Xbox 720 to ship with "insanely powerful CPU"

Yeah because an obsecure, "insanely powerful" CPu in the PS3 really helped it so much. The graphics card is the big daddy of the gaming world and it always will be. I bet that CPU couldn't outperform an I7-2600 when it comes to gaming anyway.

[QUOTE="04dcarraher"]

[QUOTE="PC_Otter"] Very much so a vector processor, but IIRC it's much more flexible. Not sure where the latest CUDA and AMD GCN architectures sit in terms of flexibility in comparison to Cell. However, Cell is still a terrible CPU for actually managing a system or server, hence why Cell based servers and supercomputers always had a general processor paired with it. The single PPE really was never enough for anything beyond a gaming console.tormentos

Problem is that also you can not gauge a CPU with GFLOPS performance charts because that performance is for singular/parallel workloads, and that a processing unit never reach's its theoretical peak. Like the PS3's Cell has a theoretical peak of 230 GFLOPS but in reality its only about 180 GFLOPS. the correct form to gauge a cpu's performance is instructions per seconds. And again the 360's cpu is nearly 2x faster then PS3's Cell which is why the PS3 has less multitasking abilites like being able to download a demo and still be able to play a game while playing music.

This is funny.... Im a Pc gamer and have no bias towards to either brand of console, so im not influenced by MS or Sony...... The Cell in some ways does kill the 360's Xenon cpu however in other ways the Xenon kills the Cell. For singular/parallel jobs =Cell, for all other CPU jubs= Xenon. The Xenon does 19,000 MIPS while the Cell only does 10,000.

Again the Cell was designed to a general purpose cpu handling all jobs including rendering, but they realized that it couldnt do the job so that is when they added the RSX to save the console they didnt want to scrap the Cell because of their investment into it. Also they had no plans to have the SPE's to supplement the RSX. until years after release. Also you need to realize also that the Cell does not do a ton of GPU tasks, The RSX still does 95% of the rendering. The Cell's SPE's supplement the RSX to allow graphical abilities to be on par with the Xenos in the 360. the Xenos is faster , but both gpu's are limited to 256mb around of video memory and a slow memory bus speed and both use Pixel Shader 3. So both do not really dont get a leg up from one another based on their specs.

Insanely powerful?

Sounds hardcore,man.

freedomfreak

Ahah, I actually laughed at this.

[QUOTE="skrat_01"]It's a mass produced consumer electronics product. Don't ever hold your breath.loosingENDS

It is not unlike xbox 360 that runs Withcer 2 with better graphics than PC though

So, will probably deliver

They would need a powerful CPU for the built in Kinect sensor. A big problem with kinect is that it's very CPU dependent and it sucks away a lot of the power from the game.

Wasdie

True but.

Kinect 2 is going to have it's own CPU built in, first Kinect would have had this except cost would have been much higher.

[QUOTE="Wasdie"]

They would need a powerful CPU for the built in Kinect sensor. A big problem with kinect is that it's very CPU dependent and it sucks away a lot of the power from the game.

Deevoshun

True but.

Kinect 2 is going to have it's own CPU built in, first Kinect would have had this except cost would have been much higher.

sweet, kinect 2 can be $250 now

[QUOTE="Wasdie"]

They would need a powerful CPU for the built in Kinect sensor. A big problem with kinect is that it's very CPU dependent and it sucks away a lot of the power from the game.

Deevoshun

True but.

Kinect 2 is going to have it's own CPU built in, first Kinect would have had this except cost would have been much higher.

[QUOTE="loosingENDS"]It is not unlike xbox 360 that runs Withcer 2 with better graphics than PC though

So, will probably deliverkuraimen

I am guessing a AMD's Trinity based APU.

I am guessing a AMD's Trinity based APU.

Xtasy26

They probably gonna use PowerPC, the reject slow cpus that Apple threw away for superior Intel chips.

[QUOTE="Xtasy26"]

I am guessing a AMD's Trinity based APU.

