Support NeoGAF

[Freki]

Killzone def looks tighter than Titanfall.

Is it true that Killzone is now running at 60fps?

Apples vs oranges... Wait for multiplat comparisons on retail systems with retail games.

 

[Beerman462]

Apples vs oranges... Wait for multiplat comparisons on retail systems with retail games.

Wait all you want. PS4 is more powerful and more efficient than the XB1. If developers don't take advantage then gamers lose.

 

[Piggus]

That's what I'm thinking. And on top of that, wasn't the PS3 more powerful than the 360? But most multi platform games were developed on the 360 and ported to PS3, right? So at the end of the day rarely did a PS3 multi platform game look better than the 360 version. In fact, a lot of times I believe the 360 version looked better!!

If the Xbox One gets out to a decent lead I'm sure that will be the lead platform for many third parties and thus you will not see much of a difference in multi platform games. First party games may be a different story, but there are only a handful of first party games on the PS3 that looked better than anything on the 360.

Okay, let's settle this once and for all.

There are a million people saying the same thing as you and you're all demonstrating that you're unfamiliar with the architecture of each system.

Let's start with PS3 and 360. You had two VASTLY different architectures. On 360 you had a PowerPC-based CPU with three cores paired with a relatively strong ATi (now AMD) GPU and a single 512 mb pool of system memory. It was a simple but effective setup at the time that allowed devs to quickly get a lot out of the hardware. The PS3's architecture overall was indeed more powerful, but wildly different and VERY difficult to master. The Cell processor is made up of one PowerPC core and 6 (available) "SPUs" that sort of act as separate smaller general purpose cores that could handle audio, physics, and even some graphics related tasks. This was paired with an Nvidia GPU that was a little weaker than the 360 GPU along with a split pool of memory - 256 mbs of GDDR3 (?) and 256 mbs of high-speed XDR memory. The Cell processor in theory is VERY powerful when used correctly, which is why we have games like The Last of Us and Killzone and Uncharted that are absolutely second to none in the console space. But porting games from 360 to PS3 was extremely difficult because so much of the game engines devs used had to be re-written to effectively work with Cell. The games that don't utilize Cell very well are the ones that look quite a bit worse on PS3. Got it?

Let's move on to PS4 and Xbone.

The difference here is PS4 and Xbone are almost completely identical in their architecture. They both have the same CPU, they both have a single/large pool of memory, and they both have an AMD GPU based on the exact same architecture. There are only TWO major differences in each system and both of them are a major advantage for the PS4. For one, the PS4's memory setup is significantly faster and more elegant. Both systems have the same amount of system memory but since Microsoft chose to have NO dedicated high speed memory for graphics they've had to use eSRAM as a sort of bandaid. This actually makes it MORE complex to develop for than PS4. Second is the GPU. The PS4 and Xbone GPUs, as stated, are nearly identical with the exception that the PS4 GPU simply has quite a bit more processing power and is also customized for being very good at general purpose computation tasks like physics. There is no complex difference in each system. If devs built their game on Xbone and ported to PS4 they would see a difference in performance right away. They wouldn't have to "unlock" any special ability or re-write huge parts of their game engine like they had to do this generation.

Think of it this way. An AMD Radeon 7850 video card is more powerful than a Radeon 7770. There's no way to spin it to make the opposite true. There isn't a single situation where the 7770 could outperform the 7850. This is not a situation where the PS4 is "theoretically" more powerful where devs have to jump through a bunch of hoops to make their games look better on it. It IS more powerful, and by a pretty wide margin. You will see this at launch.

 

[ace3skoot]

Apples vs oranges... Wait for multiplat comparisons on retail systems with retail games.

i disagree, the least powerful hardware will always determine the extent of how good multi-platform titles look, dev will play to the lowest common denominator because it saves money and makes porting easier.

the real difference will be seen in first party exclusives that push the hardware and have the biggest budget to do so.

