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

[Bencher]

As someone who is building a rig primarily for MSFS, it looks like I am going 5800X3d unless raptor lake has anything to tempt me.

If you only care about MSFS, I don't think raptorlake can touch the 3d, unless you are willing to spend 400+ on DDR5. The hynix A die kits can hit 7600+ mhz, so those on a 13900k can maybe - probably do it, but at a huge cost compared to the 3d.

 

[Joxeon]

Tech power up is garbage. So many of their results are GPU bound as they only used a 3080

What about the 720p testing which showed the 12600k pull further into the lead? Are you saying a 3080 is GPU bound at 720p?

 

[LtMatt]

I couldn't understand their results, didn't realised they only tested with a 3080.

Also one other factor, there is a difference in performance seen on Ryzen 7000 series when using an AMD GPU like a 6950 XT/6900 XT, vs using a 3090 TI/3090.

The Ryzen 7000 series is faster on the AMD GPUs, take a look.

Spoiler

AMD Ryzen 9 7950X und Ryzen 7 7700X im Test: Die Kerne der Zukunft

Diese Seite beschäftigt sich mit der Leistungsaufnahme aller Prozessoren im Test und zeigt den großen PCGH-Effizienz-Index, unterteilt in Gesamt, Spiele und Anwendungs-Effizienz. Mithilfe eines Hardware-Tools messen wir die Leistungsaufnahme der CPU allein, mit Einflussnahme der Spannungswandler...

www.pcgameshardware.de

 

[tommoT.Striker]

I only took a look at 7950x results, maybe gaming results aren't special, the 5800X3D still blows everything away, but for production if you can utilise the 7950x's cores it's a beast.

As for the temps, i'm used to it, people panic when temps get high, for me more of a concern is the power it can use when it clocks very high. But with the market being the way it is, everyone is ramping up power usage.

Shame you won't see 24+ Zen 4 cores in a non-Epyc rig running a fairly respectable clock for a while yet.

 

[uscool]

Heres another one for bencher to sweat over and make bogus metric claims before I get off for the night

 

[LtMatt]

I know how efficient 12900 is. 24.5k CBR23 at 125w (stock PL1). Just a tad less efficient than the 5950x at stock. Granted, out of the box is kind of irrelevant, at least for me, but it's actually a fact that in CBR23 the 12900k at stock PL1 is more efficient than the 7950x at stock.

I bet you must have been really impressed with the 7950X managing 31K in CBR23 at only 65W right?

AMD Zen 4 Ryzen 9 7950X and Ryzen 5 7600X Review: Retaking The High-End

www.anandtech.com

 

[Bencher]

Heres another one for bencher to sweat over and make bogus metric claims before I get off for the night

That's the one from club386 right? Thanks, that's the one im talking about. The 12900k at stock is more efficient than the 7950x at stock, LOL XD

38004/230 = 165p / watt
23976/125 = 191p / watt

Thanks a lot man

 

[tommoT.Striker]

Puget had a review of the 7950x but couldn't get the decode working properly on Davinci Resolve but looking at the specs the AV capabilities are strong , AV1 decode, H264 and HEVC encode, so just needs AMD and Black Magic to get an update out: https://videocardz.com/newz/amd-confirms-ryzen-7000-integrated-rdna2-graphics-features

 

[Bencher]

I bet you must have been really impressed with the 7950X managing 31K in CBR23 at only 65W right?

AMD Zen 4 Ryzen 9 7950X and Ryzen 5 7600X Review: Retaking The High-End

www.anandtech.com

Are we talking stock or not? Cause i remember when i was power limiting my 12900k people here said that only stock matters. So, at stock the 12900k is more efficient at long workloads after the 56 seconds of TAU2 have expired. Deal with it

 

[uscool]

I bet you must have been really impressed with the 7950X managing 31K in CBR23 at only 65W right?

AMD Zen 4 Ryzen 9 7950X and Ryzen 5 7600X Review: Retaking The High-End

www.anandtech.com

He'll just twist it regardless if it is correct or mostly incorrect but will give credit where it's due absolute relentless lol

 

[LtMatt]

Are we talking stock or not? Cause i remember when i was power limiting my 12900k people here said that only stock matters. So, at stock the 12900k is more efficient at long workloads after the 56 seconds of TAU2 have expired. Deal with it

ECO mode is not stock, but it can be applied with a click of a button in Ryzen Master Tool. I just thought you might like the efficiency of it.