PC_Otter

Even if the Xbox 720 launches with Microsoft taking a heavy lost per consoles, expect to pay that sum plus more back to Microsoft in the long run.ShadowDeathX

CELL doesn't handle "tons" of GPU task i.e. it doesn't even include GPU's fix functions that RSX provides. CELL's 1X PPE and 2X SPEs covers Xbox 360's PPE X3 while 4 SPEs "patches" RSX's aging design.XBox 360 and PS3 is about the same in power.ronvalencia

This is funny.... Im a Pc gamer and have no bias towards to either brand of console, so im not influenced by MS or Sony...... The Cell in some ways does kill the 360's Xenon cpu however in other ways the Xenon kills the Cell. For singular/parallel jobs =Cell, for all other CPU jubs= Xenon. The Xenon does 19,000 MIPS while the Cell only does 10,000. Again the Cell was designed to a general purpose cpu handling all jobs including rendering, but they realized that it couldnt do the job so that is when they added the RSX to save the console they didnt want to scrap the Cell because of their investment into it. Also they had no plans to have the SPE's to supplement the RSX. until years after release. Also you need to realize also that the Cell does not do a ton of GPU tasks, The RSX still does 95% of the rendering. The Cell's SPE's supplement the RSX to allow graphical abilities to be on par with the Xenos in the 360. the Xenos is faster , but both gpu's are limited to 256mb around of video memory and a slow memory bus speed and both use Pixel Shader 3. So both do not really dont get a leg up from one another based on their specs.04dcarraher

Stats are useless every one knows Cell actually in gaming applications actually beat the Xenon,is not about mips is about how helpful Cell is to the RSX,contrary to how little the Xenon help the Xenos...

Isnt thispoor excuse of back peddling..... performance numbers tell you in what areas where something is stronger or not. With normal computing/gaming the CPU only provides the data the GPU is to process, not to help it. However the Cell is not a normal cpu which is why it is harder to code for. Again the Cell was designed to be a general purpose cpu handling all jobs including rendering. The PPE is the cpu core the main driving force, and its slower then the Xenon. The only reason they started to use the SPE's was because the PS3 with only the PPE and RSX was inferior to the 360 in everyway. And they had to figure out a way to use those unsused co processors
Every single task the Xenos has to do for it self consume resources,the more you do the more you consume,once you start applying multiple FX like AA,HDR,and so on the performance take even an even bigger hit,which is why games like Gears of War 3 don't have AA,even that the xbox 360 has more ram available to use than the PS3,the Xenos is push in such a way doing everything that even implementing AA on its smaller form would cause problems.

Even though th Xenos is 25% faster then the RSX the SPE's + RSX do not get any clear advantage of the Xenos because of the fact that both have the same amount of memory which puts them in the same area for graphics detail and resolutions. The PS3 was not able to use much AA if any at all until they used MLAA which is inferior to to normal AA methods but is less demanding. However both consoles have issues with AA, as they push the consoles to the breaking limit in what they can do things getsacrificed. Gears 2 has 2x AA while UC2 had 2xAA. But with Gears 3 had no AA, UC 3 had only edge detect based AA which is basically improves the quality of large objects withstraight edges, but thin objects or odd shapes do get the method of AA.

The analogy is simple,Cell helps the RSX is such a way,that it actually becomes a fair fight even when the 360 GPU is stronger.

Problem is that this is not the case for most games, only a handful of games actually use the SPE's with any degree that makes a real difference


That is basically Cell help for the RSX,or the other way around if you like,since we all know cell was develop first and use as CPU because as a GPU alone it could not cut it.

Um no Again the Cell was designed to be a general purpose cpu handling all jobs including rendering, but the SPE's were designed to do all the rendering not the main cpu core the PPE. as a standard cpu the Cell sucks, so does the Xenon but not as bad as the Cell.the only reason Dev's started figure out how to code workloads for the SPE's is because they were desperate for processing powwer to offset the weak RSX.

tormentos

Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.