 

[Piggus]

i disagree, the least powerful hardware will always determine the extent of how good multi-platform titles look, dev will play to the lowest common denominator because it saves money and makes porting easier.

the real difference will be seen in first party exclusives that push the hardware and have the biggest budget to do so.

It would be so easy to port from one or the other that it doesn't really matter which one is the lead platform. I think it's more likely that PS4 will be the lead platform in most cases simply because it's easier to develop for. All devs would really have to do is scale back some effects and program in the logic for moving data through the eSRAM when porting to the Bone.

 

[Durante]

What would you choose for a HSA APU from a software development perspective:


4 Steamroller cores with 3.2GHz + 14 GCN CUs with 800MHz

or

8 Jaguar cores with 1.6GHz + 18 GCN CUs with 800MHz

That depends on the application area (that is, which software I want to run on it), and how much I want to invest into software development. If I assume my software development budget is large and my application is really well parallelizable, I'll take the latter, otherwise the former.

Absolutely. I just spend half a day tracing down a race condition and spend the other half of the day finding the bugfix with the smallest performance impact. That race condition only occurred in less than 0,05% of all cases, and since purposefully provoking it in simulated test runs is almost impossible, testing and debugging is a b*tch.

Half a day? 0.05%? How nice. I debugged a race condition at one point which occurred on average for one in ~50 million instances of the affected operation, when using 24 threads. In the end, I figured out it was simply caused by GCC reordering two assembler operations. My own fault of course, it was perfectly within its rights to do so with the way the code was originally written.

Anyway, parallelism creates interesting jobs for qualified programmers, it's great

 

[Pop]

Okay, let's settle this once and for all.

There are a million people saying the same thing as you and you're all demonstrating that you're unfamiliar with the architecture of each system.

Let's start with PS3 and 360. You had two VASTLY different architectures. On 360 you had a PowerPC-based CPU with three cores paired with a relatively strong ATi (now AMD) GPU and a single 512 mb pool of system memory. It was a simple but effective setup at the time that allowed devs to quickly get a lot out of the hardware. The PS3's architecture overall was indeed more powerful, but wildly different and VERY difficult to master. The Cell processor is made up of one PowerPC core and 6 (available) "SPUs" that sort of act as separate smaller general purpose cores that could handle audio, physics, and even some graphics related tasks. This was paired with an Nvidia GPU that was a little weaker than the 360 GPU along with a split pool of memory - 256 mbs of GDDR3 (?) and 256 mbs of high-speed XDR memory. The Cell processor in theory is VERY powerful when used correctly, which is why we have games like The Last of Us and Killzone and Uncharted that are absolutely second to none in the console space. But porting games from 360 to PS3 was extremely difficult because so much of the game engines devs used had to be re-written to effectively work with Cell. The games that don't utilize Cell very well are the ones that look quite a bit worse on PS3. Got it?

Let's move on to PS4 and Xbone.

The difference here is PS4 and Xbone are almost completely identical in their architecture. They both have the same CPU, they both have a single/large pool of memory, and they both have an AMD GPU based on the exact same architecture. There are only TWO major differences in each system and both of them are a major advantage for the PS4. For one, the PS4's memory setup is significantly faster and more elegant. Both systems have the same amount of system memory but since Microsoft chose to have NO dedicated high speed memory for graphics they've had to use eSRAM as a sort of bandaid. This actually makes it MORE complex to develop for than PS4. Second is the GPU. The PS4 and Xbone GPUs, as stated, are nearly identical with the exception that the PS4 GPU simply has quite a bit more processing power and is also customized for being very good at general purpose computation tasks like physics. There is no complex difference in each system. If devs built their game on Xbone and ported to PS4 they would see a difference in performance right away. They wouldn't have to "unlock" any special ability or re-write huge parts of their game engine like they had to do this generation.

Think of it this way. An AMD Radeon 7850 video card is more powerful than a Radeon 7770. There's no way to spin it to make the opposite true. There isn't a single situation where the 7770 could outperform the 7850. This is not a situation where the PS4 is "theoretically" more powerful where devs have to jump through a bunch of hoops to make their games look better on it. It IS more powerful, and by a pretty wide margin. You will see this at launch.

