INTEL CPUs *OC Benchmark & Discussion*

matique

Senior Member
Joined
Sep 25, 2017
Messages
797
Reaction score
786
Does this apply to AMD as well? Don't think I've seen any 2X24 with EXPO.

So far there's only kits validated for intel. Maybe agesa not ready for AMD side, idk.

for now imo the 24x2 markup not worth losing the capacity and 5-6% performance loss of minimal fps ?



amd i using cbeap kingston a-die 2x32, with low speed and ok tightness.

amd mem controller may not run as fast as intel but it can do it tight 1:1, 1T and gdm off with low voltages. so no point buyed fast 2x24. :o

https://forums.hardwarezone.com.sg/...ead-31-faster-than-adl.6753659/post-147645433

It's more for speed mainly. I don't think any gamer needs 64GB of ram lol. 48 is a decent enough bump from 32 if you need slightly more ram, and it will clock as well as 16x2. This is for intel anyway, it can be tuned to have much higher bandwidth + lower latency than AMD. But yeah i agree for now, the markup is not worth it.

2*24GB kits a bit in no man's land IMO. if 32GB isn't enough, 48GB isn't exactly that far off from 32GB and you still might have issues with not enough RAM.

The 2*48GB DIMMs are interesting though.
2*48 seems like hell to run, especially for the imc.
 

Mach3.2

Great Supremacy Member
Joined
Apr 8, 2011
Messages
72,511
Reaction score
2,488
2*48 seems like hell to run, especially for the imc.
I saw a few people reporting success with 2*48GB kits running XMP on AMD platforms, so probably isn't the absolute worse case scenario of running 4 DIMMs bah.

But since the context here is Intel, I can see why dual rank DIMMs are an issue since almost no one overclocking is going to be content with their RAM at 6000MHz/6200MHz like on AMD platforms.
 

Encrypted11

Supremacy Member
Joined
Jun 9, 2012
Messages
5,937
Reaction score
271
Looks like I can get 8000 stabilized when this BIOS finalizes.

8000 C34 at reasonable voltages with some minor timing changes, air cooled A-dies (BP LN2 RAM sink)
1.41V TX/ VDDQ
1.44V IMC/VDD2
1.58V VDIMM

Long time no see @hlots123 sifu! :o

Edit: 8000 C34 unlocked



hCyBu4x.png

HtRwSbc.png
 
Last edited:

matique

Senior Member
Joined
Sep 25, 2017
Messages
797
Reaction score
786
Looks like I can get 8000 stabilized when this BIOS finalizes.

8000 C34 at reasonable voltages with some minor timing changes, air cooled A-dies (BP LN2 RAM sink)
1.41V TX/ VDDQ
1.44V IMC/VDD2
1.58V VDIMM

Long time no see @hlots123 sifu! :o

Edit: 8000 C34 unlocked



hCyBu4x.png

HtRwSbc.png
great job esp on the unify! what stability test is that on linux?
 

Encrypted11

Supremacy Member
Joined
Jun 9, 2012
Messages
5,937
Reaction score
271
1 Hr of Google Stressapptest on Linux Mint USB stick instead of the Linux shell version inside of Windows.

I usually don’t do stress testing on OS SSD unless it is at least a migration of a known good profile from an older BIOS to protect the data.
 

matique

Senior Member
Joined
Sep 25, 2017
Messages
797
Reaction score
786
1 Hr of Google Stressapptest on Linux Mint USB stick instead of the Linux shell version inside of Windows.

I usually don’t do stress testing on OS SSD unless it is at least a migration of a known good profile from an older BIOS to protect the data.
yep that's good practice, i also do fresh testing on a test os on a separate drive to not corrupt main drive if unstable.
 

Encrypted11

Supremacy Member
Joined
Jun 9, 2012
Messages
5,937
Reaction score
271
So far I started keying the slew control section manually to stabilize 8000MHz with numbers derived from the autos trained by Z790 Apex (copied from OCN users that used various UEFI tools to read numbers trained on Apex). Glad someone's already done the homework! :s13:

The MSI BIOS doesn't expose 2 of the slew control parameters but copying the rest that are available seem to be good enough and currently info on manual ODT tuning on DDR5 is still pretty limited.