[QUOTE="PC_Otter"][QUOTE="Xtasy26"]

I am guessing a AMD's Trinity based APU.

mitu123

The only reason why they would equip their next product with a powerful CPU is because they want to implement some form of motion control technology as most games on the console market are not RTS (basically the only type of games that is CPU intensive).

When they invest in a powerful CPU and do not reciprocate with a powerful GPU, you get games that render shadows via CPU (like Skyrim).

The only reason why they would equip their next product with a powerful CPU is because they want to implement some form of motion control technology as most games on the console market are not RTS (basically the only type of games that is CPU intensive).

When they invest in a powerful CPU and do not reciprocate with a powerful GPU, you get games that render shadows via CPU (like Skyrim).

Adversary16

Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.04dcarraher

[QUOTE="04dcarraher"]Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.tormentos

Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.

Hmmm... M$ launches a super expensive console, with expensive online one year after Sony, and two years after Nintendo. I smell epic fail; maybe 30 million units sold for lifetime.

[QUOTE="tormentos"][QUOTE="04dcarraher"]Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.04dcarraher

Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.

Aside from MIPS being an archaic and useless metric that hasn't been seriously used for decades, there are some severe problems with your chart. Playstation 3 and Xbox 360 are improperly labeled leading to the false conclusion that the whole of the machine is considered. Secondly, only the PPE is considered for the PS3 and then compared to the entirety of other chips.

[QUOTE="04dcarraher"]

[QUOTE="tormentos"] Funny how Cell sucks,the 360 GPU is stronger,easier,has more available ram,and still nothing on 360 touch Uncharted 3.. It most be that the PS3 can do magic. Were ever you want to admit it or not,Cell kick the living crap out of the Xenon,and unlike what you think tons of games do use then SPE,Cell was design with graphics task in mind,sony wanted to use it as a GPU,is well know it did not live to sonys expectations,reason why sony used the RSX,and drop the idea of 2 Cell units,that say Cell still is able to handle GPU task,and it does take out load out of the RSX back,it has been stated many times. Funny how Cell sucks,the 360 has a stronger GPU and more resources available,yet they are nearly identical in power,that really does not compute.asylumni

Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.

Aside from MIPS being an archaic and useless metric that hasn't been seriously used for decades, there are some severe problems with your chart. Playstation 3 and Xbox 360 are improperly labeled leading to the false conclusion that the whole of the machine is considered. Secondly, only the PPE is considered for the PS3 and then compared to the entirety of other chips.

MIPS is not a useless nor archaic method , and its still used to gauge cpu performance, yes there are better tests to show cpu's integer processing power but for generalize usage to show processing differences MIPS is just fine. How is the chart mislabeled? its just the cpu performance based upon the specs of the processors. Also the SPE's do not include Out-of-order execution, branch prediction or caches so it is left to the programmer to effectively do those functions in code. Which means the PPE handles most of the computational workload, and leaves the SPE's to do the floating point code execution. Even with all the cell's processing power with the SPE's its still less then half the processing power of a Q6600 with Folding@home which means at best it does only 6 clock cycles per second.

But another way to look at the cpu's between the consoles is clock cycles per second where the Xenon does 6.0, while the Cell's PPE does 3.2 while a modern dual core does 9.0+, however each core on the Xenon is rated at 2.0.

16 core by itself means nothing, modern GPUs have 100smarkop2003

[QUOTE="asylumni"]

[QUOTE="04dcarraher"] Of course you bypass all the facts..... All I have to say is stop being a PS3 fanboy and take off your goggles... both the 360 and PS3 graphical abilities are nearly identical. The Cell does not kick the crap out of the Xenon in all areas as I told you. You think because the Cell's SPE's helping out RSX makes the Cell faster when doing all other normal cpu jobs? Wrong,

Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.

