Discussion AMD's Ryzen 7 5800X3D gaming CPU's "Overclocking Lock" bypassed ? CHIP spotted running at 4.8 GHz,

[Metal Messiah.]

Hi guys,

As you may already know by now, that AMD's Ryzen 7 5800X3D, the world's first 3D V-Cache chip hits retail next week and it looks like overclockers have already bypassed its OC-Lock.

The AMD Ryzen 7 5800X3D is going to be the first and only chip with 3D V-Cache based on the 7nm Zen 3 core architecture. The CPU is going to offer 8 cores, 16 threads, and 100 MB of combined cache thanks to its additional 64 MB 3D Stacked SRAM design. Clock speeds will be maintained at 3.4 GHz base and 4.5 GHz boost with a TDP of 105W.

But before continuing further, it is also worth mentioning that according to AMD's Robert Hallock, the Ryzen 7 5800X3D CPU has overclocking hard locked to avoid physical damage to the 3D V-Cache.

The reason for this lock in the first place is due to the voltage curve. While AMD's Zen 3 chips can reach up to 1.5-1.6V with overclock, the 3D V-Cache stack that sits on top of the Zen 3 cores can only do 1.35V and it is already running at its limit out of the box. If users try to overvolt the chip beyond that voltage curve, they can break the chip and hence why overclocking is not supported for the CPU.

AMD only said that it will enable memory & FLCK overclocking for the said chip.

However, Pieter of SKATTERBENCHER seems to have proven AMD wrong and bypassed this overclock lock. The overclocker shared a pictured of CPU-z in which the Ryzen 7 5800X3D can be seen with an operating frequency of 4.82 GHz, far above its 4.5 GHz max boost clock. Though please sprinkle some salt on this discovery/feat for the time being.

Anyway, the chip was running on the ASUS ROG Crosshair VIII Extreme motherboard with a 1.306V voltage and a 45.5 multiplier and a BCLK of 105.99 MHz. The overclocker did use ASUS's Voltage Suspension technology to make it work & will have a video out explaining the whole process soon.

This proves that the 3D V-Cache chip can indeed be overclocked and users can get additional performance though we don't know what impact the overclock has on the cache itself and what kind of temperatures/power it produces with such overclocks. Hopefully, we will be getting more details on overclocking once the first reviews hit the web tomorrow.

The Ryzen 7 5800X3D is turning out to be a great gaming CPU as far as its price is compared to the Intel flagship, the Core i9-12900K. AMD can have a very disruptive offering as it says its last hurrah to the AM4 platform. For good or bad only time can tell ?

https://twitter.com/i/web/status/1514063342440116229

View: https://twitter.com/skatterbencher/status/1514063342440116229

https://twitter.com/i/web/status/1514071272979931138

View: https://twitter.com/Buildzoid1/status/1514071272979931138

 

[hotaru.hino]

All this is proving is you can overclock using BCLK instead of the multiplier, much like how you can overclock Intel's non-K processors with BCLK. And while you can still throw a pie at AMD's face for saying this chip isn't overclockable for whatever reason, I think this is just proves the mess of what "overclocking" really means today.

 

[TerryLaze]

It didn't prove AMD wrong , all he did is get into windows at 4.8 and using 1.3V ... put any load on that and you will easily go above the 1.35 that will start burning the chip.

 

[hotaru.hino]

It didn't prove AMD wrong , all he did is get into windows at 4.8 and using 1.3V ... put any load on that and you will easily go above the 1.35 that will start burning the chip.

I'm pretty sure the person fixed the VID to that point, so it's not going to move from there.

 

[TerryLaze]

I'm pretty sure the person fixed the VID to that point, so it's not going to move from there.

That would mean that under any load it would run slower so it wouldn't be a overclock anymore. If power draw increases then either Vcore has to increase with it or clocks are going down.

 

[drea.drechsler]

I'm pretty sure the person fixed the VID to that point, so it's not going to move from there.

That sounds like something Asus' Voltage Suspension Technology (available only on certain BETA BIOS' of one of their super-premium Crosshair boards) would provide for. This doesn't sound like something for the masses.

That would mean that under any load it would run slower so it wouldn't be a overclock anymore. If power draw increases then either Vcore has to increase with it or clocks are going down.

