How can Nvidia be ahead of ATI but 360 GPU is on par with RSX?

[Inane_Dork]

edit - also, re. FP10 - RSX supports FP10, FP16, FP32 etc.

Whoah, where did that come from? We know almost nothing about the RSX other than clockspeed and it supports FP10? I certainly don't recall reading anything about that.

 

[Titanio]

Whoah, where did that come from? We know almost nothing about the RSX other than clockspeed and it supports FP10? I certainly don't recall reading anything about that.

Maybe I misread that, but I assumed G70 supports FP10. If not, FP16 it is, for its "min" HDR precision, which shouldn't be a problem either (especially if MSAA is not a simultaneous possibility ). Either way, placing it amongst "advances" over RSX/G70 etc. seems odd.

 

[vblh]

Acert, that is word for word, the best thing i've read with regards to this subject that i've seen on this board for a while. It was & is very tiring reading the ongoing bickering between the PS & MS camps here. People, just take each for what they are, different takes on the same question. One will be better at some somethings things & the other will be better at others.
As much as i like the new tech in the 360, i won't be getting it at launch because of the US etc but because the games appeal to me & it's going to make up my Home Theatre set (32in TV, X360 for games & movies & Logitech Z5500 for sound). Rather simple for some but suits me fine.

Just my 2cents

 

[pjbliverpool]

It's simple: RSX is NV40 with some tweaks (symmetrical ALUs in each fragment pipe being the most obvious) and running faster on 90nm. Started back in 2002.

Xenos was started back in 2002 as well, but is about 2 NVidia generations ahead. NVidia won't have anything this advanced until after G80. Call it G90, if you like - 2007.

Jawed

Wow I didn't know they had released details of the G80. Tell us all about it!

 

[Jawed]

Jawed

 

[Shifty Geezer]

G80's powering the new Sega arcade machine. TEXAN says so. And it's better than Xenos too.

 

[Alpha_Spartan]

R520.



There is a difference between features and performance.

To put it simply:

• G70 is designed with DX9/SM3.0 in mind with the desktop as the target platform.
• Xenos is designed with many DX10 elements in mind with a console as the target platform.

I listed a good number of features above (unified shaders array, MEMEXPORT, eDRAM, FP10, fully orthogonal ROPs in regards to FP10/16 and MSAA, hardware tesselation, etc) that are more advanced features. That does not mean every single one improves performance or IQ, but they are features not found on DX9 hardware (or even desktop parts in some circumstances).

Overlooking the differences is about equivlant as looking at Xenon/CELL, noting they both are PPC based, multicore, same frequency and developed on a similar time frame, and shrug. They may be comparable, and even perform similarly, but the architectures are VASTLY different. And in certain circumstances one design is advantageous over the other.

On the GPU side of the equation there is no question that Xenos is much more advanced than G70 in regards to features. Performance, of course, is the unknown.

No one is saying RSX is not next gen; but from a design perspective--and its target (DX9)--it does incorperate "older" technology. It is not a clean slate design.

But before everyone gets defensive, I will paraphrase Dave from a long time ago: Xenos is high risk, high reward. It is a TOTALLY new design. It could be extremely effecient and powerful, or it may end up having significant issues that were unaccounted for. That is what happens with first generation clean slate designs. CELL runs the same risk. Obviously these are calculated risks (e.g. Xenos is not a completely new design in that it is following the R400, so there was a test case) but risks none the less.

Sometimes the "tried and true" method is not only safer, but also more effecient/effective. So noting that RSX is from the G70 line (which NV claims it is) and Xenos shares more in common with ATI's unreleased R600 series (which cannot be released without an API! So there is more to the issue when considering the "Whys and Hows") is not some diss on RSX.

But there is no question Xenos has some features RSX does not.

Arcert, to Playstation myrmidons even the slightest assertion that something in the Xbox 360 may be "better", "faster" or "more next gen" than the almighty, unbeatable PS3, then they jump on the defensive. When you start to bring an individual down to earth, they get upset because they love how things appear in their own little dream world.

But kudos to the post. It was very well-balanced.

 

[X-AleX]

Was Xenos not in the Beta kits? I thought it was "feature complete", just underclocked? Also, I believe Xenos taped out in Q1.

Yes, but beta kits were out in August. Too late. These games are developed basically on X800s
Xenos is far ahead in regard to RSX

 

[seismologist]

You know the gist of this thread was that there has to be some tradeoffs that people aren't aware of. Xenos can't have more power, more features, same cost, and arrive 6 months earlier.
It just doesn't work that way.

