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"

 

[igor_kavinski]

My brother. And I even made sure to get him b-die back when I assembled his computer. And when I starting talking about overclocking the ram, he waved me away and said; "Don't you do any of that strange stuff on my machine!".

Rightio...

I totally understand his reluctance. He wants stability above all else. My IT guy would get super nervous whenever I was running some benchmarks on his company issued Tiger Lake HP laptop. He pleaded with me to stop. He was afraid that I would push it beyond its limits and it would die.

 

[igor_kavinski]

Yet, if you really want to know, 720p is a good way to find out.

It's good to know for CPU benchmarking but most people are interested in the practical value of V-cache and that starts at 1080p gaming. A recent and decent GPU wouldn't get bottlenecked at that resolution.

I wonder why reviewers haven't used WARP yet:

Windows Advanced Rasterization Platform (WARP) Guide - Win32 apps | Microsoft Docs

How to Use WARP
Direct3D 10, 10.1, 11, and 12 components can use an additional driver type that you can specify when you create the device (for example, when you call the D3D11CreateDevice function). This driver type is D3D10_DRIVER_TYPE_WARP or D3D_DRIVER_TYPE_WARP. When you specify this driver type, the runtime creates a WARP device and does not initialize a hardware device.
Because WARP uses the same software interface to Direct3D as the reference rasterizer does, any Direct3D application that can support running with the reference rasterizer can be tested by using WARP. To use WARP, rename D3d10warp.dll to D3d10ref.dll and place it in the same folder as the sample or application. Next, when you switch to ref, you will see WARP rendering.
If you rename WARP to D3d10ref.dll and place it in C:\Program Files (x86)\Microsoft DirectX SDK (June 2010)\Samples\C++\Direct3D\Bin\x86, you can run all the DirectX samples against WARP, either by clicking the "Toggle Ref" button in the sample, or by running the sample with /ref specified on the command line.

With WARP, the GPU's 3D pipeline is totally removed from the equation.

 

[MadRat]

I'm curious why the multiplier is 44.5 x 100. Each stage is at 100MHz still? Why not other combinations like 19x233, 27x166, or 60x75? What is so special about the 100MHz base besides the math being easy?

 

[biostud]

While any current CPU are not going to be a bottleneck in games when played 1440p+ resolution and with all eye-candy turned on, then having a fast CPU will still make it last longer until it will be a bottleneck.

 

[zebrax2]

^PCI-E by default run at 100mhz maybe that's why.

@igor_kavinski
Running a game with modern graphics with just pure CPU power I think isn't feasible. Crysis, a game from 2007, running purely on a CPU can barely reaches 20fps on a 3990x (although it seems like its limited to 24 threads).

 

[nicalandia]

4800C40 means that the 12900KS has a bunch of headroom left. The 5800X3D, not so much. Still impressive.

Let's be honest here. Nobody that is DIY runs memory below 3600mhz. So all the tests should be between 3600-3800mhz. AMD gets big performance improvements based on ram speed. I don't like seeing 3200mhz tests for any AMD CPU.

AMD 5800X3D with 3600mhz RAM is still beating(10%) Alder Lake 12900K with 5200mhz DDR5 RAM on Shadow of the Tomb Raider at 1080P Ultra settings on an Old BIOS(not 3D V Cache tuned)

 

[Hans Gruber]

AMD 5800X3D with 3600mhz RAM is still beating Alder Lake 12900K with 5200mhz DDR5 RAM on Shadow of the Tomb Raider at 1080P Ultra settings on an Old BIOS(not 3D V Cache tuned)

My point was that AMD processors Zen 2 or Zen 3 perform best with fast ram. If run head to head vs. Intel using DDR4. AMD benefits more from fast ram than Intel. 3600mhz 14-14-14 is more than good enough for the5800x3d to perform well against Alder Lake and DDR5.

 

[Timorous]

DDR5 4800 is the Intel spec.
DDR4 3200 is the AMD spec.

Testing and spec speeds (and tighter than spec timings perhaps) sets a baseline.

Then you can compare DDR5 6400 C32 vs DDR4 4000 C16 or something to see what scales better.

 

[nicalandia]

DDR5 4800 is the Intel spec.
DDR4 3200 is the AMD spec.

In AMD Internal testing at 1080P they used DDR5 5200 and DDR4 3600 with a 3080 and the 5800X3D was 10%+

There is less than 1% difference between the regular 5800X and the 12900KS at that gaming resolution(With DDR5-6000 and DDR4-3600)
so 10% boost is quite exceptional

 

[sdifox]

AMD is investing billions in data center technologies, first with Xilinx and now with Pensando. Pretty soon, us poor gamers will be their lowest priority business segment. They will throw us scraps from their Epyc leftover rejects and we will happily gobble them up for good money.

