Anyone here love DOOM!! !! !! !!

Post Reply New Topic

DOOM is one of the greatest shell-blasting games of all time.

Indeed it is. I still like to play it every now and then.

Yeah, you've convinced me

So many things wrong with that I wouldn't know where to start. It's kind of funny that someone has the time to make that though. I mean WHY?!? Why waste your time making up random videos like that?

Oh, what the heck, let's go through this. Their first "chapter"-yeah, duh it's not selling as well as Sony would no doubt like. That isn't causing Sony to lose money-they lose hundreds on each Playstation 3 produced. They're selling it at a big loss. It is massively outselling the 360 in Japan though. It overtook the 360's year lead almost instantly. Though we don't have figures yet, anecdotally it seems to have really picked up from the price cut, combined with it's library starting to fill out. This person evidently doesn't realize the 360 isn't selling as well as expected either, and many analysts say the PS3 is doing better than the 360 last year.
And the Wii is outselling both...which hardly means it's the best system.

"Chapter 2": This is even more ludicrous. Of COURSE those performance numbers are theoretical, they are on the 360 too. He randomly pulls some other number out of the air, and completely misses the point-it's MORE POWERFUL. We all know neither will get close to theoretical numbers, and no one really cares what those numbers are. What we care about is how it will affect GAMES down the line. Both CPUs are actually extremely inefficient on branching code by the way.

Then he pulls out this random "remember, the Playstation 2 has twice the FLOPS as the X-Box, and it didn't make the PS2 perform better!" Well no kidding. If he actually knew anything about these systems architecture, he'd know the Playstation 2 DOES NOT HAVE HARDWARE T&L. The X-Box does. On the Playstation 2, the vector units are used largely to set up polygons. The X-Box uses it's Geforce 3 GPU, which has a theoretical 80 Gigaflops. Not to mention it's MASSIVELY MORE MODERN GPU.

You can't just pull random numbers up and compare them like that without actually understanding the hardware.

He then goes on, supposedly going to tell us which console wins based on "real world performance", and instead goes on about how Cell is more complex to code for. Well no kidding. We know that. That doesn't change the fact that it's a more powerful chip. It's just plain got more execution hardware on it. (I'm leaving out that he's making assertions that aren't backed up-improbable things like one of the SPEs is used to run the OS-the OS would be run on the PPE, not an SPE, and of course ignores that the 360 is ALSO using resources to run the OS. Duh!). We get screen after screen of this ridiculousness. This idiot needs to read a real hardware site if he actually wants to learn about CPU architecture. But he doesn't care, he's just a fanboi who wants "his" team to win.

Next he says "The PS3 only has 512MB L2 cache, the 360 has 1MB!" Yeah, no kidding moron. That's for that ONE CPU. The 360 has 3 CPUS sharing that 1MB? Remember? Each SPE has it's own 256KB. That's 2.25x as much cache/scratch RAM on the PS3 as 360, if we want to randomly count numbers.

Next he says Microsoft made a "better decision for developers..." because of the identical cores. Maybe, honestly both of these CPUs are pretty limited next to a PC's CPU. But regardless, the PS3 has a lot more power there to tap, and won't be maxed out as fast...and plus, presumably as gamers we care about the final results, not necessarily how fast developers can get the most out of the hardware.
And also, regarding the asymetric thing, both Intel and AMD have shown interest in moving towards that, with specialized cores for different uses. So that's something that's going to come sooner or later for developers regardless.

:Chapter 3": Next up we get confused ranting about GPUs. First he compares it to a completely different achetecture. Apperently not realizing you can't just count shaders. Shaders are just execution units-basically specialized CPU cores, and just like CPUs, you can't just count the number of CPUs there and declare a winner. Heck, the Geforce 8800GTX "only" has 128 shaders, versus 320 on ATi's new part-and of course the 8800 is easily more powerful. Most of ATi's shaders on their new part aren't nearly as advanced as those in the 8800-they can't do as much work. (Plus he's once again randomly throwing out theoretical performance numbers, ignoring what we actually care about-GAME PERFORMANCE.)

Next we get a flat out lie-360's Xenos is "far more advanced...much more performance!", and throws out more random numbers (which he's uncritical of-of course-despite just attacking RSX's numbers as theoretical). Then he goes on about Xenos' "2 advanced effects it does for free!"...one of which actually isn't correct (HDR), and the other of which is only partially true (anti aliasing).

