• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

DigitalFoundry: X1 memory performance improved for production console/ESRAM 192 GB/s)

Status
Not open for further replies.

ekim

Member
Nothing in the article points to the assumption from some people here, that bandwidths were added. So stop with that "but you can't just add bandwidth" stuff.
http://www.eurogamer.net/articles/digitalfoundry-xbox-one-memory-better-in-production-hardware

Well-placed development sources have told Digital Foundry that the ESRAM embedded memory within the Xbox One processor is considerably more capable than Microsoft envisaged during pre-production of the console,
with data throughput levels up to 88 per cent higher in the final hardware.

Bandwidth is at a premium in the Xbox One owing to the slower DDR3 memory employed in the console, which does not compare favourably to the 8GB unified pool of GDDR5 in the PlayStation 4.
The 32MB of "embedded static RAM" within the Xbox One processor aims to make up the difference, and was previously thought to sustain a peak theoretical throughput of 102GB/s -
useful, but still some way behind the 176GB/s found in PlayStation 4's RAM set-up.
Now that close-to-final silicon is available, Microsoft has revised its own figures upwards significantly, telling developers that 192GB/s is now theoretically possible.

Catching up to the GDDR5 bandwidth.

So how could Microsoft's own internal tech teams have underestimated the capabilities of its own hardware by such a wide margin?
Well, according to sources who have been briefed by Microsoft, the original bandwidth claim derives from a pretty basic calculation -
128 bytes per block multiplied by the GPU speed of 800MHz offers up the previous max throughput of 102.4GB/s. It's believed that this calculation remains true for separate read/write operations from and to the ESRAM.
However, with near-final production silicon, Microsoft techs have found that the hardware is capable of reading and writing simultaneously.
Apparently, there are spare processing cycle "holes" that can be utilised for additional operations.
Theoretical peak performance is one thing, but in real-life scenarios it's believed that 133GB/s throughput has been achieved with alpha transparency blending operations (FP16 x4).

more or less debunking the downclock rumor:
The news doesn't quite square with previous rumours suggesting that fabrication issues with the ESRAM component of the Xbox One processor had actually resulted in a downclock for the GPU, reducing its overall capabilities and widening the gulf between graphical components of the Xbox One and the PlayStation 4. While none of our sources are privy to any production woes Microsoft may or may not be experiencing with its processor, they are making actual Xbox One titles and have not been informed of any hit to performance brought on by production challenges. To the best of their knowledge, 800MHz remains the clock speed of the graphics component of the processor, and the main CPU is operating at the target 1.6GHz. In both respects, this represents parity with the PlayStation 4.

Multiplatform performance / tools:
In terms of what this all means with regards multi-platform titles launching on both next-gen consoles,
our information suggests that developers may be playing things rather conservatively for launch titles while dev tools are still being worked on.
This is apparently more of an issue with Xbox One, where Microsoft developers are still in the process of bringing home very significant increases in performance from one release of the XDK development environment to the next.
Our principal source suggests that performance targets are being set by game-makers and that the drivers should catch up with those targets sooner rather than later.
Bearing in mind the stuttering performance we saw from some Xbox One titles at E3 such as Crytek's Ryse (amongst others), this is clearly good news.

As the performance levels of both next-gen consoles are something of a moving target at the moment, differences in multi-platform games may not become evident until developers are working with more mature tools and libraries.
At that point it's possible that we may see ambitious titles operating at a lower resolution on Xbox One compared to the PlayStation 4.
 

twobear

sputum-flecked apoplexy
Wait what, is this even possible? MS's hardware designers didn't even know the capabilities of something they designed? I smell bullshit.
 

LiquidMetal14

hide your water-based mammals
After listening to Mark Cerny's talk, I believe this route is not the best way. The number looks higher but it creates added complication. I do like the clock speeds staying the same though.
 

RibMan

Member
godelsmetric said:
Wait what, is this even possible? MS's hardware designers didn't even know the capabilities of something they designed? I smell bullshit.

Microsoft has revised its own figures upwards significantly, telling developers that 192GB/s is now theoretically possible.

The information about the performance improvement is coming from Microsoft. Keep that in mind.
 

Zukuu

Banned
I call BS. Hell, they even throw in another number in the same article:
Theoretical peak performance is one thing, but in real-life scenarios it's believed that 133GB/s throughput has been achieved with alpha transparency blending operations (FP16 x4).
 

Putty

Member
"telling developers that 192GB/s is now theoretically possible."

While theoretically jumping through hoops.

