine for me, have run the chip up to 220 watts (SMU in hwinfo) thus far testing out the Optimus Foundation block, nothing weird observed.

104 and 104.2 bclk are achievable, bringing 4770ish single core, but they take a positive voltage offset, not ok imho (puts single core boost max at 1.512-1.515 ish).

Scalar now actually does something paired with 103 or higher bclk. I can use it to fine tune stability and boost speeds from 103-103.8 bclk, so, while it seems like 99% placebo usually there is "some" value to it lol.

It behaves like a very mild "FIT limit" override now.

Regarding bclk, I've never had an issue with it, I don't think I've run stock FSB/Bclk since at least the Duron 600 days even, on either side, Intel or AMD.

On the other hand, it is a real issue because people do end up with corrupted drives at times.

I do tend to not use anything PCIE aside from 1 or 2 GPUs, though the $7 NVME to PCIE adapter seems flawless so far lmao.

I would never advise running any production or mission critical machine like this, at least, not without extensive testing and validation before deployment.

Nor would I advise it for anyone who can't afford to replace something that gets killed easily. I have a plethora of PC parts sitting around, backups to backups, so if something dies.... Meh, NEXT, lol.
07-03-2020 01:57 PM
nick name So how is the LN2 BIOS setting working out for everyone? I can't boot with it enabled.
This thread has more than 10 replies. Click here to review the whole thread.

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off