How important is system bandwidth

[citizen X]

: how important is system bandwidth



I know how important system bandwidth is, but I wan’t to know what u guys think, because some people don’t see that the ps3’s 48.8 GB/s is a big problem, but it is. The Xbox 360 has 5x the system bandwidth, 278.4 GB/s



: PS3's bandwidth VS Xbox 360's bandwidth



What is system bandwidth?

To put it simply, it’s like a pipeline that connects all the parts together. Think of the data as the water flowing through the pipeline.

Xbox360
The xbox360’s pipeline is, let’s say ….. 5 feet in diameter. This miens a lot of water is flowing through the pipeline at any point in time. The xbox360 has a system to make sure that the pipeline is always full of running data, that means, NO BOTTLENECKS.

ps3
The ps3’s pipeline is 1 foot in diameter because its 5x smaller. This means that there is 5x less data going through the pipeline at its best. The ps3’s design is full of bottlenecks.


PS3’s designed bottlenecks
1. The cell’s 7 SPEs need to run in order, they can’t run simoultaniously.

2. the pipeline itself, its to small

3. No way to manage a continuous flow of data through the pipeline to prevent bottlenecks.

4. The pipeline can’t be shared by the CPU and GPU at the same time, its one or the other. To me this sounds questionable, so if u have any information on this share it.


: why


because
I’m tired of people on the forum who just don’t want to talk about this big problem. They keep on saying that it’s going to be a tie. NOT IF THE PS3 ONLY HAS 48.8 GB/s.

The xbox360 uses more system bandwidth on the shading alone, then what the ps3 has all together. That means that the xbox360 uses some were Around 50 GB/s just for shading

I think that the xbox360 has a better chance of running KILL ZONE 2, at least better then the ps3

If u think I’m wrong on any of the information I shared, tell me. I know u were going to anyway






ADD ON

The CELL processor is a very powerful CPU; don’t get me wrong, I just think that this is the wrong application for the kind of processing its going to be used for. And that Sony didn’t give the new Cell technology a chance to show what it’s capable of, buy basically keeping it confined to the bandwidth barrier. Sure they can compress the dated but to what point? Compression is not the answer; compression is what u do as a last resort, not as plane B, that is meant to fix the known problem with plan A from the start.

[Grumpy Old Fart]

That's a great post cheers citizen X, and helps a non techie person like myself out.

[citizen X]

I’m glad to help but to be honest I just finished it now at 3:35. AM

and God I’m tired!!!!!!

And yes I have a life

[Grumpy Old Fart]

Originally Posted by citizen X

I’m glad to help but to be honest I just finished it now at 3:35.

God I’m tired!!!!!!

And yes I have a life

I certainly appreciate the input you tech guys are putting in on the forum. I'm certainly learning alot and really helps to put 'arguments' across on other sites regarding the next gen consoles.

oh and get to bed

[Justin Buser]

Nice article citizen X - keep an eye on the front page over the next few days

[pApA SmerF01]

Originally Posted by citizen X

: how important is system bandwidth



I know how important system bandwidth is, but I won’t to know what u guys think, because some people don’t see that the ps3’s 48.8 GB/s is a big problem, but it is. The Xbox 360 has 5x the system bandwidth, 278.4 GB/s



: PS3's bandwidth VS Xbox 360's bandwidth



What is system bandwidth?
Fist of all, if u don’t know what system bandwidth is, u don’t belong in the hardware forums .
To put it simply, it’s like a pipeline that connects all the parts together. Think of the data as the water flowing through the pipeline.

Xbox360
The xbox360’s pipeline is, let’s say ….. 5 feet in diameter. This miens a lot of water is flowing through the pipeline at any point and time. The xbox360 has a system to make shore that the pipeline is always full of running data, that miens NO BOTTLENECKS.

ps3
The ps3’s pipeline is 1 foot in diameter because its 5x smaller. This means that there is 5x less data going through the pipeline at its best. The ps3’s design is full of bottlenecks.


PS3’s designed bottlenecks
1. The cell’s 7 SPEs need to run in order, they can’t run at the same time.

2. the pipeline itself, its to small

3. No way to manage a continuous flow of data through the pipeline to prevent bottlenecks.

4. The pipeline can’t be shared by the CPU and GPU at the same time, its one or the other. To me this sounds questionable, so if u have any information on this share it.


: why


because
I’m tired of people on the forum who just don’t want to talk about this big problem. They keep on saying that it’s going to be a tie. NOT IF THE PS3 ONLY HAS 48.8 GB/s.