Very odd they only just realised. Even so, the 133 figure mentioned still falls some way short of Sony's figure, without jumping through hoops.
 

RoboPlato

I'd be in the dick
The information about the performance improvement is coming from Microsoft. Keep that in mind.
True that it is only a theoretical number and it comes from Microsoft but it is still a pretty big leap. The console will still have lower average bandwidth between the two pools than PS4 but this will help a lot.

So the XBO is more capable than the PS4 if this article is true?

Would be interesting to see.
No, just that the eSRAM pool could have more bandwidth but there is a lot less of it than the GDDR5 in PS4. PS4 still has the better GPU and more fast RAM available to games.
 
Well, it is good if their yields are higher, but why are they still adding up the two bandwidths...? It is not exactly comparable to what is going on in the PS4 if I have it right.

If the info is from MS directly, then I am probably more sceptical. Considering their recent activities they have very misleading and wishy-washy on the info front.
 

GetemMa

Member
yeah not much difference on multiplats.

It's the exclusives where the PS4's extra CU's will provide extra punch.
 
So the XBO is more capable than the PS4 if this article is true?

Would be interesting to see.

Not really. No matter what numbers or components people throw out here and on other sites, the two consoles are basically going to be near parity in the real world.

Which is a good thing save for the console warriors.
 
Are they simply adding HDD read/write speed with it? How can they increase bandwidth without changing the bit-interface or clock speed?

What is the read-write speed of the HDD?
 
However, with near-final production silicon, Microsoft techs have found that the hardware is capable of reading and writing simultaneously. Apparently, there are spare processing cycle "holes" that can be utilised for additional operations. Theoretical peak performance is one thing, but in real-life scenarios it's believed that 133GB/s throughput has been achieved with alpha transparency blending operations (FP16 x4).


So which is it?
 

Sorral

Member
Theoretical peak performance is one thing, but in real-life scenarios it's believed that 133GB/s throughput has been achieved with alpha transparency blending operations (FP16 x4).

Isn't that the real jump from 102.4GB/s to 133GB/s on ESRAM since it is not a simple combine to 192GB/s?

Weird that they didn't know from the start...
 

LiquidMetal14

hide your water-based mammals
It's the same bad math that's been used before. You can't just add the speeds of both. That's not how bottlenecks work.

That's what I was thinking. Theoretical numbers are one thing but that's not some number that will be met all the time on Xbone. This is a little misleading since you still have the bottleneck of using DDR3 in conjunction with 32MB od eDRAM.
 
look. Stop trying to downplay this people. It only benefits both consoles if the gap in power is closer.


That said, a more complicated setup :- Only first party will take advantage. That is undeniable.
 

GetemMa

Member
So the XBO is more capable than the PS4 if this article is true?

Would be interesting to see.

not really. The effective bandwidth appears to be around 133 gb/sec and the PS4 Memory bandwidth is still significantly higher plus the PS4 GPU has more CU's.
 

derFeef

Member
The GPU is still much more powerful in the PS4 though - at least that's what GAF tells me.

GDDR5 is good for graphics, less good for CPU operations due to the higher latency.
DDR3 is good for the CPU but worse for the graphics due to lower bandwith. ESRAM is there to make up for that, at least a little bit.
 

EMT0

Banned
This is still more inefficient than using GDDR5 alone and more like patching a poor decision than anything else. The GPU is also weaker than the PS4s by a noticeable margin.

The games will look the same? okay
 
So basically the so called power gap is even smaller. heck i thought the xbox cpu was at 1.2 but it is at 1.6. So basically multiplats will look the same and run smoothly on both systems it will take the exclusives to show a difference if there is one.
 

Krakn3Dfx

Member
We've been hearing about yield issues from reliable sources for months and suddenly MS is not only saying that's not true, but that they're getting even high quality yields than previously stated?

Oh yeah, this is from Microsoft.
 

ekim

Member
GDDR5 is good for graphics, less good for CPU operations due to the higher latency.
DDR3 is good for the CPU but worse for the graphics due to lower bandwith. ESRAM is there to make up for that, at least a little bit.

Not a concern for OOO CPUs which are used here.
 

SEGAvangelist

Gold Member
GDDR5 is good for graphics, less good for CPU operations due to the higher latency.
DDR3 is good for the CPU but worse for the graphics due to lower bandwith. ESRAM is there to make up for that, at least a little bit.

So Xbox One was designed for bigger worlds and PS4 for better looking worlds. 3rd parties will still make it the same for both, though.
 
Status
Not open for further replies.
Top Bottom