AMD's Robert Hallock confirms Ryzen 7 5800X3D overclocking limitations

nanoguy

Posts: 1,355   +27
Staff member
Recap: The Ryzen 7 5800X3D is a $450 CPU landing on April 20 for people with a 300, 400, or 500 series motherboard. AMD's performance claims look promising for gamers seeking an upgrade, but overclocking enthusiasts will want to wait for the Ryzen 6000 "Rembrandt" lineup.

Earlier this month, we heard that AMD's much-awaited Ryzen 7 5800X3D CPU would not support manual overclocking as the company asked motherboard manufacturers to disable the functionality in their BIOS/UEFI.

Fast forward to this week, and we now have official confirmation from Team Red that its newest Ryzen 7 series processor for gamers isn't overclocking friendly. Apparently, the main reason for this limitation is AMD's 3D V-Cache, the technology that's supposed to make this CPU up to 15 percent faster in games when compared with the Ryzen 9 5900X.

To AMD's credit, it will open up support for this processor as well as other Ryzen 5000 variants on 300 series motherboards via an upcoming BIOS update. However, AMD's Robert Hallock told HotHardware the 3D V-Cache is limited to a maximum working voltage of 1.3 to 1.35 volts, and as such it doesn't even support the typical boost voltages seen on some Ryzen CPUs, which are in the 1.45 to 1.5 volt range.

This would also explain why AMD chose to lower the base and boost clocks of the Ryzen 7 5800X3D to 3.4 GHz and 4.5 GHz, respectively. Hallock notes the Infinity Fabric and memory bus can still be manually overclocked, and that it should lead to better performance gains than an all-core overclock, anyway.

According to Hallock, the Ryzen 7 5800X3D isn't a sign that AMD is looking to neglect overclocking enthusiasts moving forward. The company is still working around the limitations of the existing manufacturing technologies that make its chiplet designs possible, so the next generation of Ryzen CPUs should have better overclocking support.

Hallock also explained during the interview that AMD chose to go with an eight-core configuration because gaming workloads have yet to benefit from adding more cores beyond that. As our own Steve Walton explained in his detailed analysis of how CPU cores & cache impact gaming performance, the CPU cache size is much more important than the number of cores, and AMD seems to be sold on this idea as well.

Permalink to story.

 
Will be interesting to see the reviews. Ryzen right now gains more performance from FLK and low latency memory more so than clocks

If this chip can do 2000 FLK with PBO+CO and DDR4000 Memory may still be plenty fast
 
Last edited:
Core OC'ing is disabled at chip level, because it could do physical damage. No bios needed.
 
Last edited:
Overclocking has technically never been an "officially" supported feature (beyond the supported boost frequencies) so it's not like we're "losing" a feature now that CPUs are already running at near full potential out of the box. Back in the Sandy Bridge days it was sooo easy to get 1Ghz+ overclocks because Intel was far more conservative with their binning and TDP figures actually meant something useful.

But now competition is fierce again so CPUs are squeezed for all they got at every binning tier. This is why Silicon Lottery shut down. The writing was on the wall, and now overclocking is basically only for breaking records on LN2.
 
If this thing isn’t faster at gaming than the cheaper Alder lake alternatives it’s DOA.

A few people will still buy it - plus new Zens are around the corner .
If I was AMD - I wouldn't be concerned about sales - like you say bragging rights .
Plus showing prove of concept to commercial samples is a big thing .
Plus now they will have real world users to "test" the product

I think the main benefit is still an unknown product that might appear if they can really improve this .
Lots of companies have vapourware they tout
 
A few people will still buy it - plus new Zens are around the corner .
If I was AMD - I wouldn't be concerned about sales - like you say bragging rights .
Plus showing prove of concept to commercial samples is a big thing .
Plus now they will have real world users to "test" the product

I think the main benefit is still an unknown product that might appear if they can really improve this .
Lots of companies have vapourware they tout
It will sell *some* if it under-performs. Some AM4 platform owners may choose to upgrade.
It will sell very well if it offers better performance than the ADL 12700, many AM4 owners + new builders will make the purchase.
 
Does anyone overclock CPUs these days?


I've been overclocking since the P mmx/K 6-2 days, and I'm not about to stop now. Sure it's much much easier and the increase in performance not nearly as brag worthy. But I also run water and figure if I spend the extra money I might as well have something to justify it other then running close to silent. So you do you, me I'll do me...
 
Just a friendly reminder - 4k gaming is GPU constraint - so the young smart kids will buy a cheaper intel/amd CPU and RTX 3080 are getting bang for bucks - Most of us here are probably well off - and pick a higher tier - simply because we can and who knows what the future may bring .
There are so many things we pay for next tier - when we see stuff all difference . It's like the story Amazon buys MGM - oh goodie I will buy a Sony QD-OLED then watch those classic movies that are in the SDR colour space ( no real world diff - except brightness or POV , speakers etc - cheaper Sony OLED can do SDR 100%.)
Looking at comments about the new IPads - some people get it right are the port options better , or the screen , memory costs , or the speakers , mic etc - The old IPad was fast enough for 95% of users
 
Back