Maybe these tradeoffs will be more apparent once we have the details on RSX.

 

[zidane1strife]

A unified shader arch, If I've remember correctly, should be able to either:

a.) Allow similar performance to a dedicated one with less transistors/lower clock

or

b.) Allow greater performance for a similar transistor/clock budget.

Yet the xenos is not only unified but on top of that has an additional chip with eDRAM which eases b/w constraints, and allows for supposedly near performance penalty free 4xAA + HDR, etc.

Given it's virtually 'free' from the penalties of AA, and that it should offer better performance for similar clock/transistor budget, we should be seeing things that blow g70 based demos out of the water, yet we are not.

What gives? That is the question.

 

[Alpha_Spartan]

And would you call one of these processors 'next-gen' and the other 'current-gen'? I don't think so. I appreciate the differences of Xenos and advanced features, but as i say I don't think it can be called 'next-gen', at least not counting a generation as the next GPU after the previous one.

I think he was referring to unified shaders and other DX10 features. For example if the RSX was based on the G80 it would be considered a "next gen" since the G80 is supposed to be DX10 part with unified shaders.

Again, I'm getting the feeling that some of you guys are putting semantics under a microscope to make sure that it doesn't "slight" the RSX in any way. Some people need to come down to earth or they will find themselves heading straight for disappointment.

 

[Jawed]

You know the gist of this thread was that there has to be some tradeoffs that people aren't aware of. Xenos can't have more power, more features, same cost, and arrive 6 months earlier.
It just doesn't work that way.

Maybe these tradeoffs will be more apparent once we have the details on RSX.

R300

Jawed

 

[Titanio]

Again, I'm getting the feeling that some of you guys are putting semantics under a microscope to make sure that it doesn't "slight" the RSX in any way. Some people need to come down to earth or they will find themselves heading straight for disappointment.

This is at least as true of the "Xenos is > all" crowd, however. The number of times I've seen the implication that Xenos is the equivalent of an R600, for example, is quite astounding. Or that "newer" features necessarily = better performance. Or that Xenos is a DX10 chip

These are less supportable positions than others, IMO (particularly the first!). If luck has been pushed with regard to the capability of any chip, it's Xenos IMO.

There are points of advantage to highlight on both chips. It comes down to what you think is more important. I've my own opinion on that, as I'm sure everyone here does. More importantly though, what devs find more important and more critical for their games will determine which is the more suitable/"better" chip, but even that is likely a case-by-case thing.

 

[Alpha_Spartan]

You know the gist of this thread was that there has to be some tradeoffs that people aren't aware of. Xenos can't have more power, more features, same cost, and arrive 6 months earlier.
It just doesn't work that way.

Maybe these tradeoffs will be more apparent once we have the details on RSX.

It's all in the planning. I doubt that as much planning went into developing the RSX than what was put into Cell. On the flipside, I doubt that as much planning went into Xenon than what was put into Xenos.

It's quite obvious that Sony built the PS3 around the Cell and Microsoft quietly built the Xbox 360 around Xenos. Thus each system's strengths will reflect this.

Anyone whose waiting for a clear-cut Xenos vs. RSX article where they can circle jerk at GAF or TXB respectively with their cohorts will have to keep on waiting. Let's just say that this irrelevant debate will still be going and just as foggy as it is now when the PS4 and Xbox 3 are launched.

However, I can look at both architectures and tell you what each company focused on: It's where they spent the most money. For Sony it is the Cell and for Microsoft it's Xenos. I mean, is it surprising that Cell is similar to chips we'll see 10 years from now? No. Is it surprising that Xenos contains next generation GPU architecture? No. Money well spent on both fronts.

 

[Alpha_Spartan]

This is at least as true of the "Xenos is > all" crowd, however. The number of times I've seen the implication that Xenos is the equivalent of an R600, for example, is quite astounding. Or that "newer" features necessarily = better performance. Or that Xenos is a DX10 chip

These are less supportable positions than others, IMO (particularly the first!). If luck has been pushed with regard to the capability of any chip, it's Xenos IMO.

I'll give you that. Anyone who thinks that Xenos is faster than a PC graphics chip coming out after it is seriously delusional. Xenos = R600 in architecture only, not in speed.