04dcarraher

Aside from MIPS being an archaic and useless metric that hasn't been seriously used for decades, there are some severe problems with your chart. Playstation 3 and Xbox 360 are improperly labeled leading to the false conclusion that the whole of the machine is considered. Secondly, only the PPE is considered for the PS3 and then compared to the entirety of other chips.

MIPS is not a useless nor archaic method , and its still used to gauge cpu performance, yes there are better tests to show cpu's integer processing power but for generalize usage to show processing differences MIPS is just fine. How is the chart mislabeled? its just the cpu performance based upon the specs of the processors. Also the SPE's do not include Out-of-order execution, branch prediction or caches so it is left to the programmer to effectively do those functions in code. Which means the PPE handles most of the computational workload, and leaves the SPE's to do the floating point code execution. Even with all the cell's processing power with the SPE's its still less then half the processing power of a Q6600 with Folding@home which means at best it does only 6 clock cycles per second.

But another way to look at the cpu's between the consoles is clock cycles per second where the Xenon does 6.0, while the Cell's PPE does 3.2 while a modern dual core does 9.0+, however each core on the Xenon is rated at 2.0.

[QUOTE="asylumni"]

[QUOTE="04dcarraher"] Of course you bypass all the facts..... All I have to say is stop being a PS3 fanboy and take off your goggles... both the 360 and PS3 graphical abilities are nearly identical. The Cell does not kick the crap out of the Xenon in all areas as I told you. You think because the Cell's SPE's helping out RSX makes the Cell faster when doing all other normal cpu jobs? Wrong,

Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.

04dcarraher

Aside from MIPS being an archaic and useless metric that hasn't been seriously used for decades, there are some severe problems with your chart. Playstation 3 and Xbox 360 are improperly labeled leading to the false conclusion that the whole of the machine is considered. Secondly, only the PPE is considered for the PS3 and then compared to the entirety of other chips.

MIPS is not a useless nor archaic method , and its still used to gauge cpu performance, yes there are better tests to show cpu's integer processing power but for generalize usage to show processing differences MIPS is just fine. How is the chart mislabeled? its just the cpu performance based upon the specs of the processors. Also the SPE's do not include Out-of-order execution, branch prediction or caches so it is left to the programmer to effectively do those functions in code. Which means the PPE handles most of the computational workload, and leaves the SPE's to do the floating point code execution. Even with all the cell's processing power with the SPE's its still less then half the processing power of a Q6600 with Folding@home which means at best it does only 6 clock cycles per second.

But another way to look at the cpu's between the consoles is clock cycles per second where the Xenon does 6.0, while the Cell's PPE does 3.2 while a modern dual core does 9.0+, however each core on the Xenon is rated at 2.0.

Well, it's mislabeled since it lists Playstation 3 and 360 but only measures one or part of one component of the whole with no indication of what is actually being compared.

And yes, MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.

This is what also makes it even more obsurd to only count the PPE. The SPE's contribute to processing power of the Cell chip, so if you actually wanted to compare the processing power among CPU's, they should be included, not excluded to make others seem better. Pure FUD.

Clock cycles per second is the metric measured in Hz and most modern cpu's can be measured in millions (or even billions) of Hz. Perhaps you're thinking of operations per clock cycle? This is ok, but you run into issues with varying architectures since it counts some of the luxries included in modern x86 cpu's that involve instruction handling whereas the numbers for the Xenon and Cell would be more skewed towards actual execution operations.

"The processor benchmark called MIPS has nothing to do with the company name. In the context of CPU performance measurement, MIPS stands for 'Million Instructions Per Second' and is probably the most useless benchmark ever invented. The rest of this page concerns MIPS as a benchmark, not the company (also discussed here are the MFLOPS and SPEC benchmarks, plus a comment on memory bandwidth).