The xbox360 uses more system bandwidth on the shading alone, then what the ps3 has all together. That miens that the xbox360 uses some were Around 50 GB/s just for shading

I think that the xbox360 has a better chance of running KILL ZONE 2, better then the ps3

If u think I’m wrong on any of the information I shared, tell me. I know u were going to anyway

also an interesting thing to point out is, that the 48.8 gps is ps3's total system bandwidth. that's for data rendering between the spe's, signaling between RAM, shader ops, die-form-routing, bit-rendering; it simply is too much to do with so little.

that's a really good analogy for bandwidth, but here's another one.

the 360's bandwidth is like a boeing 747. lots of passengers in for the trip; tons of room.

the ps3's bandwidth is like a cessna (not sure on spelling) trying to fit in all those people for a rocky trip over the grand canyon....

[pApA SmerF01]

you might like this read on the ATI xenos c1 (aka the r500)

it's really techie, and i'm just going to post the bandwidth section, but if you want the full link, pm me.

When creating a high performance computing platform bandwidth between components and operations is highly important, especially when creating a system that has to last for 3-5 years before a new version comes about, such is the world of consoles. With the Xenos processor being both a high performance graphics processing element of the XBOX 360 as well as the "Northbridge" component of the system, which is essentially the communication hub for the other components of the system, it has many interconnects and bandwidths to deal with. Below is a diagram highlighting the connection bandwidths between the most important elements it is connected to:

see picture 1

XBOX 360 Primary System Bandwidths

As we discussed earlier, the XBOX 360 carries a unified memory architecture and Xenos's parent die is acting as the Northbridge controller as well as the graphics processing device. The system memory bandwidth is 22.4GB/s courtesy of the 128-bit GDDR3 memory interface running at 700MHz. At 232M transistors the Xenos parent die isn't an enormous chip so internal memory communication isn't going to be too latency bound, hence the memory interface only needs to be a standard crossbar, which is partitioned into two 64-bit blocks. Xenos's parent die also has a 32GB/s connection to the daughter, eDRAM die Connection to the Southbridge audio and I/O controller is achieved via two PCI Express lanes which results in 500MB/s of both upstream and downstream bandwidth.

As the CPU is going to be using Xenos to handle all its memory transfers, the connection between the two has 10.8GB/s of bandwidth both upstream and downstream simultaneously. Additionally the Xenos graphics processor is able to directly lock the cache of the CPU in order to retrieve data directly from it without it having to go to system memory beforehand. The purpose of this is that one (or more, if wanted) of the three CPU cores could be generating very high levels of geometry that the developer doesn't want to, or can't, preserve in the memory footprints available on the system when in use. High-resolution dynamic geometry such as grass, leaves, hair, particles, water droplets and explosion effects are all examples of one type of scenario that the cache locking may be used in.

see picture 2

ROP Bandwidth to eDRAM

The one key area of bandwidth, that has caused a fair quantity of controversy in its inclusion of specifications, is that of bandwidth available from the ROPS to the eDRAM, which stands at 256GB/s. The eDRAM is always going to be the primary location for any of the bandwidth intensive frame buffer operations and so it is specifically designed to remove the frame buffer memory bandwidth bottleneck - additionally, Z and colour access patterns tend not to be particularly optimal for traditional DRAM controllers where they are frequent read/write penalties, so by placing all of these operations in the eDRAM daughter die, aside from the system calls, this leaves the system memory bus free for texture and vertex data fetches which are both read only and are therefore highly efficient. Of course, with 10MB of frame buffer space available this isn't sufficient to fit the entire frame buffer in with 4x FSAA enabled at High Definition resolutions and we'll cover how this is handled later in the article.

Both XBOX 360 and Playstation 3 feature UMA and graphics busses, respectively, that have been announced to use fairly fast 700MHz GDDR3 memory, but both only have a 128-bit interface. Whilst this is less of a surprise for XBOX 360 as Xenos's use of eDRAM will move the vast majority of the frame buffer bandwidth to the EDRAM interface leaving the system memory bandwidth available primarily for texturing bandwidth. It does seem odd that by the time the consoles will be released the likelihood is that high end PC graphics will using at least the same speed RAM but on double wide busses. The primary issue here is, again, one of cost - the lifetimes of a console will be much greater than that of PC graphics and process shrinks are used to reduce the costs of the internal components; 256-bit busses may actually prevent process shrinks beyond a certain level as with the number of pins required to support busses this width could quickly become pad limited as the die size is reduced. 128-bit busses result in far fewer pins than 256-bit busses, thus allowing the chip to shrink to smaller die sizes before becoming pad limited - by this point it is also likely that Xenos's daughter die will have been integrated into the shader core, further reducing the number of pins that are required.