I agree it would be interesting to see how it operates as load varies. But do remember that the point of this CPU is for gaming which presents very light loads on the CPU. Boosting to 4.8Ghz even briefly should only help performance.

 

[hotaru.hino]

That would mean that under any load it would run slower so it wouldn't be a overclock anymore. If power draw increases then either Vcore has to increase with it or clocks are going down.

Not to sound rude but have you actually tweaked with an AMD system like this before?

You can set a fixed Vcore setting and the processor can still run through its frequency steps like normal. The problem is if the voltage is too low to keep the frequency stable enough for the computer to like it. And while I don't know if the person here fixed the multiplier, you can do that too with a fixed Vcore setting. Power consumption is simply dictated from there by the current going through the processor, which fluctuates based on workload.

I mean, I should know. I did this with several Zen based AMD processors when trying to get them to be more efficient.

 

[drea.drechsler]

....
while I don't know if the person here fixed the multiplier, you can do that too with a fixed Vcore setting.
....

It's actually a bit confusing what all he did, but it seems he almost certainly could not fix the multiplier since it's overclocking locked. And of course, he could not fix VCore...except perhaps with the Asus Voltage Suspension Technology. How that could work is a bit mysterious too since it's supposedly only present in certain BETA BIOS's that shouldn't have support for the 5800X3D. That suggests a custom, hacked BIOS maybe?

But what I do see is that he raised busclock to get the higher clocks. I'd imagine it would work the same as raising busclock on any board: everything rises in measure with it. But if you don't have access to PBO and the ability to change any parameters it's just going to continue boosting only to the limits of PPT, EDC and TDC the algorithm allows. I don't think changing busclock messes with that so....will it matter much for actual performance?

 

[hotaru.hino]

It's actually a bit confusing what all he did, but it seems he almost certainly could not fix the multiplier since it's overclocking locked. And of course, he could not fix VCore...except perhaps with the Asus Voltage Suspension Technology. How that could work is a bit mysterious too since it's supposedly only present in certain BETA BIOS's that shouldn't have support for the 5800X3D. That suggests a custom, hacked BIOS maybe?

But I know some people have run their 5800X's at 4.8G, fixed overclock at voltage as low as 1.35V so it seems reasonable. But, of course, they have crap stability for heavy all-core workloads. But again, this is for gaming so does that matter?

As far as I know, a "locked processor" just means you can't tweak the multiplier beyond the specifications. You can tweak everything else about the processor to your heart's content.

The goal of a locked processor isn't so much to prevent people from tweaking with it period. It's to prevent people from taking cheaper parts, overclocking them, and then selling them as higher end parts.

 

[drea.drechsler]

As far as I know, a "locked processor" just means you can't tweak the multiplier beyond the specifications. You can tweak everything else about the processor to your heart's content.

The goal of a locked processor isn't so much to prevent people from tweaking with it period. It's to prevent people from taking cheaper parts, overclocking them, and then selling them as higher end parts.

That would suggest PBO is still allowed (it does not fiddle with clocks, only current and power limits)? I think we'll have to wait and see since the dribbles and drabbles are all very short on details.

 

[hotaru.hino]

That would suggest PBO is still allowed (it does not fiddle with clocks, only current and power limits)? I think we'll have to wait and see since the dribbles and drabbles are all very short on details.

I'm also certain PBO was meant to be a reward for using a better cooling system and not necessarily an overclocking feature.

Which again, this seems to reinforce my disdain over the use of the term "overclocking" by itself. Does it mean adjusting the multiplier beyond the specifications? Does it mean adjusting the base clock speed? What?

 

[drea.drechsler]

....
Which again, this seems to reinforce my disdain over the use of the term "overclocking" by itself. Does it mean adjusting the multiplier beyond the specifications? Does it mean adjusting the base clock speed? What?

It certainly does seem to be a term that needs better definition, but in this case it can mean anything AMD wants it to mean.

 

[Metal Messiah.]

Thanks for the input and your insight on this guys ! On some other news(a bit off topic), but obviously related to CPUs, it appears that AMD's next-generation upcoming B650 motherboard chipset, featuring the AM5 socket has allegedly been spotted running a next-gen AMD Ryzen 7000 Desktop CPU.