The MIPS rating of a CPU refers to how many low-level machine code instructions a processor can execute in one second. Unfortunately, using this number as a way of measuring processor performance is completely pointless because no two chips use exactly the same kind of instructions, execution method, etc. For example: on one chip, a single instruction may do many things when executed (CISC = Complex Instruction Set Computing), whereas on another chip a single instruction may do very little but is dealt with more efficiently (RISC = Reduced Instruction Set Computing). Also, different instructions on thesamechip often do vastly different amounts of work (eg. a simple arithmetic instruction might take just 1 clock cycle to complete, whereas doing something like floating point division or a square root operation might take 20 to 50 clock cycles)."

[QUOTE="ronvalencia"]CELL doesn't handle "tons" of GPU task i.e. it doesn't even include GPU's fix functions that RSX provides. CELL's 1X PPE and 2X SPEs covers Xbox 360's PPE X3 while 4 SPEs "patches" RSX's aging design.XBox 360 and PS3 is about the same in power.tormentos

MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.asylumni

Well that "insanely powerful cpu" ain't gonna be worth a damn if the gpu ain't insanely powerful which I doubt because I think the 360's successor will be a slight upgrade to the 360 (think Gamecube and Wii).

[QUOTE="asylumni"] MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.RyviusARC

"The processor benchmark called MIPS has nothing to do with the company name. In the context of CPU performance measurement, MIPS stands for 'Million Instructions Per Second' and is probably the most useless benchmark ever invented. The rest of this page concerns MIPS as a benchmark, not the company (also discussed here are the MFLOPS and SPEC benchmarks, plus a comment on memory bandwidth).

The MIPS rating of a CPU refers to how many low-level machine code instructions a processor can execute in one second. Unfortunately, using this number as a way of measuring processor performance is completely pointless because no two chips use exactly the same kind of instructions, execution method, etc. For example: on one chip, a single instruction may do many things when executed (CISC = Complex Instruction Set Computing), whereas on another chip a single instruction may do very little but is dealt with more efficiently (RISC = Reduced Instruction Set Computing). Also, different instructions on thesamechip often do vastly different amounts of work (eg. a simple arithmetic instruction might take just 1 clock cycle to complete, whereas doing something like floating point division or a square root operation might take 20 to 50 clock cycles)."

savagetwinkie

[QUOTE="04dcarraher"]

[QUOTE="asylumni"]

Aside from MIPS being an archaic and useless metric that hasn't been seriously used for decades, there are some severe problems with your chart. Playstation 3 and Xbox 360 are improperly labeled leading to the false conclusion that the whole of the machine is considered. Secondly, only the PPE is considered for the PS3 and then compared to the entirety of other chips.

asylumni

MIPS is not a useless nor archaic method , and its still used to gauge cpu performance, yes there are better tests to show cpu's integer processing power but for generalize usage to show processing differences MIPS is just fine. How is the chart mislabeled? its just the cpu performance based upon the specs of the processors. Also the SPE's do not include Out-of-order execution, branch prediction or caches so it is left to the programmer to effectively do those functions in code. Which means the PPE handles most of the computational workload, and leaves the SPE's to do the floating point code execution. Even with all the cell's processing power with the SPE's its still less then half the processing power of a Q6600 with Folding@home which means at best it does only 6 clock cycles per second.

But another way to look at the cpu's between the consoles is clock cycles per second where the Xenon does 6.0, while the Cell's PPE does 3.2 while a modern dual core does 9.0+, however each core on the Xenon is rated at 2.0.

Well, it's mislabeled since it lists Playstation 3 and 360 but only measures one or part of one component of the whole with no indication of what is actually being compared.

And yes, MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.

This is what also makes it even more obsurd to only count the PPE. The SPE's contribute to processing power of the Cell chip, so if you actually wanted to compare the processing power among CPU's, they should be included, not excluded to make others seem better. Pure FUD.

Clock cycles per second is the metric measured in Hz and most modern cpu's can be measured in millions (or even billions) of Hz. Perhaps you're thinking of operations per clock cycle? This is ok, but you run into issues with varying architectures since it counts some of the luxries included in modern x86 cpu's that involve instruction handling whereas the numbers for the Xenon and Cell would be more skewed towards actual execution operations.