Xenos is actually comprised of two pieces-the main GPU has been estimated at around 200 million transistors-that's INCLUDING the system's memory controller. It's at least 50% less than RSX, which should tell you something right there (same with Cell versus Xenon). That "free" effect is because the second chip contains hard wired anti aliasing effects. It's not really free though, because using it eats up memory bandwidth, which is at a premium. Turn it on? A frame can't fit entirely in that 10MB cache. Turn it up to 4x? It has to hit main system RAM even harder.
Then he claims that embedded RAM is "revolutionary". Well, no, actually it's not. The Playstation 2 had 4MB, the Gamecube 3MB, and the Playstation 3 has it's own separate video RAM, like a PC. As soon as you fill up that 10MB buffer, you're hitting main RAM again-and a single frame with AA turned on already spills in to main RAM as I recall.
He then goes on about the embedded RAM's memory bandwidth-which as I just said, helps alleviate strain on main system RAM, but doesn't even fit a single frame with any AA turned on.

There are some very bizare situations you could find, where Xenos would outperform RSX...if you somehow got it to process more vertex than pixel data. But the reality is, that's not how today's code is written and developers have no reason to write strange code like that for the Playstation 3, when they know exactly what kind of hardware they're writing for. Real world, RSX just plain does more operations per second (it also quite likely has double the number of ROPs as Xenos does, though that info has never been publicly revealed to me knowledge, but Nvidia's 7900 series GPU RSX is based on has 16 ROPs versus 8 on Xenos). (He also actually says RSX has more vertex shaders than it really does, but wouldn't comprehend that Xenos shaders DO LESS WORK than RSX's). Yeah, a unified shader architecture helps prevent hardware from sitting unused...but that's ignoring three huge issues-Xenos can't just arbitrarily devote any shaders to pixel or vertex data, it can only split it's shaders into three groups, each doing either one or the other. Second, developers KNOW WHAT HARDWARE is in the box, and will obviously code to minimize wasted time. Third, RSX just plain does more operations (and probably has double the ROPs)-so even if it actually were less efficient, it's got more brute force to make up for it.

Then he makes the completely insane assertion that Xenos would be better than RSX at 1080p. HUH!?! His "proof" is that "inferior GPUs always show their weaknesses at higher resolutions...". Well, that's nice, but RSX is only "inferior" because you said so.

In reality, RSX has over double the bandwith of Xenos, excluding that 10MB cache (I'll get to that...). Guess what higher resolutions need...that's right, MORE BANDWIDTH! Gee, RSX wins there. But Xenos does have that 10MB cache that helps alleviate strain on main system RAM-but since it can't fit an entire frame in the cache, it still ends up worse on bandwidth. And guess what happens when you turn up the resolution? EVEN LESS OF THE FRAME fits in the cache! That means that proportionally Xenos has even LESS bandwidth available (already less than RSX) at 1080p than at 720p!
But that's not all. Like I said before, RSX probably has double the ROPs of Xenos-meaning in addition to having close to double the memory bandwidth, it can output double the pixels every second. What do higher resolutions need besides more memory bandwidth? Oh, that's right MORE PIXELS.

While I actually do agree with his assertion that this generation should stick to 720p, there's no question that RSX is better equipped to handle it. That's why quite a few games actually do use 1080p. The 360 has an option in it's menu to set the output to 1080p, but all that's actually doing is taking the game's 720p output and running it through a scaler* to get to 1080p-the game isn't actually running at 1080p. In fact not all 360 games even hit 720p. Several have used 1024x720 (like Project Gotham Racing) and had the scaler up that to the full 1280x720.

The reality is, several first generation Playstation 3 games were already running at 1080p while their second generation 360 counterparts were at 720p. An example is Marvel: Ultimate Alliance. At 1080p, the frame rate on the PS3 version wasn't supposed to be as stable, but it had better special effects, and was pushing 2.25x the resolution of the 360 version.

I've suspected that as PS3 development matures, some developers may opt to run the PS3 version of the game at 1080p, since it can basically do that higher resolution for "free" (since the game will be developed for the lowest common denominator-ie the 360-although I suspect most consumers won't understand that the 360's 1080p menu setting isn't actually 1080p). Overall though, I don't think RSX is really powerful enough to push that resolution, not without massively sacrificing other graphical elements. I do think 720p will be the norm this generation-and should be.

Next he baselessly claims that Xenos is the "first step to the future..." blah blah, and RSX is "far behind". Wow, so being more powerful is far behind. Okay.

(*This hardware scaler is one area I think Microsoft did a great job. Basically it means developers don't have to do much to set up various resolutions like 720p and 1080i, the scaler just takes the 720p output, and scales it to whatever resolution the user selects. The Playstation 3, like every other platform, the developer has to actually code for the different resolutions. That's caused an issue for early games that didn't support 1080i, as some people own early HDTVs that only support that resolution.)

