AMD Ryzen 7 5800U "Cezanne" Based on "Zen 3," Geekbenched

[roberto888]

This Cezanne CPU is a joke compared to the M1. This is important information for people who want the best products they can buy.

I really don't understand where your hostility is coming from. Apple won. AMD, Intel, and Nvidia lost. Get over it.

AMD Celadon-CZN vs MacBookAir10,1 - Geekbench

browser.geekbench.com

Here is M1 running x86 based Geekbench. Not so much more powerful now, is it? Consider also that it is built on a superior node and you'll see that it's not that big of a deal. It is simply doing well in what it was designed for. No surprise there.

 

[xman2007]

AMD Celadon-CZN vs MacBookAir10,1 - Geekbench

browser.geekbench.com

Here is M1 running x86 based Geekbench. Not so much more powerful now, is it? Consider also that it is built on a superior node and you'll see that it's not that big of a deal. It is simply doing well in what it was designed for. No surprise there.

But but... Apple won apparently and amd and Intel have lost

 

[theGryphon]

This Cezanne CPU is a joke compared to the M1. This is important information for people who want the best products they can buy.

I really don't understand where your hostility is coming from. Apple won. AMD, Intel, and Nvidia lost. Get over it.

OK, let's stop making new processors because apparently race is over and Apple won. Go home people, nothing else to see here...

Are you for real??

 

[MikeMurphy]

PC users care about hardware efficiency, and Apple's M1 is the most efficient desktop chip on the market.

Enjoy your inferior technology.

The M1 is exciting and will really propel competition across all platforms.

Having said that you need to consider the entire platform as a whole before deciding what suits you best.

 

[techisfun]

Not so much more powerful now, is it?

Yes, it is. That's translated x86 code on the M1 keeping up with native x86 code on Cezanne. That's impressive.

I was going to buy a Zen 3, but I'll probably buy an M-based Mac now. AMD clearly can't compete against a company like Apple.

 

[dyonoctis]

Apple's M1 easily beats this with only 4 hp cores (~1700/7600).

Even if the scores are better when it's released, the M1 will still have more efficient cores.

The native version of Chrome scores a 207 in Speedometer 2.0 on the M1 Mac mini. Ryzen 5950X scores a 140. 10900k scores a 119.

So much for the "PC master race."

Even though I hate macOS, I'm thinking of switching to an M1 Mac. You can't buy a PC that's as efficient as an M1 Mac. The CPU and GPU are both more efficient than anything AMD, Intel, and Nvidia have to offer. Alder Lake and Cezanne won't change that. They'll also likely be going up against Apple's M2 chip since Apple probably will update their M chips annually just as the do with their A chips.

PC users are going to be raging over Apple's desktop hardware dominance for years. Apple is prioritizing efficiency over flexibility. The PC will always prioritize flexibility over efficiency.

But I bet that the cinebench scores are going to tell a slighlty different story. Geekbench scores are sometimes odd : just look at how the zen 3 core had the lead in single core for AES-XTS, but then the scaling in multicore is trough the roof for the m1, but doesn't translate for the text and image compression.

Still, I'm not going to lie, I wouldn't mind if I could replace my i5 laptop with a m1 mac, (but apple doesn't do a 15 inch macbook air, 13" is too small for my taste, but the pro 16" is too expensive).
However for my desktop, I don't see myself getting locked with the same cpu/gpu combo for years and having to buy a whole new machine once an uppgrade make sense. Going from the 1700x to the 3700x for only 360€ was so nice. (I'm doing a fair amount of 3d rendering with Arnold/maya/blender)

 

[Frick]

Yes, it is. That's translated x86 code on the M1 keeping up with native x86 code on Cezanne. That's impressive.

I was going to buy a Zen 3, but I'll probably buy an M-based Mac now. AMD clearly can't compete against a company like Apple.

So were you going to install macOS on the Zen laptop? M1 does look amazing, but it's not really comparable to a Windows laptop, good and bad, unless the software you want to use exists for the Mac. Rosetta 2 also look great, but it's basically the same deal there. We all wish MSs coming x86_64 translator for Windows ARM to be amazing, but we don't know if it is.

 

[hardcore_gamer]

M1 is good for the industry. It finally proved that ARM is a viable alternative in the laptop/desktop space after years of talk (silicon always tells the truth). Now other companies may follow suit. On the server-side, Neoverse V1 is projected to take per core and per socket crown. x86 now has competition across the board.

These ARM designs also bring their power efficiency advantage to the high-performance tier. Although the microarchitecture matters more than the ISA, it remains to be seen if x86, with its 1-15 bytes of instruction length can match the efficiency of A64's simple and clean 4-byte instruction format. Anybody with basic Verilog skills can write an A64 instruction decoder in a day. Try doing that with x86.

 

[hurakura]

Forget about that synthetic crap, which can't really be trusted, is Photoshop any faster on that M1 "wonder"? And can it run Crysis?

 

[WeeRab]

Yes, it is. That's translated x86 code on the M1 keeping up with native x86 code on Cezanne. That's impressive.

I was going to buy a Zen 3, but I'll probably buy an M-based Mac now. AMD clearly can't compete against a company like Apple.