 

[klapauzius]

Article from Germany's biggest and very reputable IT news site[/URL]:

Yes the c't magazine and the publishing house heise are reputable in the PC scene. But don't take their console articles to seriously. If someone with a little knowledge of console hardware and games read their console related articles (since years) it is obviously that there are at least one or two real Sony fanboys in the house.

In this article the author mentioned anonymous Developers who praise the PS4? Come on ... do you real take this seriously? There are well known non-anonymous sources that say that the power of Xbox One and PS4 is equal.

By the way: It seems the Wii U has its own hUMA with its complex interaction between GPU, CPU, caches, RAM and eDRAM where both processors could use the same areas?

 

[ConteZero76]

Wait all you want. PS4 is more powerful and more efficient than the XB1. If developers don't take advantage then gamers lose.

That's a misconcepiton... developers doesn't have to take advantage of anything, the whole system is itself more powerful PER SE.

 

[Beerman462]

The heart of the matter...

You summed up what I have the urge to post in everyone of these threads. I am primarily a PC gamer, but the advantages of PS4 over XB1 are so easy to understand a cave man could do it. Same CPU, more powerful GPU, faster memory and now we know more efficient memory management.

That's a misconcepiton... developers doesn't have to take advantage of anything, the whole system is itself more powerful PER SE.

I get what your saying, but after the EA BF4 thread, where they hint at a more powerful platform and say we are aiming for cross platform parity. I assume that there will be cases where the difference will be negligible, because the developer just can't be bothered. Even if it's a simple as locking the frame rate at 30fps, even though it could go higher.

 

[FranXico]

That isn't always possible. Imagine having to parse a big binary file. That is really hard to divide on multiple threads. But maybe you have to do it right now, and the rest of the program must wait. In that case you must have one strong core to do all that parsing. You could assign it to some other thread, but it wouldn't "free up" the main thread, if the continuation of the program relies on the result of that file.

You misunderstood my post. I meant having one dedicated thread that waits for a certain subset of tasks, and then executes one task at a time. A command processor of sorts, which lives separate from the main thread. But I see where you are going with your essential file example.

 

[Raytow]

Yes the c't magazine and the publishing house heise are reputable in the PC scene. But don't take their console articles to seriously. If someone with a little knowledge of console hardware and games read their console related articles (since years) it is obviously that there are at least one or two real Sony fanboys in the house.

In this article the author mentioned anonymous Developers who praise the PS4? Come on ... do you real take this seriously? There are well known non-anonymous sources that say that the power of Xbox One and PS4 is equal.

By the way: It seems the Wii U has its own hUMA with its complex interaction between GPU, CPU, caches, RAM and eDRAM where both processors could use the same areas?

Equal? really?

 

[Piggus]

You summed up what I have the urge to post in everyone of these threads. I am primarily a PC gamer, but the advantages of PS4 over XB1 are so easy to understand a cave man could do it. Same CPU, more powerful GPU, faster memory and now we know more efficient memory management.

Unfortunately I'm sure we'll continue to see plenty more "well dey said da ps3 was more powerful and look wut happened lololol" posts from people who are totally clueless. -_-

 

[ElTorro]

In the end, I figured out it was simply caused by GCC reordering two assembler operations.

That can make your day. Worst case if the race condition is influenced by target hardware (cache sizes and probability of paging, etc.) and only starts showing up on your continuos integration server.

 

[demolitio]

You summed up what I have the urge to post in everyone of these threads. I am primarily a PC gamer, but the advantages of PS4 over XB1 are so easy to understand a cave man could do it. Same CPU, more powerful GPU, faster memory and now we know more efficient memory management.


I get what your saying, but after the EA BF4 thread, where they hint at a more powerful platform and say we are aiming for cross platform parity. I assume that there will be cases where the difference will be negligible, because the developer just can't be bothered. Even if it's a simple as locking the frame rate at 30fps, even though it could go higher.

