*** AMD "Zen 4" thread (inc AM5/APU discussion) ***

[g67575]

"DDR5 RAM automatically runs in 1:1 mode between UCLK and MCLK up to a clock rate of 6000 Mbps and in 2:1 mode above that. However, the 1:1 mode can still be set manually and operated stably with higher clock rates like DDR5-6400, even in dual-rank with 4x 16 GB".

So, there's the answer I was looking for.
From here:

The big Ryzen 7000 Memory and OC Tuning Guide - Infinity Fabric, EXPO, Dual-Rank, Samsung and Hynix DDR5 in Practice test with Benchmarks and Recommendations | Page 10 | igor´sLAB

AMD’s new Ryzen 7000 desktop CPUs, based on the Zen4 micro-architecture, still use the same chiplet design as their predecessors, with a few small but not negligible changes. Igor already gave details…

www.igorslab.de

A bit more commentary
'Unlike Zen 3, the 2:1 mode does not have any major performance disadvantages, so the advantage of the higher clock rate could already outweigh from DDR5-7000 onwards'

 

[brostradamus]

Does anyone know if we'd need new motherboards, in case if ddr5 spec changes to accomodate higher clock speeds?

 

[Armageus]

Does anyone know if we'd need new motherboards, in case if ddr5 spec changes to accomodate higher clock speeds?

No - DDR5 is DDR5, the spec cannot suddenly change.

JEDEC may certify higher speeds at some point, but they will be backwards compatible at a lower speed even if boards/chips aren't capable of running at a new higher frequency

 

[Andromeda1983]

So will there be any way to run say a 7600X under a NH-D15s and still keep your fans at about 600rpm with these chips or is the fan curve right out the f*ning window?

 

[Michael Nager]

That Skaterbench guy says he think all core 6ghz might be possible on the 7950x just with custom loop water cooling by doing manual tuning

First of all define, "all core 6GHz". Is that for a single benchmark run or for 24/7 operation?

Unless you define that point then the statement is totally meaningless.

A custom loop isn't a magic bullet, it only defers the time until the cooling system reaches saturation if you are running the system under load.

 

[brostradamus]

No - DDR5 is DDR5, the spec cannot suddenly change.

JEDEC may certify higher speeds at some point, but they will be backwards compatible at a lower speed even if boards/chips aren't capable of running at a new higher frequency

Thanks. Can the speed deficit be corrected through a cpu or bios upgrade? or are we still looking at new motherboards.

 

[Steedie]

Has anyone from OcUK said what time they go live yet? Is it the usual 2pm?

 

[Armageus]

Thanks. Can the speed deficit be corrected through a cpu or bios upgrade? or are we still looking at new motherboards.

What deficit?

It will be exactly the same as present with boards/cpu's - higher speeds may not be supported out of the box, but will be via overclocking. Whether higher speeds can be made stable depends on the quality of the motherboard (less so now since the vrm has been moved to the memory modules), and the quality of the CPU's memory controller.

You do not need to hold off for a later motherboard

 

[Dave2150]

Probably posted already (not scrolling through all the pages to check lol) - 20DC improvement in temps by delidding - why AMD used such a terrible IHS contact method is very strange, as there's clearly no need for such high temperatures.

Roman also got a 15W power consumption reduction from the lower temperatures, crazy!

Good luck to all attempting this!

 

[Gibbo]

Has anyone from OcUK said what time they go live yet? Is it the usual 2pm?

Yes 2pm and we have several hundred of each CPU in stock.

 

[uscool]

Probably posted already (not scrolling through all the pages to check lol) - 20DC improvement in temps by delidding - why AMD used such a terrible IHS contact method is very strange, as there's clearly no need for such high temperatures.

Roman also got a 15W power consumption reduction from the lower temperatures, crazy!

Good luck to all attempting this!

because AMD wanted AM4 coolers to be compatible and the height needed to be higher thats why the lid is thicker its not a terrible contact its because it needed to be thicker if you watched the video he explained it good decision or not I dont know , also it doesnt throttle and not any worse than intel ? also I like to see what the mounting pressure is like where we know intel wasnt great with

 

[Steedie]

Yes 2pm and we have several hundred of each CPU in stock.