Seems to me all that 'advanced' technology is wasted on Apple stuff.
A $300 basic x86 laptop can write documents and browse the web just as fast.

 

[Caring1]

Apple is prioritizing efficiency a walled garden over flexibility. The PC will always prioritize flexibility over efficiency.

Fixed it for you.

 

[btarunr]

PC users care about hardware efficiency, and Apple's M1 is the most efficient desktop chip on the market.

Enjoy your inferior technology.

But can it run Crysis?

inb4 emulators with half-baked D3D to Metal translation.

 

[Frick]

Forget about that synthetic crap, which can't really be trusted, is Photoshop any faster on that M1 "wonder"?

When compared to what? A 5950X runs circles around it. Here they say that the M1 runs Photoshop roughly as fast as a decently configured PC from 2017, which doesn't sound impressive at all until you realise this is the smaller chip in the family which is focused on mobility and that you can actually use it when in a laptop for content creation for eight hours and that the entire laptop weighs < 3 pounds and still badly outperforms bigger and heavierIntel based Macbook Pros.

 

[dyonoctis]

When compared to what? A 5950X runs circles around it. Here they say that the M1 runs Photoshop roughly as fast as a decently configured PC from 2017, which doesn't sound impressive at all until you realise this is the smaller chip in the family which is focused on mobility and that you can actually use it when in a laptop for content creation for eight hours and that the entire laptop weighs < 3 pounds and still badly outperforms bigger and heavierIntel based Macbook Pros.

To stay fair, those results are with roseta, the beta version of photoshop compatible with the m1doesn't offer all the functionality yet. But the thing that surprise me is how cinema 4d and octane X are already m1 compatible, but nobody cared enough to make some benchmarks with real projects.

 

[techisfun]

AMD Zen 3 = Evolutionary
Apple M1 = Revolutionary

Anybody who thinks Zen 3 is a bigger accomplishment than the M1 is an idiot. PC gamers are dismissing the M1 because they're dumbshits.

 

[Chrispy_]

AMD Zen 3 = Evolutionary
Apple M1 = Revolutionary

Anybody who thinks Zen 3 is a bigger accomplishment than the M1 is an idiot. PC gamers are dismissing the M1 because they're dumbshits.

You're intent on baiting this thread aren't you.
The M1 is impressive but if you are naive enough to think that AMD or Intel are incapable of designing something like the M1, then you are deluded and don't understand the x86 ecosystem.

Current-gen CPUs like Zen3 are designed to do a large range of things using a very large instruction set. RISC, the basis of ARM processors literally stands for 'reduced instruction set'.

Comparing ARM vs x86 is like comparing a 50-foot speedboat with a half-million-ton container ship. Yes, the speedboat is faster and more efficient but there are lots of things it simply cannot carry. If you are going to be imprisoned in Apple's walled garden and only allowed to play with things that are within the limits of their little speedboat then great, the M1 is a fantastic processor. If you need to move 1200 shipping containers halfway around the world then you're up shit creek, so to speak...

Yes it's true that a lot of the x86 code that is bogging down current PC CPUs is legacy and should be stripped out now, but it's a chicken-and-egg problem and the first person to break that cycle is going to pay dearly for the privilege with no return on the investment for years, perhaps even decades. Apple, the single wealthiest organisation on the planet may actually be the only sensible company to pull this off as they have a captive audience that will buy their stuff no matter how inconvenient the headaches new tech causes, and a closed ecosystem that reduces the scope of Apple's risk by an almost infinite amount compared to the completely open and unregulated x86 software market.

 

[dyonoctis]

You're intent on baiting this thread aren't you.
The M1 is impressive but if you are naive enough to think that AMD or Intel are incapable of designing something like the M1, then you are deluded and don't understand the x86 ecosystem.

Current-gen CPUs like Zen3 are designed to do a large range of things using a very large instruction set. RISC, the basis of ARM processors literally stands for 'reduced instruction set'.

Comparing ARM vs x86 is like comparing a 50-foot speedboat with a half-million-ton container ship. Yes, the speedboat is faster and more efficient but there are lots of things it simply cannot carry. If you are going to be imprisoned in Apple's walled garden and only allowed to play with things that are within the limits of their little speedboat then great, the M1 is a fantastic processor. If you need to move 1200 shipping containers halfway around the world then you're up shit creek, so to speak...

Yes it's true that a lot of the x86 code that is bogging down current PC CPUs is legacy and should be stripped out now, but it's a chicken-and-egg problem and the first person to break that cycle is going to pay dearly for the privilege with no return on the investment for years, perhaps even decades. Apple, the single wealthiest organisation on the planet may actually be the only sensible company to pull this off as they have a captive audience that will buy their stuff no matter how inconvenient the headaches new tech causes, and a closed ecosystem that reduces the scope of Apple's risk by an almost infinite amount compared to the completely open and unregulated x86 software market.

Lots of people don't see that Apple can do what they are doing because they are Apple. If microsoft ever dared to take out all the legacy code, the amount of bitching worlwide would be enourmous. Apple going 64bits only already broke a bunch of plug-ins that were used by professional. Apple is literally saying : "adapt or disapear, we don't care that much about compatibility".