OS: Next he claims the PS3's OS takes up more RAM and CPU power. No source for this one of course. The claim that the OS takes up 32MB of video RAM is especially strange. Doing what, exactly? Why would an OS store part of itself in video RAM? His 360 figures naturally don't jive with what I've seen on tech sites either.

And even if this were true, since the PS3 doesn't have as much overhead for networking stuff, I have a hard time believing Sony can't get it's memory usage lower, and regardless, PS3 has more CPU power to start with even if it was wasting more (and again, the assertion that the OS is running on an SPE is just silly).

Blu-Ray: Let's see how he manages to lie about this. Ah, PS3 can only have 256MB of textures at most! Oh, except that's a lie. RSX can use the entire 512MB available for textures, same as the 360.

Next he goes on about various games having long play times on DVD. Well duh, there are NES and Gameboy games with super long play times.

My personal thought is as a GAME MACHINE Sony would have been better off releasing PS3 without Blu Ray. They could have launched a year earlier, at probably $400 or less, but with the same better hardware (plus the better reliability, larger installed base, etc.). The 360 couldn't have stood up against that.

But that said, there's no question Blu Ray is an advantage. Blue Dragon on the 360 already uses three discs, and it won't be the last. And creativity is going to be stifled somewhat by a lack of storage space, as developers will have to shoot for 9GB max for most projects. His assertion that it doesn't benefit the consumer is a flat out lie.

Then he sums it up, claiming the 360 won three of his four categories (even though it didn't win any, here in the real world).

And of course his idiotic comparison failed to mention the Playstation 3's standard hard drive. 360 games have to be coded without the assumption that a drive won't be present. That gives PS3 developers the chance to use virtual memory among other benefits, effectively giving the PS3 more RAM and faster load times. (Oh, and that hard drive is even a normal off the shelf drive, easily user replaceable.)

And of course it fails to mention two huge benefits-the PS3 runs dramatically quieter than the 360. It's a night and day difference, despite the PS3 being more powerful/using larger chips. So far it's failure rate is beneath the radar-no mass failures at all, versus a 33%+ failure rate according to analysts on the 360.

But oh no, fanboi here thinks the 360 WINZ DUDEZ!! !

How does someone delude themselves so much? And to what end? If you own a 360 and can't afford both systems, just enjoy your 360, no reason to make up random stuff.

Yeah, you've convinced me

So many things wrong with that I wouldn't know where to start. It's kind of funny that someone has the time to make that though. I mean WHY?!? Why waste your time making up random videos like that?

Oh, what the heck, let's go through this. Their first "chapter"-yeah, duh it's not selling as well as Sony would no doubt like. That isn't causing Sony to lose money-they lose hundreds on each Playstation 3 produced. They're selling it at a big loss. It is massively outselling the 360 in Japan though. It overtook the 360's year lead almost instantly. Though we don't have figures yet, anecdotally it seems to have really picked up from the price cut, combined with it's library starting to fill out. This person evidently doesn't realize the 360 isn't selling as well as expected either, and many analysts say the PS3 is doing better than the 360 last year.
And the Wii is outselling both...which hardly means it's the best system.

"Chapter 2": This is even more ludicrous. Of COURSE those performance numbers are theoretical, they are on the 360 too. He randomly pulls some other number out of the air, and completely misses the point-it's MORE POWERFUL. We all know neither will get close to theoretical numbers, and no one really cares what those numbers are. What we care about is how it will affect GAMES down the line. Both CPUs are actually extremely inefficient on branching code by the way.

Then he pulls out this random "remember, the Playstation 2 has twice the FLOPS as the X-Box, and it didn't make the PS2 perform better!" Well no kidding. If he actually knew anything about these systems architecture, he'd know the Playstation 2 DOES NOT HAVE HARDWARE T&L. The X-Box does. On the Playstation 2, the vector units are used largely to set up polygons. The X-Box uses it's Geforce 3 GPU, which has a theoretical 80 Gigaflops. Not to mention it's a MASSIVELY MORE MODERN GPU.

You can't just pull random numbers up and compare them like that without actually understanding the hardware.

He then goes on, supposedly going to tell us which console wins based on "real world performance", and instead goes on about how Cell is more complex to code for. Well no kidding. We know that. That doesn't change the fact that it's a more powerful chip. It's just plain got more execution hardware on it. (I'm leaving out that he's making assertions that aren't backed up-improbable things like one of the SPEs is used to run the OS-the OS would be run on the PPE, not an SPE, and of course ignores that the 360 is ALSO using resources to run the OS. Duh!). We get screen after screen of this ridiculousness. This idiot needs to read a real hardware site if he actually wants to learn about CPU architecture. But he doesn't care, he's just a fanboi who wants "his" team to win.