 

[Shifty Geezer]

Again, I'm getting the feeling that some of you guys are putting semantics under a microscope to make sure that it doesn't "slight" the RSX in any way. Some people need to come down to earth or they will find themselves heading straight for disappointment.

Considering I've gone on record several times as saying I expect Xenos could well be better than RSX, you're aiming this statement at the wrong guy, mister! My response was to dukmahsik's assertion that RSX is current gen, Xenos is next gen, which is a wild and unsubstantiable claim given no criteria for judging what is current-gen and what is next-gen. Acert's smart observations on features were, as he recognised, somewhat out of context in his reply. I'm not daft enough to not recognise the differences between the hardwares, nor am I blighted enough to care which platform has a 'better' GPU as long as the system does a good job of producing the games I want to play. A carpet-bomb statement like 'this GPU is a generation ahead of that GPU' shouldn't go uncontested in a smart forum though. It needs explaining, and the impact of being a generation ahead if true should be clarified.

 

[Jawed]

Ah yes, it's that argument all over again about the useless features of Xenos:

• increased efficiency (50%+) for all shader code
• increased efficiency (100%+) for all texture operations (out of order texturing)
• zero-latency dynamic branching (shader pipeline never needs to be flushed)
• free 4xAA
• realworld fill-rate of 4G pixels/s as memory is fast enough
• a shadow rendering capability that runs at roughly 6x the speed of RSX
• ROPs that can populate the framebuffer with shadow data at three to four times RSX speed (32G samples per second) - RSX is bandwidth limited
• in-built tessellation which increases the efficiency of both the incoming vertex stream and level of detail manipulation of all geometry - meaning that none of this slows down the CPU
• particles rendered solely on the GPU, no CPU overhead, no CPU-GPU bandwidth saturation

Jawed

 

[ihamoitc2005]

unified shaders misunderstood

A unified shader arch, If I've remember correctly, should be able to either:

a.) Allow similar performance to a dedicated one with less transistors/lower clock

or

b.) Allow greater performance for a similar transistor/clock budget.

Unified shaders do not = better performance/clock-transistor at all times

7800GTX overclocked @ 500mhz, ~304 Million transistors
Xenos @ 500mhz, ~337 Million transistors

Situation where 0 vertex shader required, o.c. 7800GTX = Xenos
Situation where 1 vertex shader required, o.c. 7800GTX > Xenos
Situation where 8 vertex shader required, o.c. 7800GTX >>>> Xenos

But, Xenos has AA & HDR while mutually exclusive on 7800GTX ... trade-off.

 

[mckmas8808]

Question?

If the RSX is designed in such a way that it allows great connectivity between the GPU and CPU; which gives the devs the ability to make their games look as if they were made on the Xenos wouldn't that be a positive for Nvidia and Sony? I guess I'm thinking like Titanio when he said...

More importantly though, what devs find more important and more critical for their games will determine which is the more suitable/"better" chip, but even that is likely a case-by-case thing.

Could it be that the connectivity that the CELL can pump to the RSX could put the overall graphics ahead what the Xenon and Xenos could do? If and I mean IF the RSX is weaker overall than Xenos, maybe Sony didn't mind because they helped designed the RSX to communicate with the CELL in such a great way to help out. Maybe they just looked at things differently than MS?

But we won't know for sure until somebody releases some damn information about the RSX.

 

[Titanio]

Given it's virtually 'free' from the penalties of AA, and that it should offer better performance for similar clock/transistor budget, we should be seeing things that blow g70 based demos out of the water, yet we are not.

Thing is, Xenos is not using a similar transistor budget for its shaders as G70/RSX.

Your first point "Allow similar performance to a dedicated one with less transistors/lower clock" is also built on the assumptions that a) the dedicated chip is being underutilised to a degree that negates the advantages in terms of clockspeed/greater logic and b) that a unified shader remains as efficient at a specific task as a dedicated shader does. The first is a questionable assumption in a closed system with good performance analysis tools. The second, and the quality of ATi's implementation etc. remains an unknown - but we do know that NVidia's implementation is very refined and mature at this stage.

Also, with regard to "performance free" HDR + 4xAA - I'm not sure how free that is. 2xAA with HDR, perhaps, but I know some games (PGR3?) are using 2xAA, and I assume the reason is performance.

On a general bandwidth note, also, I think once you get over the mental hurdle of accepting that RSX won't do HDR and AA simultaneously, it likely turns the bandwidth situation in many cases to PS3's favour for everything else. Well, unless the figures I'm thinking of are wrong