Exactly. I'm a PC gamer but also a big console gamer and the differences aren't lost on me. Will we see it in the majority of third-party games? Probably not unless they actually want to because it's easier not to. Will some games be noticeable? I think so. Will first-party games a few years from now show it as well? I think we all know that will happen given just how talented ICE is at making the most of the hardware and sharing it with all the first-party teams.

It's not a night and day difference, but it's pretty clear that the PS4 is in better shape from all we know.

 

[ElTorro]

YThere are well known non-anonymous sources that say that the power of Xbox One and PS4 is equal.

I am pretty sure that everybody with real information and knowledge on that topic is bound by NDAs and will express himself only in the most diplomatic and nonbinding way. ;-)

 

[dragonelite]

Absolutely. I just spend half a day tracing down a race condition and spend the other half of the day finding the bugfix with the smallest performance impact. That race condition only occurred in less than 0,05% of all cases, and since purposefully provoking it in simulated test runs is almost impossible, testing and debugging is a b*tch.

Its been almost an year since i last did some programming for multi threaded(cpu)..
Its all an big black hole so much horror of finding bugs.

I just started reading into win32/C++ threading API literature right now sort of preparation for next project i want to do in minor period. I hope i can at least make an nice basic job based multi threaded game engine. And a bit of preparation for XNA 2.0 next year if Microsoft doesn't fucks that up.

 

[benny_a]

Got a response from the author of the heise.de article.

Edit: Article is accurate as it is presented.

 

[ConteZero76]

You summed up what I have the urge to post in everyone of these threads. I am primarily a PC gamer, but the advantages of PS4 over XB1 are so easy to understand a cave man could do it. Same CPU, more powerful GPU, faster memory and now we know more efficient memory management.


I get what your saying, but after the EA BF4 thread, where they hint at a more powerful platform and say we are aiming for cross platform parity. I assume that there will be cases where the difference will be negligible, because the developer just can't be bothered. Even if it's a simple as locking the frame rate at 30fps, even though it could go higher.

Shelf (v)power : if they stick with parity we'll stick with other games less concerned about "parity".

 

[Perkel]

Riiiiiiight.

XPS12 is i7-3517U - that's dual core Ivy Bridge at 1.9 GHz.
Kabini A4-5000 is a 4 core Jaguar @ 1.5 GHz.
You can pretty much double the single threading result to see how PS4/XBO CPUs will compare to a 4 core Haswell. Basically - a 4 core Haswell is 6+ times faster on a single thread and ~2 times faster in multithreading than an 8 core Jaguar.

Did you see that anadtech comparison i mentioned ? i5 vs FX ? I said already that Jaguar isn't competition to Haswell in other post

 

[lherre]

A wild lherre appears!

This thread looks to be heating up!

And he's gone again...

Oh well. Good to see you're alive and lurking, lherre!

Thanks !!!

 

[ElTorro]

Got a response from the author of the heise.de article.

Edit: Article is accurate as it is presented.

Did he also clarify if the third sentence of the article was his own interpretation?

 

[driver116]

There are well known non-anonymous sources that say that the power of Xbox One and PS4 is equal.

But even at the highest level:

Sony: "We targeted the most powerful hardware"
Microsoft: "We deliberately didn't target the most powerful hardware"

 

[benny_a]

Did he also clarify if the third sentence of the article was his own interpretation?

Edit: I was mostly asking what he said about AMD's supposed denial.

 

[googleplex]

Killzone def looks tighter than Titanfall.

Is it true that Killzone is now running at 60fps?

No but i think its locked at 30fps

 

[googleplex]

Unfortunately I'm sure we'll continue to see plenty more "well dey said da ps3 was more powerful and look wut happened lololol" posts from people who are totally clueless. -_-

What happened is that the Best looking games of the generation are PS3 exclusives.

 

[Nags]

i disagree, the least powerful hardware will always determine the extent of how good multi-platform titles look, dev will play to the lowest common denominator because it saves money and makes porting easier.

the real difference will be seen in first party exclusives that push the hardware and have the biggest budget to do so.