Next he says "The PS3 only has 512MB L2 cache, the 360 has 1MB!" Yeah, no kidding moron. That's for that ONE CPU. The 360 has 3 CPUS sharing that 1MB? Remember? Each SPE has it's own 256KB. That's 2.25x as much cache/scratch RAM on the PS3 as 360, if we want to randomly count numbers.

Next he says Microsoft made a "better decision for developers..." because of the identical cores. Maybe, honestly both of these CPUs are pretty limited next to a PC's CPU. But regardless, the PS3 has a lot more power there to tap, and won't be maxed out as fast...and plus, presumably as gamers we care about the final results, not necessarily how fast developers can get the most out of the hardware.
And also, regarding the asymetric thing, both Intel and AMD have shown interest in moving towards that, with specialized cores for different uses. So that's something that's going to come sooner or later for developers regardless.

Chapter 3": Next up we get confused ranting about GPUs. First he compares it to a completely different architecture. Apparently not realizing you can't just count shaders. Shaders are just execution units-basically specialized CPU cores, and just like CPUs, you can't just count the number of CPUs there and declare a winner. Heck, the Geforce 8800GTX "only" has 128 shaders, versus 320 on ATi's new part-and of course the 8800 is easily more powerful. Most of ATi's shaders on their new part aren't nearly as advanced as those in the 8800-they can't do as much work. (Plus he's once again randomly throwing out theoretical performance numbers, ignoring what we actually care about-GAME PERFORMANCE.)

Next we get a flat out lie-360's Xenos is "far more advanced...much more performance!", and throws out more random numbers (which he's uncritical of-of course-despite just attacking RSX's numbers as theoretical). Then he goes on about Xenos' "2 advanced effects it does for free!"...one of which actually isn't correct (HDR), and the other of which is only partially true (anti aliasing).

Xenos is actually comprised of two pieces-the main GPU has been estimated at around 200 million transistors-that's INCLUDING the system's memory controller. It's at least 50% less than RSX, which should tell you something right there (same with Cell versus Xenon). That "free" effect is because the second chip contains hard wired anti aliasing effects. It's not really free though, because using it eats up memory bandwidth, which is at a premium. Turn it on? A frame can't fit entirely in that 10MB cache. Turn it up to 4x? It has to hit main system RAM even harder.
Then he claims that embedded RAM is "revolutionary". Well, no, actually it's not. The Playstation 2 had 4MB, the Gamecube 3MB, and the Playstation 3 has it's own separate video RAM, like a PC. As soon as you fill up that 10MB buffer, you're hitting main RAM again-and a single frame with AA turned on already spills in to main RAM as I recall.
He then goes on about the embedded RAM's memory bandwidth-which as I just said, helps alleviate strain on main system RAM, but doesn't even fit a single frame with any AA turned on.

There are some very bizarre situations you could find, where Xenos would outperform RSX...if you somehow got it to process more vertex than pixel data. But the reality is, that's not how today's code is written and developers have no reason to write strange code like that for the Playstation 3, when they know exactly what kind of hardware they're writing for. Real world, RSX just plain does more operations per second (it also quite likely has double the number of ROPs as Xenos does, though that info has never been publicly revealed to me knowledge, but Nvidia's 7900 series GPU RSX is based on has 16 ROPs versus 8 on Xenos). (He also actually says RSX has more vertex shaders than it really does, but wouldn't comprehend that Xenos shaders DO LESS WORK than RSX's). Yeah, a unified shader architecture helps prevent hardware from sitting unused...but that's ignoring three huge issues-Xenos can't just arbitrarily devote any shaders to pixel or vertex data, it can only split it's shaders into three groups, each doing either one or the other. Second, developers KNOW WHAT HARDWARE is in the box, and will obviously code to minimize wasted time. Third, RSX just plain does more operations (and probably has double the ROPs)-so even if it actually were less efficient, it's got more brute force to make up for it.

Then he makes the completely insane assertion that Xenos would be better than RSX at 1080p. HUH!?! His "proof" is that "inferior GPUs always show their weaknesses at higher resolutions...". Well, that's nice, but RSX is only "inferior" because you said so.

In reality, RSX has over double the bandwith of Xenos, excluding that 10MB cache (I'll get to that...). Guess what higher resolutions need...that's right, MORE BANDWIDTH! Gee, RSX wins there. But Xenos does have that 10MB cache that helps alleviate strain on main system RAM-but since it can't fit an entire frame in the cache, it still ends up worse on bandwidth. And guess what happens when you turn up the resolution? EVEN LESS OF THE FRAME fits in the cache! That means that proportionally Xenos has even LESS bandwidth available (already less than RSX) at 1080p than at 720p!
But that's not all. Like I said before, RSX probably has double the ROPs of Xenos-meaning in addition to having close to double the memory bandwidth, it can output double the pixels every second. What do higher resolutions need besides more memory bandwidth? Oh, that's right MORE PIXELS.