 

[uscool]

Mind blowing he won't give credit to that cb23 score at 65 watts

 

[Bencher]

ECO mode is not stock, but it can be applied with a click of a button in Ryzen Master Tool. I just thought you might like the efficiency of it.

Power limit could also be applied with a click, you don't even have to get into the bios, but people kept insisting that comparisons should be done stock out of the box and that's why the 12900k is a furnace. How times have changed, right?

 

[LtMatt]

As someone who is building a rig primarily for MSFS, it looks like I am going 5800X3d unless raptor lake has anything to tempt me.

This might be of interest to you. I will assume the 7950X will be a bit faster too, but not found any other testing of this game yet.

AMD Ryzen 9 7900X CPU Review - Page 7 of 9

AMD's new Zen 4 based Ryzen 9 7900X CPU is here, and we have all the benchmarks, synthetic, rendering, and gaming with a 10 CPU comparison.

www.thefpsreview.com

I have not fully disected the results difference between this and Linus review so might want to take a close look before drawing conclusions.

To be fair, I don't think you can go wrong with a 5800X3D, especially if you are already on AM4.

 

[Calin Banc]

@Calin Banc The thing with TPU CPU reviews is they are all pretty GPU heavy, its been like that for a few years, they seem to think its fashionable now, they ain't wrong but whats fashionable is not necessarily right.

The 7600X is way more than 13% faster than the 2600X, in reality, that slide is accurate to those in the hear and now who run games in such a way where its only ever a GPU bottleneck, outside of whatever he's doing to get a result so tightly clustered you will find that your 2600X is actually quite old and outdated now.

I'm not saying go out and buy a 7600X, or any other CPU, but that slide is a nonsense for the reality of top of the range GPU's, especially for those that are coming, if you think a 7600X is 13% faster than a 2600X then there is the reason reviewers like this are a joke. the reality is the 7600X is about twice as fast.

@Calin Banc reality vs Tech Power Up, the Ryzen 3600 is actually quite a good chunk faster than your Ryzen 2600X. the difference between your 2600X and the 7600X, outside of TPU's weird testing, is probably around 80%+

There's also the 720p results.
7600x is 87% faster than 2600 - so in line with others I saw.

5600 is about 55% faster than 2600.

7600x is about 21% faster than 5600.

A big slice of that performance gap between 2600 and 7600x is offset by the 5600. And if I'd go for a whole system upgrade then I'd go for a 7700x which would put the whole thing over $1000. Vs $175.
I could also look at at 5700x or 5900x and still end up way better.

Performance is not so bad after all, just the price is. So for now is not that bad to wait. Probably I'll skip again this gen in CPU and GPU tech and slide in some 5xxx or something.

 

[timorous]

That's the one from club386 right? Thanks, that's the one im talking about. The 12900k at stock is more efficient than the 7950x at stock, LOL XD

38004/230 = 165p / watt
23976/125 = 191p / watt

Thanks a lot man

Since when is the 125W score stock for the 12900K?

Either you compare the stock vs stock numbers of 38k and 27k and the respective power draw, whatever that may be, or you compare the 125W numbers to each other.

Comparing stock vs the 125W 12900K score is disingenuous, and you know it.

 

[=XDC=FluphyBunny]

They used a 3080 at 720p. That's better than running a 3090ti at 1080p like hwunboxed. So wtf are you talking about? LOL

Give over. Results even at 1440 were massively gpu limited hence why they don’t match others results and hence why their roundups are skewed.

 

[SKILL]

That's the one from club386 right? Thanks, that's the one im talking about. The 12900k at stock is more efficient than the 7950x at stock, LOL XD

38004/230 = 165p / watt
23976/125 = 191p / watt

Thanks a lot man

lulwat

You're not comparing stock to stock there are you? You're comparing 7950X 'stock' with 12900k 125W... Either you're being dense or intentionally misrepresenting things, I think we all know which one it is.

If you need help the 12900K at stock is 27,694, divided by 240W, gives 115p / watt vs 165 for the 7950X.

Or you could compare 125W vs 125W in which case the 7950X is 289p / watt vs 191 for the 12900k.

 

[uscool]

Wow that mfs benchmark is impressive 6 cores beating 12900k

 

[LtMatt]

Blimey this thread is updating fast. Made a post and there's half a dozen to read after mine.