Where is this assumption that the dev "will play to the lowest common denominator" coming from?

 

[Coulomb_Barrier]

BTW how do we know that support ticket is real. Consumer service talking about what an exec says in Germany? Seems quite odd.

If that is made up then god that is such a sad act it is actually funny.

 

[ElTorro]

By the way, the "reporting" in this topic is hilarious. I found articles that report the story by citing other websites which eventually got all their information from this thread. However, none of them was able to quote the original source. And those articles already claim that Marc Diana said that "it is a gamechanger"… which was actually written by W!ICKED in this thread.

http://www.psxextreme.com/ps4-news/614.html

Just incredible.

 

[benny_a]

By the way, the "reporting" in this topic is hilarious. I found articles that report the story by citing other websites which eventually got all their information from this thread.

In before half the article is deleted and replaced, just like Gameplay did when Microsoft PR dude said that PlayStation isn't a real eight-generation console.

At least those were actual quotes at first.

 

[cebri.one]

Got a response from the author of the heise.de article.

This should go into the OP

 

[ElTorro]

This should go into the OP

I will wait for a mods confirmation to be fair, as I also didn't include that supposed AMD support mail.

 

[Perkel]

By the way, the "reporting" in this topic is hilarious. I found articles that report the story by citing other websites which eventually got all their information from this thread. However, none of them was able to quote the original source. And those articles already claim that Marc Diana said that "it is a gamechanger"… which was actually written by W!ICKED in this thread.

http://www.psxextreme.com/ps4-news/614.html

Just incredible.

What was that kid game ? Tomato ?

 

[Soul Destroyer]

Killzone def looks tighter than Titanfall.

Is it true that Killzone is now running at 60fps?

According to N4G it is 60FPS for multiplayer.

I'm sceptical though, would wait till we hear from GG.

 

[cebri.one]

I will wait for a mods confirmation to be fair, as I also didn't include that supposed AMD support mail.

ok ;D

 

[haterofpants]

Has anyone posted this article from June where AMD supposedly told The Verge that both consoles take advantage of hUMA?

That might sound suspiciously vague, but we spoke to AMD and it's actually true. The AMD chips inside the PlayStation 4 and Xbox One take advantage of something called Heterogeneous Unified Memory Access (HUMA), "Good for gaming, good for AMD"which allows both the CPU and GPU to share the same memory pool instead of having to copy data from one before the other can use it. Diana likened it to driving to the corner store to pick up some milk, instead of driving from San Francisco to Los Angeles. It's one of AMD's proposed Heterogeneous System Architecture (HSA) techniques to make the many discrete processors in a system work in tandem to more efficiently share loads.

http://www.theverge.com/2013/6/21/4452488/amd-sparks-x86-transition-for-next-gen-game-consoles

 

[benny_a]

I will wait for a mods confirmation to be fair, as I also didn't include that supposed AMD support mail.

Yeah, I send a PM to a mod that was online. Hope he has time to confirm it.

Because then the discussion changes, because there is no confusion by the author.
If you claim it's wrong then you're basically calling c't-Editors which I'm sure you're aware of are very well respected in Germany's IT-industry of inventing stories.

 

[Daffy Duck]

By the way, the "reporting" in this topic is hilarious. I found articles that report the story by citing other websites which eventually got all their information from this thread. However, none of them was able to quote the original source. And those articles already claim that Marc Diana said that "it is a gamechanger"… which was actually written by W!ICKED in this thread.

http://www.psxextreme.com/ps4-news/614.html

Just incredible.

That's fantastic. LOL.

 

[News Bot]

By the way, the "reporting" in this topic is hilarious. I found articles that report the story by citing other websites which eventually got all their information from this thread. However, none of them was able to quote the original source. And those articles already claim that Marc Diana said that "it is a gamechanger"… which was actually written by W!ICKED in this thread.

http://www.psxextreme.com/ps4-news/614.html

Just incredible.