While I actually do agree with his assertion that this generation should stick to 720p, there's no question that RSX is better equipped to handle it. That's why quite a few games actually do use 1080p. The 360 has an option in it's menu to set the output to 1080p, but all that's actually doing is taking the game's 720p output and running it through a scaler* to get to 1080p-the game isn't actually running at 1080p. In fact not all 360 games even hit 720p. Several have used 1024x720 (like Project Gotham Racing) and had the scaler up that to the full 1280x720.

The reality is, several first generation Playstation 3 games were already running at 1080p while their second generation 360 counterparts were at 720p. An example is Marvel: Ultimate Alliance. At 1080p, the frame rate on the PS3 version wasn't supposed to be as stable, but it had better special effects, and was pushing 2.25x the resolution of the 360 version.

I've suspected that as PS3 development matures, some developers may opt to run the PS3 version of the game at 1080p, since it can basically do that higher resolution for "free" (since the game will be developed for the lowest common denominator-ie the 360-although I suspect most consumers won't understand that the 360's 1080p menu setting isn't actually 1080p). Overall though, I don't think RSX is really powerful enough to push that resolution, not without massively sacrificing other graphical elements. I do think 720p will be the norm this generation-and should be.

Next he baselessly claims that Xenos is the "first step to the future..." blah blah, and RSX is "far behind". Wow, so being more powerful is far behind. Okay.

(*This hardware scaler is one area I think Microsoft did a great job. Basically it means developers don't have to do much to set up various resolutions like 720p and 1080i, the scaler just takes the 720p output, and scales it to whatever resolution the user selects. The Playstation 3, like every other platform, the developer has to actually code for the different resolutions. That's caused an issue for early games that didn't support 1080i, as some people own early HDTVs that only support that resolution.)

OS: Next he claims the PS3's OS takes up more RAM and CPU power. No source for this one of course. The claim that the OS takes up 32MB of video RAM is especially strange. Doing what, exactly? Why would an OS store part of itself in video RAM? His 360 figures naturally don't jive with what I've seen on tech sites either.

And even if this were true, since the PS3 doesn't have as much overhead for networking stuff, I have a hard time believing Sony can't get it's memory usage lower, and regardless, PS3 has more CPU power to start with even if it was wasting more (and again, the assertion that the OS is running on an SPE is just silly).

Blu-Ray: Let's see how he manages to lie about this. Ah, PS3 can only have 256MB of textures at most! Oh, except that's a lie. RSX can use the entire 512MB available for textures, same as the 360.

Next he goes on about various games having long play times on DVD. Well duh, there are NES and Gameboy games with super long play times.

My personal thought is as a GAME MACHINE Sony would have been better off releasing PS3 without Blu Ray. They could have launched a year earlier, at probably $400 or less, but with the same better hardware (plus the better reliability, larger installed base, etc.). The 360 couldn't have stood up against that.

But that said, there's no question Blu Ray is an advantage. Blue Dragon on the 360 already uses three discs, and it won't be the last. And creativity is going to be stifled somewhat by a lack of storage space, as developers will have to shoot for 9GB max for most projects. His assertion that it doesn't benefit the consumer is a flat out lie.

Then he sums it up, claiming the 360 won three of his four categories (even though it didn't win any, here in the real world).

And of course his idiotic comparison failed to mention the Playstation 3's standard hard drive. 360 games have to be coded without the assumption that a drive won't be present. That gives PS3 developers the chance to use virtual memory among other benefits, effectively giving the PS3 more RAM and faster load times. (Oh, and that hard drive is even a normal off the shelf drive, easily user replaceable.)

And of course it fails to mention two huge benefits-the PS3 runs dramatically quieter than the 360. It's a night and day difference, despite the PS3 being more powerful/using larger chips. So far it's failure rate is beneath the radar-no mass failures at all, versus a 33%+ failure rate according to analysts on the 360.

But oh no, fanboi here thinks the 360 WINZ DUDEZ!! !

How does someone delude themselves so much? And to what end? If you own a 360 and can't afford both systems, just enjoy your 360, no reason to make up random stuff.

Yeah, you've convinced me

So many things wrong with that I wouldn't know where to start. It's kind of funny that someone has the time to make that though. I mean WHY?!? Why waste your time making up random videos like that?

