Developers Claim Xbox One Much Weaker Than PS4, We Evaluate The Evidence

rated by 0 users
This post has 6 Replies | 0 Followers

Top 10 Contributor
Posts 25,803
Points 1,164,115
Joined: Sep 2007
ForumsAdministrator
News Posted: Mon, Sep 16 2013 1:43 PM
One of the ongoing debates between the Xbox One and PS4 revolves around the expected performance difference between the two consoles -- or lack thereof. According to Edge, developers have privately stated that the PS4 is significantly faster. A console agnostic game can, according to their sources, run at 30 FPS in 1920x1080 on the PS4, but in the mid-20s on the Xbox One, even with a lower 1600x900 target resolution.

"Xbox One is weaker and its a pain to use its ESRAM," concluded one developer.

Microsoft's response to this controversy has been straightforward and expected. It told Kotaku, "Xbox One architecture is much more complex than what any single figure can convey. "

So whose telling the truth? Let's examine the broader context of the question.

The Weaker Company Always Claims The Specs Don't Count. Sometimes, They're Right

Here's a fact of marketing as old as time. The company with the spec advantage plays that card, the commpany that lacks an advantage always attempts to dodge the issue. What's notable in this case is that Microsoft doesn't actually deny the performance data.

From a game developer's perspective, the question of whether the PS4 is theoretically faster than the Xbox One is less important than whether the PS4's performance is easier to utilize. The Xbox 360 had a noted advantage over the PS3 in precisely this area -- Sony's console was much more difficult to program and the company's dev tools weren't as good. Thus, many cross-platform titles looked better on the Xbox 360, particularly in the early years, because it took less effort to customize the game to run on esoteric hardware.


Original comparison by Digital Foundry

But there's another wrinkle here, and it's every bit as important. If Game X looks slightly better on PS4 than Xbox One, but you have to put the two versions side-by-side to tell, then Microsoft is going to walk away without a problem. Modern gamers aren't stupid -- how much you like a console will depend on everything from your opinion of the UI, additional apps, streaming capabilities, brand loyalty, Kinect, and general experience with the product.

In other words, if you love Kinect, Halo, and other MS tie-ins, a better-looking Game X on PS4 won't be enough to sway you if the difference isn't egregious. If game developers can optimize a game by removing a few trees, or pushing a bit less detail to the sides of the primary environment, gamers won't notice the difference most of the time. The only problem would be if games look so poor on the Xbox One that it sends people searching for comparisons. That didn't happen last generation, and it's unlikely to happen this time around.

Microsoft's 32MB of ESRAM Isn't An Ace Card

Much has been made of the fact that Microsoft offers a high-speed ESRAM cache. The first thing to understand is that the purpose of a cache is to hide main memory access latencies and improve overall performance. That's true for all architectures, including the Xbox One's.

Despite Microsoft's attempts to claim otherwise, the 32MB of ESRAM cache is not a substitute for a high-speed memory bus and its performance cannot be estimated by adding the 68GB/s bandwidth of the SoC's DDR3 + the ESRAM cache bandwidth itself. This doesn't mean the cache can't be used to substantial benefit. Developers, however, are likely to prefer simple solutions to complex ones. That puts more pressure on Microsoft to create tools that make leveraging the Xbox One's features as easy as possible.

It's not surprising to hear that the Xbox One's cache is esoteric and difficult to use -- but it's probably not as difficult to leverage as Cell was. Having dug into Cell on multiple occasions, the chip was truly a pain to program effectively. The performance benefit of the ESRAM cache, whatever it ends up being, will almost certainly be easier to extract than Cell's full power ever was.

Performance Alone Will Not Determine The Long-Term Winner

Ultimately, whether the Xbox One is faster or slower than the PS4 is one factor among many that will determine its success. Do gamers care about performance? Absolutely. But if Microsoft can deliver games that are equally fun, that's what matters.

There are a number of tricks developers can use to optimize a scene for one architecture over another. If those optimizations make the Xbox One version noticeably worse than the PS4 version, than gamers will prefer the PS4. If they result in changes that you only notice when you pause both games side-by-side and stare at the two frames against each other, than gamers likely won't care.

I think it's fair to say that Microsoft has been fighting from a tough position this cycle, and that its pricing and messaging are largely responsible for the uphill struggle. But the final outcome, even if the PS4 really is the faster console, simply isn't assured.
  • | Post Points: 80
Top 500 Contributor
Posts 161
Points 1,100
Joined: Jul 2013

and both combined dont even equal one desktop computer by the time they both come out

  • | Post Points: 5
Top 100 Contributor
Posts 1,070
Points 11,570
Joined: Jul 2009
Joel H replied on Mon, Sep 16 2013 11:11 PM

That's not actually true.

A combined PS4 + Xbox One would be a 16-core system with 16GB of RAM -- 8GB for the video card (we'll use the PS4's memory bus for that) and 8GB of system RAM. The 8GB of system RAM would be a 68GB/s memory bus -- faster than anything but the highest-end Ivy Bridge-E system. Graphics memory, at 8GB, would be larger than anything you can buy today.

Since most PC games are still 32-bit, you'd wind up with a 3GB ceiling on those applications, which means even 8GB of RAM is overkill. True, the multi-threading issue is still present -- single-thread performance on Jaguar isn't great -- but a 16-core Jaguar should be able to hit about 6.4x in multi-threaded Cinenbench. That's comparable to Intel Core i7 quad-cores from a few years back and more than capable of driving decent gaming. Floating point performance will be excellent.

Meanwhile, the combined GPU would be roughly similar to a Radeon 7970. Clock speeds are a bit lower, but you've got the 32MB SRAM cache (thanks to Xbox One) and a good GDDR5 interface thanks to the PS4. Toss in the additional RAM, the lightweight operating systems, highly tuned architectures, and I'd wager that our combination Xbox 4 leaves the comparable PC in the dirt.

  • | Post Points: 5
Top 50 Contributor
Posts 3,093
Points 38,115
Joined: Aug 2003
Location: Texas
acarzt replied on Tue, Sep 17 2013 12:51 PM

... Except that they wont be combined and as it stands they would both be outperformed by any new PC with a Geforce GTX 760 or equivalent.

  • | Post Points: 5
Not Ranked
Posts 1
Points 20
Joined: Sep 2013
LonBeyer replied on Tue, Sep 17 2013 4:49 PM

Actually, you CAN plug the PS4 into the Xbox One - so essentially they CAN be combined.

http://thegamingprophet.blogspot.com/2013/09/xbox-one-vs-ps4-real-story.html

  • | Post Points: 20
Top 100 Contributor
Posts 983
Points 9,085
Joined: Mar 2012
Location: LA, CA
sevags replied on Tue, Sep 17 2013 11:49 PM

LOL Joel, you already thought about that is advance didn't you hehe

  • | Post Points: 5
Top 100 Contributor
Posts 1,070
Points 11,570
Joined: Jul 2009
Joel H replied on Thu, Sep 19 2013 11:55 AM

Acarzt,

Nope. Thanks to the enormous performance inefficiency baked into the Windows driver model, a GTX 760 won't outperform a console GPU. PC operating systems are tuned for compatibility and security, which directly impacts things like draw calls per second.

LonBeyer,

That's a terrible blog post. Specs and interpretation are ridiculously wrong.

  • | Post Points: 5
Page 1 of 1 (7 items) | RSS