Support NeoGAF

[strata8]

Its only 15W less than i5 4 core ivy bridge clocked at 2.3ghz.

And would perform the same in multithreaded tasks and cost far, far less to produce. You can fit 4 Jaguar cores in the same amount of silicon as 1 Ivy Bridge core.

 

[shuri]

These benchmarks are juste a waste of time and total clickbaits. It's impossible to really extrapolate accurate performance benchmarks from comparing them like this. You gotta wait for the final product to see how it reacts when its all together.

Analtech is better than this.

 

[strata8]

These benchmarks are juste a waste of time and total clickbaits. It's impossible to really extrapolate accurate performance benchmarks from comparing them like this. You gotta wait for the final product to see how it reacts when its all together.

Analtech is better than this.

It's a very misleading thread. The benchmarks are from a Kabini with 4 cores and 1.5Ghz clock speed. PS4/Xbox One will have 8 cores and (possibly) 2Ghz clock speed.

 

[KKRT00]

And would perform the same in multithreaded tasks and cost far, far less to produce. You can fit 4 Jaguar cores in the same amount of silicon as 1 Ivy Bridge core.

Yep, die size is the problem here, i agree.

BTW with Haswell they would get double the vector performance compared to Ivy Bridge, which i think would be worth die size.

 

[Momentary]

I also wanted to throw out an "lol" at the "hardware" people who threw around suggestions like "Oh ARM can compete with that" or "Project Denver blah blah";

Not much is known about project Denver, so...

 

[Projectjustice]

I upgraded to a FX-8350 8 core 4Ghz CPU this year. I over killed it. lol

 

[Stumpokapow]

While this isn't an exact dupe, I feel like the subject matter and source is identical to this thread from yesterday:
http://www.neogaf.com/forum/showthread.php?t=562297

 

[artist]

AMD delivering weak performance to gamers everywhere now.

See;

In its cost and power band, Jaguar is presently without competition. Intel’s current 32nm Saltwell Atom core is outdated, and nothing from ARM is quick enough. It’s no wonder that both Microsoft and Sony elected to use Jaguar as the base for their next-generation console SoCs, there simply isn’t a better option today. As Intel transitions to its 22nm Silvermont architecture however Jaguar will finally get some competition. For the next few months though, AMD will enjoy a position it hasn’t had in years: a CPU performance advantage.

Ridiculous. Even with that increase 8 cores @ 2 Ghz would only be ~30W. That's peanuts compared to any desktop CPU or the last generation.

Only is relative wrt to the thermal and power design.

Its only 15W less than i5 4 core ivy bridge clocked at 2.3ghz. At 35W You can have 2 cores 2 threads ivy bridge clocked at 2.9ghz.
Haswell will have 35W model that has 4 cores, 8 threads and is clocked at 2ghz.

Not sure what Ivy Bridge comparison is but the Ivy Bridge individual cores are way larger and when you include the caches the die size difference becomes huge. Add to that Intel has no viable alternative to a ~2TFlop APU in the current launch window.

Not much is known about project Denver, so...

We know this ... it's slated for 2015. Hence its not something that Nvidia could've offered to either Microsoft or Sony.

 

[something fishy]

Cool. a new gpu will be all that's needed to keep my pc able to keep playing latest stuff then.

 

[KKRT00]

It's a very misleading thread. The benchmarks are from a Kabini with 4 cores and 1.5Ghz clock speed. PS4/Xbox One will have 8 cores and (possibly) 2Ghz clock speed.

So You cannot double the test scores Yourself? Wow...
And no, consoles wont have 2ghz clocks.

 

[TheD]

It's a very misleading thread. The benchmarks are from a Kabini with 4 cores and 1.5Ghz clock speed. PS4/Xbox One will have 8 cores and (possibly) 2Ghz clock speed.

Nothing misleading about the thread at all, they are the same cores as has been rumored.
The article points out that 2Ghz is unlikely and you can get a very good idea about how fast 8 cores would be from knowing how fast 4 cores are.

 

[maabus1999]

So You cannot double the test scores Yourself? Wow...
And no, console wont have 2ghz clocks.

Performance isn't completely linear like that.

 

[Durante]

As always when looking at post-Bulldozer AMD designs, it's important to realize that AMD's "8 cores" consists of 8 INT and 4 FP units. Intel's 4-core designs have 4 INT and 4 FP units. Guess what games like to hammer on? That's right, the FP units. So really what you're saying is,

What you are saying is that you don't know what Jaguar is . It's not a module design (well, actually I believe they conceptually consider 4 Jaguar cores a "module").