Oh, what the heck, let's go through this. Their first "chapter"-yeah, duh it's not selling as well as Sony would no doubt like. That isn't causing Sony to lose money-they lose hundreds on each Playstation 3 produced. They're selling it at a big loss. It is massively outselling the 360 in Japan though. It overtook the 360's year lead almost instantly. Though we don't have figures yet, anecdotally it seems to have really picked up from the price cut, combined with it's library starting to fill out. This person evidently doesn't realize the 360 isn't selling as well as expected either, and many analysts say the PS3 is doing better than the 360 last year.
And the Wii is outselling both...which hardly means it's the best system.

"Chapter 2": This is even more ludicrous. Of COURSE those performance numbers are theoretical, they are on the 360 too. He randomly pulls some other number out of the air, and completely misses the point-it's MORE POWERFUL. We all know neither will get close to theoretical numbers, and no one really cares what those numbers are. What we care about is how it will affect GAMES down the line. Both CPUs are actually extremely inefficient on branching code by the way.

Then he pulls out this random "remember, the Playstation 2 has twice the FLOPS as the X-Box, and it didn't make the PS2 perform better!" Well no kidding. If he actually knew anything about these systems architecture, he'd know the Playstation 2 DOES NOT HAVE HARDWARE T&L. The X-Box does. On the Playstation 2, the vector units are used largely to set up polygons. The X-Box uses it's Geforce 3 GPU, which has a theoretical 80 Gigaflops. Not to mention it's a MASSIVELY MORE MODERN GPU.

You can't just pull random numbers up and compare them like that without actually understanding the hardware.

He then goes on, supposedly going to tell us which console wins based on "real world performance", and instead goes on about how Cell is more complex to code for. Well no kidding. We know that. That doesn't change the fact that it's a more powerful chip. It's just plain got more execution hardware on it. (I'm leaving out that he's making assertions that aren't backed up-improbable things like one of the SPEs is used to run the OS-the OS would be run on the PPE, not an SPE, and of course ignores that the 360 is ALSO using resources to run the OS. Duh!). We get screen after screen of this ridiculousness. This idiot needs to read a real hardware site if he actually wants to learn about CPU architecture. But he doesn't care, he's just a fanboi who wants "his" team to win.

Next he says "The PS3 only has 512MB L2 cache, the 360 has 1MB!" Yeah, no kidding moron. That's for that ONE CPU. The 360 has 3 CPUS sharing that 1MB? Remember? Each SPE has it's own 256KB. That's 2.25x as much cache/scratch RAM on the PS3 as 360, if we want to randomly count numbers.

Next he says Microsoft made a "better decision for developers..." because of the identical cores. Maybe, honestly both of these CPUs are pretty limited next to a PC's CPU. But regardless, the PS3 has a lot more power there to tap, and won't be maxed out as fast...and plus, presumably as gamers we care about the final results, not necessarily how fast developers can get the most out of the hardware.
And also, regarding the asymetric thing, both Intel and AMD have shown interest in moving towards that, with specialized cores for different uses. So that's something that's going to come sooner or later for developers regardless.

Chapter 3": Next up we get confused ranting about GPUs. First he compares it to a completely different architecture. Apparently not realizing you can't just count shaders. Shaders are just execution units-basically specialized CPU cores, and just like CPUs, you can't just count the number of CPUs there and declare a winner. Heck, the Geforce 8800GTX "only" has 128 shaders, versus 320 on ATi's new part-and of course the 8800 is easily more powerful. Most of ATi's shaders on their new part aren't nearly as advanced as those in the 8800-they can't do as much work. (Plus he's once again randomly throwing out theoretical performance numbers, ignoring what we actually care about-GAME PERFORMANCE.)

Next we get a flat out lie-360's Xenos is "far more advanced...much more performance!", and throws out more random numbers (which he's uncritical of-of course-despite just attacking RSX's numbers as theoretical). Then he goes on about Xenos' "2 advanced effects it does for free!"...one of which actually isn't correct (HDR), and the other of which is only partially true (anti aliasing).

Xenos is actually comprised of two pieces-the main GPU has been estimated at around 200 million transistors-that's INCLUDING the system's memory controller. It's at least 50% less than RSX, which should tell you something right there (same with Cell versus Xenon). That "free" effect is because the second chip contains hard wired anti aliasing effects. It's not really free though, because using it eats up memory bandwidth, which is at a premium. Turn it on? A frame can't fit entirely in that 10MB cache. Turn it up to 4x? It has to hit main system RAM even harder.
Then he claims that embedded RAM is "revolutionary". Well, no, actually it's not. The Playstation 2 had 4MB, the Gamecube 3MB, and the Playstation 3 has it's own separate video RAM, like a PC. As soon as you fill up that 10MB buffer, you're hitting main RAM again-and a single frame with AA turned on already spills in to main RAM as I recall.
He then goes on about the embedded RAM's memory bandwidth-which as I just said, helps alleviate strain on main system RAM, but doesn't even fit a single frame with any AA turned on.