[Grindstone]

Well, okay, we've all heard about the system bandwidth problem and how the PS3 architecture "bottlenecks" the performance. My question is, how much bandwidth do we need? We've all heard the PS3 fanboi arguments about terabytes, but the reality is, we don't need 2.18 terabytes for the type of gaming we'll be doing. So, comparatively speaking, do we need that much bandwidth for gaming?

I basically know part of the answer (more bandwidth = no bottlenecking), but I'd like to solidify my answer, and I am sure others would like to know the answer to this question as well.

[citizen X]

i did look for it on the forums

[pApA SmerF01]

Originally Posted by Grindstone

Well, okay, we've all heard about the system bandwidth problem and how the PS3 architecture "bottlenecks" the performance. My question is, how much bandwidth do we need? We've all heard the PS3 fanboi arguments about terabytes, but the reality is, we don't need 2.18 terabytes for the type of gaming we'll be doing. So, comparatively speaking, do we need that much bandwidth for gaming?

I basically know part of the answer (more bandwidth = no bottlenecking), but I'd like to solidify my answer, and I am sure others would like to know the answer to this question as well.

just a quick correction, they're bragging about 2.18 teraflops.

which is basically the amount of floating-point operations the system does per clock cycle.

the only problem is that the 2.18 is what it "can" do running at 100% efficiency. but considering almost 70% of their flop count comes from the gpu, and it's only 55-70% efficient, my equations have given me the more realistic number of 1.543, .17 latent for error-correction efficiency.

with the 360, they underestimated their flop count, and with my calculations (seeing how with usa their gpu is 95-98% efficient) their actual Tflop count is 1.413,with .17 latent for error-correction efficiency.

you also have to look at this. floating point operations only count for about 20% of total system operations. the bulk of it is general and integer operations, which the xbox360 excels at.

at last calculation, the 360 is 3x more potent in general operations than the ps3 and almost 2x more efficient in its integer operations.

next time you come across a ps fanboy bragging about TFLOPS just repost this, lol.

[Grindstone]

Dangit, I meant teraflops, not terabytes (maybe next generation console, eh?). And yes, that is why I said that the supposed 2.18 teraflops were not necessary. Thanks for your breakdown, although could you explain a little more on the loss of teraflops for the two systems. Seems like you need to make a thread about this. We'll stick it to PS3 on multiple levels: teraflops, bandwidth, HD drive, etc...

Information is power.

[citizen X]

Originally Posted by Grindstone

Dangit, I meant teraflops, not terabytes (maybe next generation console, eh?). And yes, that is why I said that the supposed 2.18 teraflops were not necessary. Thanks for your breakdown, although could you explain a little more on the loss of teraflops for the two systems. Seems like you need to make a thread about this. We'll stick it to PS3 on multiple levels: teraflops, bandwidth, HD drive, etc...

Information is power.

They counted the teraflops if the consol is running at 100%. The PS3’s system architecture will never allow for full 100% efficiency, it’ll only run at around 50 to 70 % tops. So that means that the teraflop count posted is over exaggerated. Like I said before and I’ll say it again, PS3 is all hype. Or just 99.9%

[Man Whore]

Okay lets see here. Okay the GPU is the A** and the CELL is the toliet bowl. So when the plumbers (Sony) laid the pipes in the house one of the plumbers hooked up 1inch pvc pipe to the toliet. So now when the GPU take a S*** it going to get clog up in the toliet.

[Grindstone]

Uh, thank you Man Whore, for that forensic explanation and permanent image engrained in my head.

And thanks to citizen X, for skipping all toilet explanations and just providing the basic inforamtion that I needed.

Now if you'll excuse me, I'm heading to the bathroom for some experiments on "teraflops" or is it tera-plops...?

[zCraven15]

i think ps3 is gonna be screwed when it comes to memory bandwidth. the memory isnt just for gaming its for videos, mp3s, all sorts of xbox live downloads. neither a game or a console is made up of all graphics like ps3 plans to make their console. the 360 is gonna win the war because of the general hardware.