It's a very misleading thread. The benchmarks are from a Kabini with 4 cores and 1.5Ghz clock speed. PS4/Xbox One will have 8 cores and (possibly) 2Ghz clock speed.

It's not like they are hiding this fact -- it's only misleading if you don't know how to read benchmark results. And if that is the case, everything is misleading.

 

[sholvaco]

Looking at Kabini, we have a good idea of the dynamic range for Jaguar on TSMC’s 28nm process: 1GHz - 2GHz. Right around 1.6GHz seems to be the sweet spot, as going to 2GHz requires a 66% increase in TDP.

Not entirely true as the author failed to account for difference in GPU speeds:

There is no telling how much TDP increased from CPU clock increase alone (anywhere north of 33% assuming voltage remains the same).

 

[X05]

No offense, but some of you guys, really don't have any clue about those stuff.

But about this stuff some really do

 

[TheD]

Performance isn't completely linear like that.

It is less than linear.

 

[KKRT00]

Performance isn't completely linear like that.

It isnt, but its not for everything, thats why You check multithreaded benchmark divide it by cores and then fit to Your core+clock setup, its easy and quite accurate comparison method for theoretical maximum performance.

For something like i3, which has 2 cores and 4 threads, You use 4 cores clocked at half-speed in calculations.

 

[shandy706]

also their release date....

I'm guessing the X1 comes out in October. Just a guess...we'll know in a few weeks.

 

[Unknown Soldier]

What you are saying is that you don't know what Jaguar is . It's not a module design (well, actually I believe they conceptually consider 4 Jaguar cores a "module").

Yeah, people corrected me on this earlier.

 

[chris0701]

So PS4/Xbone 8 core CPU = The low-end Ultrabook CPU we had in 2012.

http://community.futuremark.com/hardware/cpu
i5-3317u :2460
i3-3220 : 4100 (~100 USD price)

 

[strata8]

Nothing misleading about the thread at all, they are the same cores as has been rumored.
The article points out that 2Ghz is unlikely and you can get a very good idea about how fast 8 cores would be from knowing how fast 4 cores are.

It's still misleading, since the title said 'performance analysis on the CPU powering the PS4...' and then chucked a benchmark there without clarifying anything.

Case in point:

So PS4/Xbone 8 core CPU = The low-end Ultrabook CPU we had in 2012.

http://community.futuremark.com/hardware/cpu
i5-3317u :2460
i3-3220 : 4100 (~100 USD price)

 

[Shayan]

It's a very misleading thread. The benchmarks are from a Kabini with 4 cores and 1.5Ghz clock speed. PS4/Xbox One will have 8 cores and (possibly) 2Ghz clock speed.

exactly and also A4 was a much lower end first gen APU

Xone/PS4 have custom built top of the line APUs. The article is totally misleading

So PS4/Xbone 8 core CPU = The low-end Ultrabook CPU we had in 2012.

http://community.futuremark.com/hardware/cpu
i5-3317u :2460
i3-3220 : 4100 (~100 USD price)

AMD themselves said that the APU for PS4/Xone are the most powerful that they have ever built . It is not fully based on A series architecture

 

[strata8]

exactly and also A4 was a much lower end first gen APU

Xone/PS4 have custom built top of the line APUs. The article is totally misleading

Well, not exactly. The A4-5000 is a Jaguar CPU. It's just low end compared to the APU in the PS4/XONE.

AMD themselves said that the APU for PS4/Xone are the most powerful that they have ever built .

That's not saying much Their top end APU (A10-5800K), gets 3 points in Cinebench vs. possibly 3.2-4 for the PS4 CPU. But that's mostly because their top end APU sucks.

 

[aeolist]

These benchmarks are juste a waste of time and total clickbaits. It's impossible to really extrapolate accurate performance benchmarks from comparing them like this. You gotta wait for the final product to see how it reacts when its all together.

Analtech is better than this.

there's two articles here, one is an architecture deep-dive with no hard performance numbers but some general comparisons and conclusions that can be drawn

the other is a quantitative benchmark of the first jaguar-based silicon amd has sent to reviewers, it has nothing to do with the consoles beyond a loose comparison of int/fp single and multithreaded performance numbers

anandtech doesn't really care much about consoles, they're pc/mobile focused most of the time and you're not going to see them get drawn into brand wars

 

[cRIPticon]

Hehehe... such crap CPUs. Pity really.... missed chance for both of them to be really next gen.