The slide looks legit, and the leaker has a decent track record when it comes to leaks, though we need to exercise some caution here.

Anyway, in a picture tweeted by @9550Pro (HXL), we can see an off-screen BIOS display that shows the motherboard with the MAG B650 naming scheme. The MAG series motherboards are designed exclusively by MSI and are the most entry-level in their motherboard family. While you can't say which specific MAG variant this is (Mortar, Bazooka, Tomahawk), you can see that it is based on AMD's B650 chipset which will be part of the next-generation AM5 platform.

https://twitter.com/i/web/status/1514620039789363201

View: https://twitter.com/9550pro/status/1514620039789363201


The CPU isn't mentioned but it should be an AMD Ryzen 7000 'Zen 4' engineering sample/ES. Another thing that is very interesting is the Vcore voltage. The CPU can be seen running at 1.532V which is considered really high and mostly only seen in LN2-OC use cases.

Such a high voltage seen here could be due to the chip still being an early sample that has not yet been optimized. But what do you think about this ? Could this also be a bug ?

Also, AMD usually provides its board partners with test samples several quarters prior to launch, so it migth be possible that motherboard vendors are already prepped up with their initial design, if they are already testing these chips out. This aligns with the recent reports/rumors of mass production commencing this month and a launch happening as early as Q3 2022.

 

[hotaru.hino]

My best guess is, if the reading is accurate, they're doing electrical stress testing to really determine the limits of the CPU. Granted results of such testing should be taken with a grain of salt unless AMD vetted that it was basically a gold standard chip.

 

[drea.drechsler]

Since we're open to guesses... mine is it's the result of a faulty BIOS preset. The sort of screw-up you try to find in alpha testing leading up to release.

 

[Metal Messiah.]

It appears that Tom's HW have covered this topic today. Just saw this now. As noted BCLK OC can be prone to RTC Bugs as well.

https://www.tomshardware.com/news/ryzen-7-5800x3d-overclocked-to-4-74ghz.

 

[Metal Messiah.]

I think select MSI motherboard users are in for a treat. The company plans to enable AMD Ryzen 7 5800X3D BCLK overclocking with the upcoming AGESA 1.2.0.7 BIOS. It's not a very big deal, but at least one of the AIBs is giving some sort of official OC benefit on their flagship motherboards.

It's the same method used by the original overclocker Pieter of SKATTERBENCHER who did with his CPU, BCLK overclocking, as outlined in the first post/OP under this thread. Also, I presume this can only be done on motherboards sporting an external clock generators, as other models lack this feature.

Anyways, in a few days time, MSI will be rolling out the latest AGESA 1.2.0.7 BIOS Firmware for its motherboards that will allow overclocking support on the new AMD Ryzen 7 5800X3D 3D V-Cache CPU. The latest BIOS comes with a range of features.

MSI's MEG X570 GODLIKE motherboard & other options with external clock generators have the ability to allow overclocking on the 3D V-Cache chip and MSI will be releasing a new BIOS that will let users do just that.

The new MSI BIOS is also a special one since it is based on the upcoming AMD AGESA 1.2.0.7 BIOS firmware. From the changelog provided, the AGES 1.2.0.7 offers:

1. Fix the fTPM issue that causes game stuttering.
2. 300 series chipsets support the latest Ryzen CPU/APU such as Ryzen 7 5800X3D

MSI shared a screenshot showcasing up to a 107 MHz BCLK overclock. Default value should be 100 MHz. The CPU hits 4.86 GHz clock speed due to this OC.

One cool feature about the upcoming BIOS, more specifically for the MSI MEG X570 GODLIKE, is that it will offer a new 'CPU Base Clock (MHz)' tuning option within it. This will be available on the AGESA 1.2.0.7 based 1H1 BIOS and users can adjust the base clock of their chip for overclocking. Following are the MSI boards that will offer BCLK overclocking for AMD Ryzen 7 5800X3D CPU:

• MEG X570 GODLIKE
• MEG X570 ACE
• MEG X570 Unify

As expected, the AMD Ryzen 7 5800X3D offered a 5% gain in multi-threaded and 7% gain in single-threaded tests. This should even out the performance disparity that the AMD Ryzen 7 5800X3D had against the Ryzen 7 5800X due to its lower clocks.