[shaun]

I'm not trying to be flamebait here, but I saw a couple of mistakes in your original post --

One: 7SPE's . SPE stands for 'Synergystic Processing Element' The PS3 processor has 7 of these. Each SPE has an SPU, and each SPU is capable of eight floating point instructions per clock cycle. Source, also the Cell processor is capable of SMT, such as the Pentium 4 or IBM POWER5.

The core design of the cell allows for multiple SPE's to process at once.

2. Total system processing power. Great, the xbox has more system bandwidth, probably due largely to the integrated on die memory for the video chipset. While this is a welcome performance gain, the CELL processor is capable of assisting the RSX with graphics (Note that all PS3 rendered images at E3 were running off another cell processor, not RSX)

And having said that, we've all seen taht the totall system processing power (including GPU)

Microsoft XBox360 - 1 Teraflop
Sony PS3 - 2 Teraflops

That can account for a prettty hefty chunk of graphics processing.

However, the XBox360's GPU is unique, and I do believe that the 256gb/sec bandwidth will be a welcome increase, and largely due to the on die memory. With this setup 16 or 32 megabytes would be equivalent to a 256 or 512mb card nowadays.

I realize this is an Xbox forum, but if you're going to educate people, at least do it correctly. The xbox definately has its strengths, but know your enemy before yuo bash it.

I'm not going to sit here and argue aimlessly with xbox fanboys or ps3 fanboys, i just found it necessary to add my two cents.

[pApA SmerF01]

Originally Posted by shaun

I'm not trying to be flamebait here, but I saw a couple of mistakes in your original post --

One: 7SPE's . SPE stands for 'Synergystic Processing Element' The PS3 processor has 7 of these. Each SPE has an SPU, and each SPU is capable of eight floating point instructions per clock cycle. Source, also the Cell processor is capable of SMT, such as the Pentium 4 or IBM POWER5.

The core design of the cell allows for multiple SPE's to process at once.

2. Total system processing power. Great, the xbox has more system bandwidth, probably due largely to the integrated on die memory for the video chipset. While this is a welcome performance gain, the CELL processor is capable of assisting the RSX with graphics (Note that all PS3 rendered images at E3 were running off another cell processor, not RSX)

And having said that, we've all seen taht the totall system processing power (including GPU)

Microsoft XBox360 - 1 Teraflop
Sony PS3 - 2 Teraflops

That can account for a prettty hefty chunk of graphics processing.

However, the XBox360's GPU is unique, and I do believe that the 256gb/sec bandwidth will be a welcome increase, and largely due to the on die memory. With this setup 16 or 32 megabytes would be equivalent to a 256 or 512mb card nowadays.

I realize this is an Xbox forum, but if you're going to educate people, at least do it correctly. The xbox definately has its strengths, but know your enemy before yuo bash it.

I'm not going to sit here and argue aimlessly with xbox fanboys or ps3 fanboys, i just found it necessary to add my two cents.

i'll give it that you have some good tips, but your analytical statements are somewhat misleading.

to have the cell assist the rsx isn't as simple as running a channel between the two to feed, the cell has to reprogram itself to calculate the exact way the rsx does, because in the end, it all pumps out of the rsx. therefore, there's little the cell could do to help. there's a reason there aren't two cells powering the ps3. the cell wasn't designed for graphical calculations or any operations similar.

yes, the "TFLOP count" for the ps3 is about 2 (actually 1.87), but that's at MAX efficiency. due to the inefficiencies of the die cast on the gpu, the non-usa, and the in-order processes of the spes, you need to cut that number down at least 30%, while the 360 you only need to take off about 2%.

another thing to point out is that for system bandwidth, yes the 360 does have more than 5x the amount for ps3; that's partly because the ps3 bandwidth is point-to-point, while the 360's is system sharing. don't let this mislead you, because just because it's point-to-point, don't think they've given enough, because they very much haven't.

there are bottlenecks on every system. the ps had the fan problem and a little bandwidth problem that could be worked around. the xbox had gpu burnouts and die cast problems. but the ps3 bandwidth problem is A HUGE PROBLEM. you can't run next-gen elements on such a small amount. it's just not very feasible if you want to get the maximum power out of the hardware.

look, i might sound like a fanboy, but i'm not. in case you didn't know, i dink around with the alpha's and beta's over at my friend's house periodically because he's a moron (no offense, chris, but it's true).

if you want to find non-biased remarks about the hardware, you can trust me. it's good to see someone who is more technical than most finally join the forums, maybe we can debate a little later.

but for now, cheers!