Question - AMD aquires Pensando

https://ir.amd.com/news-events/press-releases/detail/1057/amd-expands-data-center-solutions-capabilities-with Anyone knows them ? I certainly dont.

forums.anandtech.com

Eh? Gamers have never been their target. They are just doing this chip to one up Intel marketing.

 

[Makaveli]

You've said this before and I still disagree. I think there are plenty of people that run Zen/Zen+ that run 3200, myself included. The memory controller just wasn't as good back then, so I'd be careful saying "any Zen CPU".

If you want to say anybody who has built since Zen 2, then I'd probably agree with you.

DYI and 3200 CL14 user here.

 

[Thunder 57]

I totally understand his reluctance. He wants stability above all else. My IT guy would get super nervous whenever I was running some benchmarks on his company issued Tiger Lake HP laptop. He pleaded with me to stop. He was afraid that I would push it beyond its limits and it would die.

What were you doing running benchmarks on your IT guys company laptop?

 

[Makaveli]

I totally understand his reluctance. He wants stability above all else. My IT guy would get super nervous whenever I was running some benchmarks on his company issued Tiger Lake HP laptop. He pleaded with me to stop. He was afraid that I would push it beyond its limits and it would die.

I use to be that guy in the IT department and I don't blame him if you blow up that machine that is work for him not you. He is the one that has to setup a new machine, copy all your data, reinstall apps etc. I would be pissed if I had to give up hours of my day doing that because you were running non work related benchmarks on the machine that caused the issue

 

[nicalandia]

Is it just me or many of you are pressing F5 every minute on this Cocktease page?

XanxoGaming - Portal de PC Gaming, noticias, review y más.

XanxoGaming, portal de PC GAMING, últimas noticias, reviews, guías y más. Portal especializado en partes de hardware.

xanxogaming.com

 

[igor_kavinski]

First time I'm hearing that benchmarks pose a threat to a PC's health. It wasn't anything strenuous like Prime95 or Furmark. Just regular Geekbench / Excel / PCMARK etc.

Why? Because when you are poor and have an opportunity, you SEIZE IT!

 

[Insert_Nickname]

You've said this before and I still disagree. I think there are plenty of people that run Zen/Zen+ that run 3200, myself included. The memory controller just wasn't as good back then, so I'd be careful saying "any Zen CPU".

If you want to say anybody who has built since Zen 2, then I'd probably agree with you.

You're quite right about the original Zen1 memory controller. I can push my 1700 to 2933MHz, but only with single rank Samsung (I'm not sure what die it is, haven't checked). It can sort of do 3066MHz, but it's not stable enough for daily use. Dual rank? Refuses to even boot at anything above 2400MHz.

I totally understand his reluctance. He wants stability above all else. My IT guy would get super nervous whenever I was running some benchmarks on his company issued Tiger Lake HP laptop. He pleaded with me to stop. He was afraid that I would push it beyond its limits and it would die.

Whaaat...?

While it may not be a good idea to run benchmarks on a company provided laptop, it really shouldn't be possible for it to suffer any damage. OEM/ODMs would get deep-fried if it failed after a few benchmarks. What about if someone decided to do a spot of video editing or rendering on the go? Heck, it might just happen...

 

[Makaveli]

First time I'm hearing that benchmarks pose a threat to a PC's health. It wasn't anything strenuous like Prime95 or Furmark. Just regular Geekbench / Excel / PCMARK etc.

Why? Because when you are poor and have an opportunity, you SEIZE IT!

Most of that stuff would be blocked by IT policy where I am so you wouldn't even be able to download and install it. We prevent these situations from happening before they do.

 

[Thunder 57]

First time I'm hearing that benchmarks pose a threat to a PC's health. It wasn't anything strenuous like Prime95 or Furmark. Just regular Geekbench / Excel / PCMARK etc.

Why? Because when you are poor and have an opportunity, you SEIZE IT!

You have no business doing that. At least it sounds like you did not. I'm surprised the IT guy let you do anything to their computer.

You're quite right about the original Zen1 memory controller. I can push my 1700 to 2933MHz, but only with single rank Samsung (I'm not sure what die it is, haven't checked). It can sort of do 3066MHz, but it's not stable enough for daily use. Dual rank? Refuses to even boot at anything above 2400MHz.



Whaaat...?

While it may not be a good idea to run benchmarks on a company provided laptop, it really shouldn't be possible for it to suffer any damage. OEM/ODMs would get deep-fried if it failed after a few benchmarks. What about if someone decided to do a spot of video editing or rendering on the go? Heck, it might just happen...