There are some very bizarre situations you could find, where Xenos would outperform RSX...if you somehow got it to process more vertex than pixel data. But the reality is, that's not how today's code is written and developers have no reason to write strange code like that for the Playstation 3, when they know exactly what kind of hardware they're writing for. Real world, RSX just plain does more operations per second (it also quite likely has double the number of ROPs as Xenos does, though that info has never been publicly revealed to me knowledge, but Nvidia's 7900 series GPU RSX is based on has 16 ROPs versus 8 on Xenos). (He also actually says RSX has more vertex shaders than it really does, but wouldn't comprehend that Xenos shaders DO LESS WORK than RSX's). Yeah, a unified shader architecture helps prevent hardware from sitting unused...but that's ignoring three huge issues-Xenos can't just arbitrarily devote any shaders to pixel or vertex data, it can only split it's shaders into three groups, each doing either one or the other. Second, developers KNOW WHAT HARDWARE is in the box, and will obviously code to minimize wasted time. Third, RSX just plain does more operations (and probably has double the ROPs)-so even if it actually were less efficient, it's got more brute force to make up for it.

Then he makes the completely insane assertion that Xenos would be better than RSX at 1080p. HUH!?! His "proof" is that "inferior GPUs always show their weaknesses at higher resolutions...". Well, that's nice, but RSX is only "inferior" because you said so.

In reality, RSX has over double the bandwith of Xenos, excluding that 10MB cache (I'll get to that...). Guess what higher resolutions need...that's right, MORE BANDWIDTH! Gee, RSX wins there. But Xenos does have that 10MB cache that helps alleviate strain on main system RAM-but since it can't fit an entire frame in the cache, it still ends up worse on bandwidth. And guess what happens when you turn up the resolution? EVEN LESS OF THE FRAME fits in the cache! That means that proportionally Xenos has even LESS bandwidth available (already less than RSX) at 1080p than at 720p!
But that's not all. Like I said before, RSX probably has double the ROPs of Xenos-meaning in addition to having close to double the memory bandwidth, it can output double the pixels every second. What do higher resolutions need besides more memory bandwidth? Oh, that's right MORE PIXELS.

While I actually do agree with his assertion that this generation should stick to 720p, there's no question that RSX is better equipped to handle it. That's why quite a few games actually do use 1080p. The 360 has an option in it's menu to set the output to 1080p, but all that's actually doing is taking the game's 720p output and running it through a scaler* to get to 1080p-the game isn't actually running at 1080p. In fact not all 360 games even hit 720p. Several have used 1024x720 (like Project Gotham Racing) and had the scaler up that to the full 1280x720.

The reality is, several first generation Playstation 3 games were already running at 1080p while their second generation 360 counterparts were at 720p. An example is Marvel: Ultimate Alliance. At 1080p, the frame rate on the PS3 version wasn't supposed to be as stable, but it had better special effects, and was pushing 2.25x the resolution of the 360 version.

I've suspected that as PS3 development matures, some developers may opt to run the PS3 version of the game at 1080p, since it can basically do that higher resolution for "free" (since the game will be developed for the lowest common denominator-ie the 360-although I suspect most consumers won't understand that the 360's 1080p menu setting isn't actually 1080p). Overall though, I don't think RSX is really powerful enough to push that resolution, not without massively sacrificing other graphical elements. I do think 720p will be the norm this generation-and should be.

Next he baselessly claims that Xenos is the "first step to the future..." blah blah, and RSX is "far behind". Wow, so being more powerful is far behind. Okay.

(*This hardware scaler is one area I think Microsoft did a great job. Basically it means developers don't have to do much to set up various resolutions like 720p and 1080i, the scaler just takes the 720p output, and scales it to whatever resolution the user selects. The Playstation 3, like every other platform, the developer has to actually code for the different resolutions. That's caused an issue for early games that didn't support 1080i, as some people own early HDTVs that only support that resolution.)

OS: Next he claims the PS3's OS takes up more RAM and CPU power. No source for this one of course. The claim that the OS takes up 32MB of video RAM is especially strange. Doing what, exactly? Why would an OS store part of itself in video RAM? His 360 figures naturally don't jive with what I've seen on tech sites either.

And even if this were true, since the PS3 doesn't have as much overhead for networking stuff, I have a hard time believing Sony can't get it's memory usage lower, and regardless, PS3 has more CPU power to start with even if it was wasting more (and again, the assertion that the OS is running on an SPE is just silly).