What a silly post. How many times does it need to be mentioned that you can not examine a system by any one single component but need to look at the system in the aggregate to determine its performance spectrum. And, these are not meant to replace high performing PCs, these are closed box, mass consumption devices.

 

[TheD]

A slow CPU is a slow CPU, anything else in the system can not really help with that unless the code is massively parallel and thus map to the GPU well, but then you are taking away GPU resources for rendering.

 

[c0de]

I am still wondering about the main-chip... It contains the APU and ESRAM for sure but it should also contain the "mysterious" move-engines as I can't see any chip which is "near" enought to CPU and RAM. So perhaps this is why MS talked 5 bn transistors.

 

[MadScientist]

I wonder what the typical technical differences we can expect between multiplatform games next gen.

While current gen the PS3 is arguably the more powerful console, its special architecture has prevented a noticable lead of PS3 versions over their Xbox counterparts. Many times the XBox versions even showed a small lead.

But next gen architectures seem to be pretty similar, with the PS4 being significantly more poerful than the XBox One. Consequently I expect PS4 versions of games to show a technical lead more often. Do you agree?

I also wonder where current high end PCs will end up. Their main deficit is their lack of a comparable amount of DDR5 RAM compared to the PS4. How will that show?

 

[Iced_Eagle]

I wonder what the typical technical differences we can expect between multiplatform games next gen.

While current gen the PS3 is arguably the more powerful console, its special architecture has prevented a noticable lead of PS3 versions over their Xbox counterparts. Many times the XBox versions even showed a small lead.

But next gen architectures seem to be pretty similar, with the PS4 being significantly more poerful than the XBox One. Consequently I expect PS4 versions of games to show a technical lead more often. Do you agree?

I also wonder where current high end PCs will end up. Their main deficit is their lack of a comparable amount of DDR5 RAM compared to the PS4. How will that show?

I think the differences are going to be minimal honestly. The easiest thing for a developer to tweak is the resolution. That's something that is really hard to tell though when you're playing a game at home.

For example, I think Halo 4 looks gorgeous, but it's still running natively at 720p and it's upscaled to fit my 1080p TV. I think if the PS4 does have this noticeable advantage, it could be something like PS4 version is at 1080p and Xbox One runs at 900p. Unless you're digital foundry though, you aren't likely to notice.

Another thing we can't forget to mention is that some games are also inherently CPU-bound, not GPU-bound. In these cases, from what I can tell the PS4 won't have the edge in this scenario, but will be essentially tied with the Xbox One since they have the same CPU. It doesn't matter how much faster your GPU is when your limiting factor is the CPU. Maybe the custom move engines and SHAPE audio hardware will allow CPU-bound games to run even faster on the Xbox One even. The proof will be in the games of course.

 

[Respawn]

people forget the ESRAM, didn't they say this will give them 200 GB/s?

200 GB/s > 176 GB/s

Please understand technology before posting such tomfoolery

 

[James Sawyer Ford]

I think the differences are going to be minimal honestly. The easiest thing for a developer to tweak is the resolution. That's something that is really hard to tell though when you're playing a game at home.

For example, I think Halo 4 looks gorgeous, but it's still running natively at 720p and it's upscaled to fit my 1080p TV. I think if the PS4 does have this noticeable advantage, it could be something like PS4 version is at 1080p and Xbox One runs at 900p. Unless you're digital foundry though, you aren't likely to notice.

Another thing we can't forget to mention is that some games are also inherently CPU-bound, not GPU-bound. In these cases, from what I can tell the PS4 won't have the edge in this scenario, but will be essentially tied with the Xbox One since they have the same CPU. It doesn't matter how much faster your GPU is when your limiting factor is the CPU. Maybe the custom move engines and SHAPE audio hardware will allow CPU-bound games to run even faster on the Xbox One even. The proof will be in the games of course.

You're going to be able to notice an increase in resolution, AA, texture resolution, effects, and possibly framerate stability on the PS4 without much tweaking. Digital Foundry won't be the only ones able to notice a difference, since these differences should be far more apparent than the PS3/360 comparisons of this gen.

Not to mention, first party exclusives are where you will truly see a rather large divide, since the core game engine will be built from the ground up for a system that has substantially more power under the hood.

 

[nelsonroyale]

The difference between PS4 and Xbone does seem to be significantly greater than between the 360 and PS4 though, so difference in multiplats games should definitely be more obvious than this gen.

 

[Biggzy]

The difference between PS4 and Xbone does seem to be significantly greater than between the 360 and PS4 though, so difference in multiplats games should definitely be more obvious than this gen.

Most definitely. The only question is how much of a difference.