Yes and I think some people forget this. Zen in particular was bad, so I have heard. Zen+ stepped it up a good bit. It wasn't until Zen 2 that Zen had a pretty good memory controller.

 

[Insert_Nickname]

Most of that stuff would be blocked by IT policy where I am so you wouldn't even be able to download and install it. We prevent these situations from happening before they do.

Which is exactly what you should be doing in a secure software environment.

Smaller companies might not really have the resources do that properly however. I should know. I've dealt with plenty of those.

 

[igor_kavinski]

You have no business doing that. At least it sounds like you did not. I'm surprised the IT guy let you do anything to their computer.

Because he leans on me for advice from time to time. By the way, nice job chastising me, members of the IT Police

Fun fact: Back when we didn't have an Assistant IT manager, I used to have access to the server (a Dell VRTX). One day, the COO caught me with a lot of tabs opened on Amazon. We had just one IT team ID and for some reason (what business does a COO have logging into the server through RDP, I've no clue), he decided to login that day. He thought the server had been hacked! I told him sheepishly, oh no. Don't worry. It was I.

Long story short: my access was revoked. I was using the server to browse because I didn't want to ask for a "full internet access" exception in the firewall for my PC. Get asked too many questions on what you want to do. Luckily, through a series of events, I later got a full access IP. I access the server through the Assistant IT manager's credentials. And I still work there. I have no idea how I'm lucky enough to still be working there.

I know what y'all gonna say. We would have fired yo arse and made sure you never worked in the industry again!

 

[Thunder 57]

Because he leans on me for advice from time to time. By the way, nice job chastising me, members of the IT Police

OK so it's a work buddy. I don't mean to chastise you. It's just that people are generally overly protective of their work equipment ant letting others fiddle with it.

 

[Makaveli]

Long story short: my access was revoked. I was using the server to browse because I didn't want to ask for a "full internet access" exception in the firewall for my PC. Get asked too many questions on what you want to do. Luckily, through a series of events, I later got a full access IP. I access the server through the Assistant IT manager's credentials. And I still work there. I have no idea how I'm lucky enough to still be working there.

lol damn dude those are all red flags would be lucky to still be employed at most places I worked.

And sounds like a hackers dream. I'm doing security now and you would clearly be the target for an outside attacker

 

[igor_kavinski]

So whose gonna be buying a new mobo for the 5800X3D?

No one's recommended a mini ITX AMD mobo for me yet. I'm guessing mini-ITX mobos with four RAM slots don't exist? Anyone know the smallest micro ATX mobo with four RAM slots?

 

[nicalandia]

More news from ChanCho Gaming

AGESA 1207 optimizaría rendimiento de procesadores con AMD 3D V-CACHE

El usuario 1usmus afirma que el nuevo código AGESA 1207 optimizaría el rendimiento de procesadores AMD Ryzen usando 3D V-CACHE. Más info aquí.

xanxogaming.com

 

[Ranulf]

Considering we are talking about a Zen 3 part. I think it's fair to say that most with Zen 2 or Zen 3 CPU's run their memory at 3600mhz or higher. AMD said the sweet spot for Zen 2 memory was 3733mhz but the maximum memory speed without using memory dividers 1:1 ratios was 3800mhz. At the time, 3733mhz 3800mhz was not as stable with current motherboards. It's not the infinity fabric within the memory controller that is the problem. It's the motherboard. Most motherboards B450 and newer can run 3800mhz stable without any issues.

I think there are a good number of people who paid a kings ransom for old B-die sticks that can run 14-14-14 3200mhz all day but cannot run 3600-3800mhz at all or with significantly loosened timings. The newest B-die sticks can run 14-16-16 3800mhz all day.
I am thinking those who paid for the old B-die don't want to pony up the cash for the newer B-Die and stick with 3200mhz because they can't let go of their old B-die sticks.

Hans, bubbie... darling... its more a matter of why would I bother ponying up the cash for newer stuff for a 1700 on x370? Up until very recently they were not supported for Zen3 chips. So why bother? Maybe if I put my 3700x in there or 2700x. I suppose I will finally be motivated to test that mobo with anything besides the $200 16gb FlareX I ended up buying for it. Hell, even my x570 mobo with the 3700x seemed to be cranky with the ddr4-3000 ram I installed. I think default had it at 2666mhz.

After awhile you just lose interest in wasting more time on it. Maybe with all the parts out of cases I should take each gen cpu I have and test it in the x370, b450 and x570 boards I have.

Also, honestly, name me a good kit of 3600 speed ram or better, 32gb (16x2) that is worth the upgrade from 3200 gskill rgb stuff I have. RGB preferred since this will ultimately go into the bling box with disco lights running the 5900x.