The AMD Ryzen All In One Thread /Overclocking/Memory Speeds & Timings/Tweaking/Cooling Part 2

That single core is crazy. Probably higher than Rocket lake will achive.

 

F12 BIOS. (Gigabyte B550 aorus elite) . Much better PBO :
Single core ( two best cores ) -20 negative + 300 mhz PBO. ( sigle core boost is flat 4950 mhz).
-10 on other cores . Voltage is AUTO , shows 1.36v during AVX stress test.
Funny thing is temperatures , before (even stock) it was around 72-74C, now it peaks 64C ...67C AVX, despite better performance WTF ?

 

Here's my 3600. I have been unsuccessful at purchasing a 5600x @ retail..... so this will hold me over until such point I can buy one. This is 4400 @ 1.31875 set & with vdroop getting ~ 1.28

 

AMD Ryzen & EPYC CPUs Based on Next-Gen 5nm Zen 4 Cores Rumored To Feature Over 25% IPC Gain

The latest rumor comes from ChipsandCheese whose sources have reported them that Genoa, AMD's next-generation server family, is going to be insanely fast compared to existing AMD EPYC offerings.
The source also talks about Zen 3+ and Zen 5 but most of the information that has been shared is related to the Zen 4 core architecture.

As per the rumored information, AMD's Zen 4 core architecture is going to offer a huge leap in IPC. Their sources report that existing Genoa engineering samples with the same number of cores and clocks as Milan CPUs are up to 29% faster.
This is a huge jump even when considering the jump to Milan from Rome. The AMD Zen 4 core architecture will be fabricated on TSMC's 5nm process node and is expected to launch sometime in 2022.

Zen 4 is what a lot of people are waiting for, and, if the info I have is accurate, that wait will prove to be even more worth it. It is important to note that the one common thread in all Zen 4 chatter I have heard is resounding positivity.
From IPC gains over 25%, a total performance gain of 40%, and even possibly (finally) 5GHz all-core thanks to the new (full node) N5 fabrication at TSMC!

Now, I can’t say what is true and what is an over-exaggeration, however I was told from a trusted source that a Genoa engineering sample (Zen 4 server chip) was 29% faster than a Milan (Zen 3) chip with the same core config at the same clocks.
Factor this in with what I have heard about the possible clock gains that N5 will enable over N7 and Zen 4 sounds like it is going to be a monster of a CPU.

via ChipsandChesse
The overall performance gain for the Genoa CPUs could very well be around 40% considering that the CPUs are going to feature more than just a core architecture upgrade.
There will be increased clock speeds and also key improvements to the IMC along with support for faster DDR5 DIMMs and an improved interconnect fabric that will communicate with the several cores.
The Zen 4 CPUs are stated to offer an increase in core count which was hinted by AMD's CEO, Dr. Lisa Su, in a previous interview.
The 4th Gen EPYC Genoa CPUs are expected to feature up to 96 cores though we could end up with an even higher number if AMD really wants to push the boundaries in the server CPU segment.

‘Mark, Mike, and the teams have done a phenomenal job. We are as good as we are with the product today, but with our ambitious roadmaps, we are focusing on Zen 4 and Zen 5 to be extremely competitive.

via Anandtech

There are also details regarding Zen 3+ which will act as an intermediate solution and the first line of chips supported by the brand new AM5 platform.
We know that Zen 3+ will act as a refresh of sorts and is internally known as Warhol. It is expected to be fabricated on TSMC's N7 process that is a variant of N7 but utilizes 5 layers of EUV.
So while it ain't technically a new node, it will offer a range of improvements, and the source reports that we can expect nominal IPC gains of 4-7% and that alone is better than the 3% IPC gains we got with Zen+.

More -> https://wccftech.com/amd-5nm-next-generation-zen-4-ryzen-epyc-cpus-to-feature-over-25-ipc-increase/

 

I hope Zen 3+ doesn’t get more expensive.

 

Either way I'm skipping Zen 3+

 

I probably am too but another increase will price AMD into oblivion. Then again rumor has it the 8c rocket lake will be north of $600 USD.

 

BIOS 5809-AGESA V2 PI 1.1.8.0.
This BIOS allowed PBO boost as high as CPU was stable. Was able to push 5800X to 5075mhz rock stable.



BIOS 5837-AGESA V2 PI 1.2.0.0
This BIOS PBO boost is locked to 5025mhz no matter the setting. Even though AMD says PBO max should be 5050mhz(4850mhz+200mhz boost= 5050mhz)


Still great performance but not happy about being locked OC wise. AGESA V2 PI 1.2.0.0 did allow me to increase curve op. which helped get gap between two closer. Feel that this BIOS is more stable but its neutered aliitle.

 

Finally was able to put together my first custom loop. Just cpu for now. Also switched motherboard manufacturers; my first MSI board. All based around a 5800X. Waiting on leak test run to finish before can see what the cpu can be pushed to.

 

The universe demands pictures posted specifically to this forum.

 

fiddled with CTR and found that CCX1 can't even run past 4025mhz so there is literally no point in using it as my cores are almost always running at 4.2-4.3

 

AMD needs to make curve optimizer available for the 3000 series. So much easier to setup and use.

 

5950X and 1080Ti ,nice combo. [/ATTACH]

 

It is not a pretty build. It is functional. Will post a picture of it tonight. The 1080 is still being used till the new gpu arrives Friday. The 5800X does run hot. Before dialing in the curve optimizer the cpu would actually hit 73-80. Hit 90 once. Now idle temp according to Hardware info is 26 and Ryzen Master says it is 23. Ambient is like 15C/60F.

Was regularly seeing Core VID registering 1.5v+. Nothing seemed to work till I lowered the scaler multiplier under PBO from 10 to 9. Now the max Core VID seen is 1.469 to 1.481.

Spoiler: Rig Picture

 

From:
https://www.reddit.com/r/Amd/comments/lnkb8k/ryzen_master_2611797_update/

https://www.amd.com/en/technologies/ryzen-master

https://www.amd.com/en/support/kb/release-notes/rn-ryzen-master-2-6-1-1797

Ryzen Master version 2.6.1.1797

Sounds like this improves support for Dual CCD 5600's and 5800's the ones that shipped with two of these on the CPU even though one is disabled.

 

https://twitter.com/9550pro/status/1364550020540768260

From:
https://www.reddit.com/r/Amd/comments/lrf0cr/agesa_1201_fix_l3_cache_performance/

Looks like MSI has posted a new 1201 AGESA code for a few motherboards and one of the fixes is improved L3 cache performance?

Copying that image off of Twitter.


And from Reddit.

Also reads like there's not much actual performance to be gained but the throughput is now much higher.

 

if anyone still needs some hynix cjr timings, this are mine working for like 1 year without any problems so far solid stable.

 

It also locked the max PBO boost to 200MHz and cpu will now boost to that full 200MHz (5050MHz).
Temps are better by 1-2c.
L3 cache throughput did increase but no noticeable difference in game.

 

Full changelog now.
https://twitter.com/AMDRyzen/status/1364971430337740804

From:
https://www.tweaktownforum.com/foru...abyte/28656-gigabyte-latest-beta-bios/page918