Actually, SGEMM benchmark is better than just theoretical FLOPs. For example, AMD RV770 reaching 1000GFLOPs in SGEMM SGEMM benchmark was ported to AMD Radeon HD.

If one going to include CELL's SPUs, then one must factor in AMD Llano's 400 stream processors or AMD Trinity's 384 stream processors. It's funny that PPC fanboys repeatedly doesn't factor in AMD's Radeon HD math array processors.

From the certain Linux based application benchmarks, IBM PPE @ 3.2Ghz performs like PowerPC 970 @ 1.6 Ghz. In terms of CPU design, IBM PPE is like Intel Atom i.e. both are in-order dual instruction issue per cycle and designed to be cheap.

PS; I have tangled with Blachford on serveral occasions (not in this forum).

[QUOTE="04dcarraher"]Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.tormentos

[QUOTE="tormentos"][QUOTE="04dcarraher"]Needless to say it seems you dont understand that both the PS3 and 360 are nearly identical when all resources are being used. also with the PS4 it looks like to use a normal setup because the Cell is a pain in the butt to code for and more expensive to use.ronvalencia

Doubt it.

[QUOTE="asylumni"] MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.RyviusARC

Not to mention the various ways to measure FLOPS and the tendancy for hardware venders to choose methods favoring their own equipment (hence Sony's 2 teraFLOPS rating for the PS3). I'm not saying it's perfect, just a better option than MIPS and there are ways it can be used to get a general idea of comparative performance.

[QUOTE="asylumni"]

[QUOTE="04dcarraher"]

MIPS is not a useless nor archaic method , and its still used to gauge cpu performance, yes there are better tests to show cpu's integer processing power but for generalize usage to show processing differences MIPS is just fine. How is the chart mislabeled? its just the cpu performance based upon the specs of the processors. Also the SPE's do not include Out-of-order execution, branch prediction or caches so it is left to the programmer to effectively do those functions in code. Which means the PPE handles most of the computational workload, and leaves the SPE's to do the floating point code execution. Even with all the cell's processing power with the SPE's its still less then half the processing power of a Q6600 with Folding@home which means at best it does only 6 clock cycles per second.

But another way to look at the cpu's between the consoles is clock cycles per second where the Xenon does 6.0, while the Cell's PPE does 3.2 while a modern dual core does 9.0+, however each core on the Xenon is rated at 2.0.

ronvalencia

Well, it's mislabeled since it lists Playstation 3 and 360 but only measures one or part of one component of the whole with no indication of what is actually being compared.

And yes, MIPS is useless. It's inerantly undefined and doesn't give an actual indication of the performance of the chip. That is why FLOPS is used, it's an actual defined task instead of just anything that the chip does.

This is what also makes it even more obsurd to only count the PPE. The SPE's contribute to processing power of the Cell chip, so if you actually wanted to compare the processing power among CPU's, they should be included, not excluded to make others seem better. Pure FUD.

Clock cycles per second is the metric measured in Hz and most modern cpu's can be measured in millions (or even billions) of Hz. Perhaps you're thinking of operations per clock cycle? This is ok, but you run into issues with varying architectures since it counts some of the luxries included in modern x86 cpu's that involve instruction handling whereas the numbers for the Xenon and Cell would be more skewed towards actual execution operations.

Actually, SGEMM benchmark is better than just theoretical FLOPs. For example, AMD RV770 reaching 1000GFLOPs in SGEMM SGEMM benchmark was ported to AMD Radeon HD.

If one going to include CELL's SPUs, then one must factor in AMD Llano's 400 stream processors or AMD Trinity's 384 stream processors. It's funny that PPC fanboys repeatedly doesn't factor in AMD's Radeon HD math array processors.