 

[Chrispy_]

Apple is literally saying : "adapt or disapear, we don't care that much about compatibility"

Apple can afford to piss off their user base. They've been doing it for years and they're still the richest company on the planet so there's no denying that their marketing works, even if all the professionals jumped ship long ago.

 

[wgoldwing]

Hey, everyone, instead of arguing the merits of each CPU or APU, can't we just celebrate the fact that we are seeing real innovation in that area for a change, after years of Intel pushing out minimal upgrades with their Core series and AMD being pretty much in coma until their Zen series? Of course, with the "big" CPU makers taking their sweet time, Apple has been busy developing their iPhone CPU which in the end led us to the M1 chip. A smartphone today has more processing power than a supercomputer from the 90's! And what's best? An 8 core cpu does NOT cost USD 2000+!

Now, we know that AMD has been busy for years developing their Zen. Apple certainly has been busy and they have the money to invest in heavy research and technology to develop their CPUs. We can only hope that Intel is doing something behind the curtains as well and will eventually come up with something that's actually competitive. Fact is, we can get amazing tech these days at a reasonably affordable price! So Apple's M1 is better in some scenario situations? Great! Let's rejoice! We have options now! Remember when in the past we could choose between, I don't know....Intel, Intel or Intel?

 

[btarunr]

Anybody who thinks Zen 3 is a bigger accomplishment than the M1 is an idiot. PC gamers are dismissing the M1 because they're dumbshits.

In the absence of any PC game that actually runs on M1, since it's not an x86 processor to begin with, they're not.

What you're basically saying is "The Rolls Royce Trent 1000 is a revolutionary railway locomotive prime mover!"

Different architectures, incompatible applications. Literally.

 

[AxilomX]

Gang please stop feeding the troll

 

[Nihilus]

Cezanne should do very well against Tiger Lake. Using Vega is a bummer but I question if RDNA 2 would actually help in the graphics department as it will still be bandwidth limitation.

The larger cache should help make the most of the Vega gpu that is already there.

 

[Chrispy_]

Cezanne should do very well against Tiger Lake. Using Vega is a bummer but I question if RDNA 2 would actually help in the graphics department as it will still be bandwidth limitation.

The larger cache should help make the most of the Vega gpu that is already there.

I dunno, if RDNA is a GPU arch designed to use cache to compensate for lower VRAM bandwidth then it sounds exactly like the silver bullet AMD APUs need to stay ahead.

 

[R0H1T]

The RDNA(2) you're talking about uses 128MB cache, this will be nowhere near that amount. Though I do wonder if AMD will use HBM or some form of unified cache for the APUs in the future, it'll probably alleviate a lot of bandwidth issues & could be the silver bullet that brings the IGP even closer to ever moving target of mid range dGPU.

 

[ValenOne]

M1 is good for the industry. It finally proved that ARM is a viable alternative in the laptop/desktop space after years of talk (silicon always tells the truth). Now other companies may follow suit. On the server-side, Neoverse V1 is projected to take per core and per socket crown. x86 now has competition across the board.

These ARM designs also bring their power efficiency advantage to the high-performance tier. Although the microarchitecture matters more than the ISA, it remains to be seen if x86, with its 1-15 bytes of instruction length can match the efficiency of A64's simple and clean 4-byte instruction format. Anybody with basic Verilog skills can write an A64 instruction decoder in a day. Try doing that with x86.

Apple's ARM clone CPU doesn't benefit the rest of the ARM ecosystem when this CPU design is unique only for Apple. This is similar to the PowerPC situation when Apple's PowerPC machines didn't benefit the PowerPC based clone ecosystem.

AMD Celadon-CZN vs MacBookAir10,1 - Geekbench

browser.geekbench.com

Here is M1 running x86 based Geekbench. Not so much more powerful now, is it? Consider also that it is built on a superior node and you'll see that it's not that big of a deal. It is simply doing well in what it was designed for. No surprise there.

Geekbench runs better on MacOS, hence why some Zen 3 benchmarks are running with HackinTosh MacOS X.

But I bet that the cinebench scores are going to tell a slighlty different story. Geekbench scores are sometimes odd : just look at how the zen 3 core had the lead in single core for AES-XTS, but then the scaling in multicore is trough the roof for the m1, but doesn't translate for the text and image compression.

Still, I'm not going to lie, I wouldn't mind if I could replace my i5 laptop with a m1 mac, (but apple doesn't do a 15 inch macbook air, 13" is too small for my taste, but the pro 16" is too expensive).
However for my desktop, I don't see myself getting locked with the same cpu/gpu combo for years and having to buy a whole new machine once an uppgrade make sense. Going from the 1700x to the 3700x for only 360€ was so nice. (I'm doing a fair amount of 3d rendering with Arnold/maya/blender)

View attachment 177120
View attachment 177121

CZN's MT scores at 6.21 GB/s while ST scores have 5.58 GB/s is strange. AES should be able to scale with core count on CZN's side.

AES-XTS is pointless for Blender3D type workloads.