Some of 7empe's posts on OCN are pretty helpful.
 

matique

Senior Member
Joined
Sep 25, 2017
Messages
797
Reaction score
786
So far I started keying the slew control section manually to stabilize 8000MHz with numbers derived from the autos trained by Z790 Apex (copied from OCN users that used various UEFI tools to read numbers trained on Apex). Glad someone's already done the homework! :s13:

The MSI BIOS doesn't expose 2 of the slew control parameters but copying the rest that are available seem to be good enough and currently info on manual ODT tuning on DDR5 is still pretty limited.

Some of 7empe's posts on OCN are pretty helpful.

yeah some slew controls can be quite beneficial. As it is i think your timings are pretty great for what it is, if it's stable don't need to play around tbh, unless you can hit 8200 hehehe then let me know. I only got 8200c36 + looser timings stable but since i prefer to run my kits tight i stuck with 8000c32.
 

Encrypted11

Supremacy Member
Joined
Jun 9, 2012
Messages
5,937
Reaction score
271
tested, no diff haha. 8000c32/8200c36 max.
Stabilizing the last 100-200MHz (Well currently 8GHz) was a combination of the following since I'm limited by air cooled DIMMs:
I think I'm done tuning with this platform. :)

This platform had plenty of time to mature unlike the likes of 11th gen or the Skylake variants that came with Ryzen 1000-2000 especially when it seemed like every other BIOS is buggy when MB vendors had to split bios dev resources between AMD and Intel early on.
Z690 going from 6000MHz+ to 8000MHz+ today is a really good statement on how the platform aged.

So far every BIOS is better than the last one which is really refreshing just like the pre-skylake times. :)
 

matique

Senior Member
Joined
Sep 25, 2017
Messages
797
Reaction score
786
Stabilizing the last 100-200MHz (Well currently 8GHz) was a combination of the following since I'm limited by air cooled DIMMs:
I think I'm done tuning with this platform. :)

This platform had plenty of time to mature unlike the likes of 11th gen or the Skylake variants that came with Ryzen 1000-2000 especially when it seemed like every other BIOS is buggy when MB vendors had to split bios dev resources between AMD and Intel early on.
Z690 going from 6000MHz+ to 8000MHz+ today is a really good statement on how the platform aged.

So far every BIOS is better than the last one which is really refreshing just like the pre-skylake times. :)

Nice data points! And impressive to get the UITX up to 8000 as well.

I am sticking with my 8000c32. In most games I am limited to my screen refresh anyway :) But the ram OC does help with keeping frametimes under control as well as improving min fps in games. So tempted to get the 24x2 full mdie kits, but then again 14th gen is looming, and there's going to be CKD in the newer ram kits that will improve signal integrity, so who knows 10000 daily might be a thing :)
 

Encrypted11

Supremacy Member
Joined
Jun 9, 2012
Messages
5,937
Reaction score
271
8GHz holding up well so far, even for cold boots. Pretty good deal going up for Unify ITX lately.

Triple M2, 8GHz RAM, maxed out rear IO, TB4.0 (true 40Gbps than pseudo), aluminium passive cooling backplate for US$309! :)

Has hlots sifu upgraded the VIII Hero with the 5GHz Skylake V1? :s34: Pretty amazing how a pretty heavy OCed (above the typical hwzer standard) PC is still healthy after 7 years.
 
Last edited:

SEvsNK

Arch-Supremacy Member
Joined
Mar 31, 2004
Messages
22,599
Reaction score
517
14900k





14th Gen ready



Sadly only the Dark Hero Z790 caught my eye
 
Last edited:

Uramasa

Master Member
Joined
Jul 26, 2005
Messages
2,728
Reaction score
719
I think the most value on the new "refresh" line up will be 14700k.
Perhaps, will camp for price drop on the current 13700k and get those instead.
 
Important Forum Advisory Note
This forum is moderated by volunteer moderators who will react only to members' feedback on posts. Moderators are not employees or representatives of HWZ Forums. Forum members and moderators are responsible for their own posts. Please refer to our Community Guidelines and Standards and Terms and Conditions for more information.
Top