Gaming journalism.

 

[Billen]

Games journalism at its finest. ^_^

You are Marc Diana, evidently. This is the only logical conclusion.

 

[gofreak]

As far as I can tell from the VG Leaks documents:

1) Xbox One memory isn't physically unified, let alone logically (unified address space). The CPU cannot directly access the eSRAM. The only direct clients are the GPU and Move Engines. If the CPU wants something from eSRAM it has to put in a request to the Move Engines and wait for it to be copied out. I don't know if the DDR3 has a single address space or not.

2) Xbox One's memory does not support a coherent view of memory between GPU and CPU without cache flushes. The GPU can snoop the CPU's cache but if the GPU modifies data that the CPU is to view, caches have to be flushed.

In contrast with PS4:

1) Chris Norden apparently confirmed in his GDC talk that the memory space is a unified address space

2) GPU can modify data to be worked with by the CPU without cache flushes

Open to correction, new info etc.

 

[ElTorro]

Gaming journalism.

Headline tomorrow: AMD representative said that only Infamous Second Son is next-gen and showed impressive exclusive gif.

 

[wildfire]

Obviously Jaguar is far cheaper and more efficient than the i5. He was just dispelling the notion that an 8 core CPU is automatically faster than a 4 core CPU.

Wait. Why is everyone in these last 2 pages making AMD cores the equivalence of Intel cores?

Ever since they moved to this new architecture, cores are bundled as modules and 1 AMD module is roughly equal to 1 Intel core.

The difference between the 2 is that an AMD module has dedicated hardware when you want to do threading while Intel just tries to maximize the usage of the same core to handle 2 threads.

 

[Daffy Duck]

They talk about 3D performance, is this in-game 3D performance or 3D (the type you wear glasses) performance?

 

[benny_a]

They talk about 3D performance, is this in-game 3D performance or 3D (the type you wear glasses) performance?

Regular rendering. Nothing stereoscopic 3D specific.

 

[Daffy Duck]

That makes this article: http://www.psxextreme.com/ps4-news/614.html even funnier then.

If you're interested in all sorts of technology, and you have a special interest in 3D tech, check this out-

I have no idea what any of that means, but Diana adds that all of this is a "game changer," so Sony is going in the right direction. The only question is whether or not 3D will be an oft-used feature on the PS4; one would assume it would be, what with all technology getting better over time, but 3D-enabled games haven't exactly exploded on the PS3.

 

[Perkel]

Obviously Jaguar is far cheaper and more efficient than the i5. He was just dispelling the notion that an 8 core CPU is automatically faster than a 4 core CPU.
.

Automatically ? No. But you don't create multicore if you expect same performance as less cores.

My original point was that AMD in PC gaming market is worse not because their hardware is worse but becAuse games are currently created with single core performance where in multicore apps amd cpu wins with i7 being cheaper. NG should change that a lot same as use of gpu

 

[Daffy Duck]

So which one is wrong, the latest article saying only PS4 has it or the Verge article saying that both new consoles have it.

 

[TheGreyHulk]

Wait. Why is everyone in these last 2 pages making AMD cores the equivalence of Intel cores?

Ever since they moved to this new architecture, cores are bundled as modules and 1 AMD module is roughly equal to 1 Intel core.

The difference between the 2 is that an AMD module has dedicated hardware when you want to do threading while Intel just tries to maximize the usage of the same core to handle 2 threads.

jaguar breaks away from the"module" approach they used for their bulldozer/piledriver chips. Each jaguar core is it's own core it doesn't share an fpu between two integer cores like done on a module for bulldozer/piledriver. Each jaguar core is it's own full core. 4 cores make up their compute unit and share a block of l2 cache. Ps4/xbone have two jaguar compute units for 8 FULL cores. Do not compare these with the cores modules used on their fx chips

 

[Vestal]

This should go into the OP

the problem is there is no direct quotes on the article which obviously causes suspicion.

An article with this type of allegation should include direct quotes