Info 64MB V-Cache on 5XXX Zen3 Average +15% in Games

[Kedas]

Well we know now how they will bridge the long wait to Zen4 on AM5 Q4 2022.
Production start for V-cache is end this year so too early for Zen4 so this is certainly coming to AM4.
+15% Lisa said is "like an entire architectural generation"

 

[nicalandia]

[URL
THG confirms that the 5800X3D is indeed the fastest gaming CPU*

*when using DDR4

OR when USING DDR5 6400Mhz Super Expensive RAM.




Look there is not way around this. The 5800X3D beats the 12900KS with RAM that is more expensive that the CPU itself.

 

[tamz_msc]

https://www.tomshardware.com/news/amd-ryzen-7-5800x3d-review

THG confirms that the 5800X3D is indeed the fastest gaming CPU*

*when using DDR4

That's what I said a few pages back, and people laughed at me.

OR when USING DDR5 6400Mhz Super Expensive RAM

View attachment 60033

Why are you fixated on Far Cry 6? Look at the averages.

 

[epsilon84]

OR when USING DDR5 6400Mhz Super Expensive RAMs...

Sorry, I'm not sure if you are being funny, or if you are actually ignoring the HWUB/techspot data?

 

[tamz_msc]

Sorry, I'm not sure if you are being funny, or if you are actually ignoring the HWUB/techspot data?

He only looks at those data points that are in favour of his narrative that the '5800X3D is the gaming king'. It's not, though it comes close, and is a good CPU.

 

[nicalandia]

Sorry, I'm not sure if you are being funny, or if you are actually ignoring the HWUB/techspot data?

Are you really Grasping at Straws at this stage? Are you going out of your way to purchase DDR5 6400 RAM to prove that you are less than 1% higher on an average of 8 games? Really?

 

[nicalandia]

He only looks at those data points that are in favour of his narrative that the '5800X3D is the gaming king'. It's not, though it comes close, and is a good CPU.

Grasping at Straws Again?

 

[epsilon84]

He only looks at those data points that are in favour of his narrative that the '5800X3D is the gaming king'. It's not, though it comes close, and is a good CPU.

Yeah, I'd agree. IMO the 5800X3D is a good 'proof of concept' CPU. I'm far more excited about what AMD can do with this tech going forward with Zen 4 and DDR5. Imagine Intel with the DDR5 advantage nullified!

 

[Karnak]

He only looks at those data points that are in favour of his narrative that the '5800X3D is the gaming king'. It's not

Depends entirely on the games and benchmark suite. You can test 10 games where the 5800X3D is only like 5% faster than the 5800X because in those specific games more cache doesn't matter a lot. But most sites I'am looking at, the 5800X3D is as fast or a bit faster than the fastet product Intel has to offer right now.

And to add something: What I care about is how fast a product is within it's official specifications. Cause everything else is OC and it's not guaranteed that RAM X at speed Y will work for everyone. Not to mention perf/$. That's another topic we don't want to talk about because it ends really bad for Intel.

 

[krumme]

Looking at Tech Jesus vid now, with its beatifull frame to frame consistency for this cpu, and i think some of the benefits of this huge cache is that reliable perf. I like it. To hell with the averages.

 

[nicalandia]

Imagine Intel with the DDR5 advantage nullified!

Imagine? You don't have to do that. the 5800X3D is slapping you on the face and telling you. Look what I can do with your DDR5 sticks...

 

[Timorous]

Are you really Grasping at Straws at this stage? Are you going out of your way to purchase DDR5 6400 RAM to prove that you are less than 1% higher on an average of 8 games? Really?

View attachment 60034

Technically they tested 9 games but they don't include Factorio in the average. The 5800X3D was 26% faster than the 12900KS in that game and if you include that average it probably tops the charts.

 

[coercitiv]

So, Schrödinger's CPU...?

As you can see, it definitely is Schrödinger's Gaming CPU: measuring it only makes the debate worse ))

 

[epsilon84]

As you can see, it definitely is Schrödinger's Gaming CPU: measuring it only makes the debate worse ))

Worse only if its not unaminously the gaming king...

Thou shall bow down. NOW!

 

[nicalandia]

Looking at Tech Jesus vid now

I am on it right now..

 

[epsilon84]

Imagine? You don't have to do that. the 5800X3D is slapping you on the face and telling you. Look what I can do with your DDR5 sticks...

OK, things just got weird. I'm out. LOL

 

[nicalandia]

OK, things just got weird. I'm out. LOL

This is not a Democracy, This is $Dictatorship. I for one welcome our $$$ Overlord.. That is until Zen4/Raptor Lake drops

 

[Borealis7]

I doubt that will happen. Devs will optimise for consoles because that is where the engines get used the most.

exactly! this CPU (or rather, CPUs with 3D V-cache) should be in consoles where gaming is the main focus and resolution isnt very high (yet)

 

[Makaveli]

exactly! this CPU (or rather, CPUs with 3D V-cache) should be in consoles where gaming is the main focus and resolution isnt very high (yet)

Maybe in the PS5 Pro only time will tell.

 

[tamz_msc]

Are you really Grasping at Straws at this stage? Are you going out of your way to purchase DDR5 6400 RAM to prove that you are less than 1% higher on an average of 8 games? Really?

For people who are purchasing a 3090 Ti to have the best GPU, purchasing a $450 kit of fast memory and overclocking it isn't really a big deal.

 

[epsilon84]

Are you really Grasping at Straws at this stage? Are you going out of your way to purchase DDR5 6400 RAM to prove that you are less than 1% higher on an average of 8 games? Really?

View attachment 60034

OK, I'm no maths genius, but I'm certain there is more than a 1% difference there! LOL at least I know you're just taking the piss now

 

[nicalandia]

For people who are purchasing a 3090 Ti to have the best GPU, purchasing a $450 kit of fast memory and overclocking it isn't really a big deal.

That is a good point...!

Perhaps the 5800X3D Combo with DDR4-3800 with 6900XT will give the same results at a 12900KS with 3090 Ti in games..

 

[Panino Manino]

OK, not gonna lie... this is unexpected. 12900K with DDR5 outperforms the 5800X3D in games.

Depends on the list of games.
The performance boost of 3D is massive, but there are games where Zen 3 massively performances for some reason. When you test the newer and stronger Intel architecture with those games even 3D can't compensate for the difference. You could call those "outliers", just like you can say that Borderlands 3 is an "outlier" for Zen3D.
Both are technically tied, it just depends on the games you want to play.

 

[epsilon84]

That is a good point...!

Perhaps the 5800X3D Combo with DDR4-3800 with 6900XT will give the same results at a 12900KS with 3090 Ti in games..

Why would you even bother with high end DDR4-3800 when its literally 1-2% faster than 3200 but costs way more?

I would say a better argument in favour of the 5800X3D is that it can achieve near parity with a 12900K/DDR5 combo while *only* using DDR4-3200. So the overall platform cost is much cheaper.

 

[gdansk]

I wonder how many will be available considering the supply competition from Milan X.

But I shall not FOMO myself into a rash decision. Nope. Definitely not.

 

[epsilon84]

Depends on the list of games.
The performance boost of 3D is massive, but there are games where Zen 3 massively performances for some reason. When you test the newer and stronger Intel architecture with those games even 3D can't compensate for the difference. You could call those "outliers", just like you can say that Borderlands 3 is an "outlier" for Zen3D.
Both are technically tied, it just depends on the games you want to play.

I'd agree. We need one of those 30 game comparisons that HWUB is known for to truly grasp how different games react to the cache. Right now, the results (and averages) can swing so wildly depending on the games tested.