From the certain Linux based application benchmarks, IBM PPE @ 3.2Ghz performs like PowerPC 970 @ 1.6 Ghz. In terms of CPU design, IBM PPE is like Intel Atom i.e. both are in-order dual instruction issue per cycle and designed to be cheap.

PS; I have tangled with Blachford on serveral occasions (not in this forum).

I've got not problem with including the graphics hardware integrated in APU's. If it's on the die it should be accounted. It's the same issue I had with that chart, excluding part of the processor to make others appear better.

And of course, as with any comparison, a bit of rationality helps. Ie., even with the Cell able to outperform, say the FX-60, in processing power, that doesn't mean it would be suitable for a PC environment or is a direct threat to the x86 architecture, there are always more factors involved.

Of course you bypass all the facts..... All I have to say is stop being a PS3 fanboy and take off your goggles... both the 360 and PS3 graphical abilities are nearly identical. The Cell does not kick the crap out of the Xenon in all areas as I told you. You think because the Cell's SPE's helping out RSX makes the Cell faster when doing all other normal cpu jobs? Wrong, Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.04dcarraher

I just want to hear a logical explanation on how the PS3 has the best looking game on consoles,when the 360 has everything going for it,including a GPU that almost double the PS3 GPU in polygons...tormentos

360 has demonstrated superiority over PS3 concerning multiplats. So, look inward and ask yourself why a system capable of pushing such exquisite gfx in what, 2-3 games, can barely keep pace with the 360 in scores(hundreds?) of games.

As for your assertion that PS3's 2 or 3 amazing games are sooo much better than what's on the 360...

[QUOTE="04dcarraher"] Of course you bypass all the facts..... All I have to say is stop being a PS3 fanboy and take off your goggles... both the 360 and PS3 graphical abilities are nearly identical. The Cell does not kick the crap out of the Xenon in all areas as I told you. You think because the Cell's SPE's helping out RSX makes the Cell faster when doing all other normal cpu jobs? Wrong, Its no different , then someone comparing an Intel i5 750 with a 6670 gpu vs an AMD A8 3850 with a 6570 gpu just because the AMD cpu has gpu that works with that 6570 does not mean that 3850 cpu is faster overall then the i5 750.tormentos

[QUOTE="Adversary16"]The only reason why they would equip their next product with a powerful CPU is because they want to implement some form of motion control technology as most games on the console market are not RTS (basically the only type of games that is CPU intensive).When they invest in a powerful CPU and do not reciprocate with a powerful GPU, you get games that render shadows via CPU (like Skyrim).ronvalencia

What would've been a superior setup to Cell + RSX would've been a truly combined single chip system sporting a large array of SPEs to operate in lieu of the pixel and vertex shaders on the RSX as well as how SPEs are used now. TMUs and ROPs would've been directly on the processor itself, with a 128 bit memory bus running to 512 MB of XDR2 memory.

Such a processor would've gone against the Cell consortium in terms of making a single processor made for many uses, as the TMUs and ROPs would've gone to waste in some applications, but for a console or visual computing system, a "graphics Cell processor" would've been amazing. For example it could have 2 PPEs feeding 16 SPEs running at 3.2 GHz tied to 16 TMUs and 8 ROPs running at I guess quarter speed (I'm not too sure how different clock domains would've effected the actual processing).

Developers could allocate the 16 SPEs as they desired for graphics or whatever else like physics/animation/sound/etc.

So not only a single VERY powerful processor capable of handling the expected loads encountered for a console of this generation (and would've been superior to Xenon + Xenos), but with greater flexibility and functionality for developers, and a greatly simplified motherboard for Sony, instead of the patch work that began with the failed dual-Cell concept that Sony threw out the window.

And I got bored so I made this monster:

360 has demonstrated superiority over PS3 concerning multiplats. So, look inward and ask yourself why a system capable of pushing such exquisite gfx in what, 2-3 games, can barely keep pace with the 360 in scores(hundreds?) of games. As for your assertion that PS3's 2 or 3 amazing games are sooo much better than what's on the 360...gamecubepad