Thanks Gibbo

 

[tedrodgers]

hi guys

im currently running an i5-10400F @2.9ghz (& 3060ti) for my gaming pc.

im thinking of upgrading soon.... possibly to the Ryzen 7 7700X.

thoughts? I presume that would be a decent upgrade for $$$.

cheers

 

[Curlyriff]

That i believe is why motherboards are being priced through the roof. They feel their part will last longer in a new build, so they make less money overall as people don't upgrade motherboards as much. Now with AM5 having DDR5 and PCIe5 and AMD supporting it for years, same as AM4 really. Their just won't be money in new motherboard next year, or the year after or even the year after.

But they are showing to still be on par cost wise with Intel versions. It isn't like it AMD specific there. The leaked Euro top end mobos were showing more for Intel compared to latest AMD also. And do we expect the new Mobo generation Intel will drop with next launch to be cheaper?

There is always someone looking to upgrade refresh from say a current gen 3000/5000 or 11/12 gen for Intel etc that want the latest board when do upgrade.

 

[Grim5]

because AMD wanted AM4 coolers to be compatible and the height needed to be higher thats why the lid is thicker its not a terrible contact its because it needed to be thicker if you watched the video he explained it good decision or not I dont know , also it doesnt throttle and not any worse than intel ? also I like to see what the mounting pressure is like where we know intel wasnt great with

Would have been nice if he tested stock rather than fixed clock. Yes 20c off is nice, but I want to know how much extra performance I'll get.

Thanks to other testing he's done on LN2, I know that every 20c off gets an extra 100mhz boost clock when it's stock. Hopefully the clock gains are more impressive with a full overclocking test run to see how high it can go

 

[TrojanWhore]

is there any benefit to using EXPO if standard kits works?

As Gibbo said, maybe not directly from the EXPO aspect, but looks like a better bet just on the side of freq/timings for the price so far:

Corsair Vengeance EXPO 32GB (2X16GB) DDR5 PC5-48000C36 6000MHz Dual Channel Kit - Black (CMK32GX5M2D6000Z36)

Order Corsair Vengeance EXPO 32GB (2X16GB) DDR5 PC5-48000C36 6000MHz Dual Channel Kit - Black (CMK32GX5M2D6000Z36) now online and benefit from fast delivery.

www.overclockers.co.uk

although does seem odd that those vengeance kits have better stats than any of the dominator EXPO kits listed so far! Maybe more to come? (there are some 6200 C36 non-EXPO there, but more expensive!)

 

[Grim5]

My local store has started listing expo kits, weird clocks and timing on them - standard xmp3 kits running higher clocks and tighter timings for cheaper

 

[timorous]

For gaming my take away from these results is wait for zen4 3d (and i didnt realise quite how good the 5800x3d was). I so wanted to like ZEN4, but feels a little flat tbh, unless you need ultra fast for productivity. I was hoping for some good old AMD price / performance stars but the 7600x doesn't seem like thats it. In fact, AMD and INTEL seem to have really very similar offerings right now, hot (even if by design), power hungry (although different reviews have very different values for the 7950x) similar perf and around the same price.

Indeed. The ComputerBase results showing that at 4.4Ghz the 5800X3D is 29% faster than a 5800X is pretty wow. Provided the 7800X3D does not have a clock speed regression vs the 7700X then that is likely to be a ~30% gain making Zen 4 3D the clear gaming champ.

I see no way of RPL competing with the performance of Zen 4 3D unless they price it really low but Intel are already hurting on margins so they probably don't want to do that.

 

[Curlyriff]

because AMD wanted AM4 coolers to be compatible and the height needed to be higher thats why the lid is thicker its not a terrible contact its because it needed to be thicker if you watched the video he explained it good decision or not I dont know , also it doesnt throttle and not any worse than intel ? also I like to see what the mounting pressure is like where we know intel wasnt great with

Honestly with most blocks etc needing a new bracket anyways it seems like making it thicker was wasted. They should have just got companies to make a new mounting system for the lower Z height and be done with it since the holes otherwise are same and it honestly wouldn't of been any different, made chips about 10degree cooler and saved 5-10watt still.

 

[killswitch73]

call me dumb but i cant seem to find any info in where to buy the new 7000 series in the UK . I'm assuming we a a later release date than that of the 27th September (today).