Blu-Ray: Let's see how he manages to lie about this. Ah, PS3 can only have 256MB of textures at most! Oh, except that's a lie. RSX can use the entire 512MB available for textures, same as the 360.

Next he goes on about various games having long play times on DVD. Well duh, there are NES and Gameboy games with super long play times.

My personal thought is as a GAME MACHINE Sony would have been better off releasing PS3 without Blu Ray. They could have launched a year earlier, at probably $400 or less, but with the same better hardware (plus the better reliability, larger installed base, etc.). The 360 couldn't have stood up against that.

But that said, there's no question Blu Ray is an advantage. Blue Dragon on the 360 already uses three discs, and it won't be the last. And creativity is going to be stifled somewhat by a lack of storage space, as developers will have to shoot for 9GB max for most projects. His assertion that it doesn't benefit the consumer is a flat out lie.

Then he sums it up, claiming the 360 won three of his four categories (even though it didn't win any, here in the real world).

And of course his idiotic comparison failed to mention the Playstation 3's standard hard drive. 360 games have to be coded without the assumption that a drive won't be present. That gives PS3 developers the chance to use virtual memory among other benefits, effectively giving the PS3 more RAM and faster load times. (Oh, and that hard drive is even a normal off the shelf drive, easily user replaceable.)

And of course it fails to mention two huge benefits-the PS3 runs dramatically quieter than the 360. It's a night and day difference, despite the PS3 being more powerful/using larger chips. So far it's failure rate is beneath the radar-no mass failures at all, versus a 33%+ failure rate according to analysts on the 360.

But oh no, fanboi here thinks the 360 WINZ DUDEZ!! !

How does someone delude themselves so much? And to what end? If you own a 360 and can't afford both systems, just enjoy your 360, no reason to make up random stuff.

Yeah, that's convinced me

So many things wrong with that I wouldn't know where to start. It's kind of funny that someone has the time to make that though. I mean WHY?!? Why waste your time making up random videos like that?

Oh, what the heck, let's go through this. Their first "chapter"-yeah, duh it's not selling as well as Sony would no doubt like. That isn't causing Sony to lose money-they lose hundreds on each Playstation 3 produced. They're selling it at a big loss. It is massively outselling the 360 in Japan though. It overtook the 360's year lead almost instantly. Though we don't have figures yet, anecdotally it seems to have really picked up from the price cut, combined with it's library starting to fill out. This person evidently doesn't realize the 360 isn't selling as well as expected either, and many analysts say the PS3 is doing better than the 360 last year.
And the Wii is outselling both...which hardly means it's the best system.

"Chapter 2": This is even more ludicrous. Of COURSE those performance numbers are theoretical, they are on the 360 too. He randomly pulls some other number out of the air, and completely misses the point-it's MORE POWERFUL. We all know neither will get close to theoretical numbers, and no one really cares what those numbers are. What we care about is how it will affect GAMES down the line. Both CPUs are actually extremely inefficient on branching code by the way.

Then he pulls out this random "remember, the Playstation 2 has twice the FLOPS as the X-Box, and it didn't make the PS2 perform better!" Well no kidding. If he actually knew anything about these systems architecture, he'd know the Playstation 2 DOES NOT HAVE HARDWARE T&L. The X-Box does. On the Playstation 2, the vector units are used largely to set up polygons. The X-Box uses it's Geforce 3 GPU, which has a theoretical 80 Gigaflops. Not to mention it's a MASSIVELY MORE MODERN GPU.

You can't just pull random numbers up and compare them like that without actually understanding the hardware.

He then goes on, supposedly going to tell us which console wins based on "real world performance", and instead goes on about how Cell is more complex to code for. Well no kidding. We know that. That doesn't change the fact that it's a more powerful chip. It's just plain got more execution hardware on it. (I'm leaving out that he's making assertions that aren't backed up-improbable things like one of the SPEs is used to run the OS-the OS would be run on the PPE, not an SPE, and of course ignores that the 360 is ALSO using resources to run the OS. Duh!). We get screen after screen of this ridiculousness. This idiot needs to read a real hardware site if he actually wants to learn about CPU architecture. But he doesn't care, he's just a fanboi who wants "his" team to win.

Next he says "The PS3 only has 512MB L2 cache, the 360 has 1MB!" Yeah, no kidding moron. That's for that ONE CPU. The 360 has 3 CPUS sharing that 1MB? Remember? Each SPE has it's own 256KB. That's 2.25x as much cache/scratch RAM on the PS3 as 360, if we want to randomly count numbers.

Next he says Microsoft made a "better decision for developers..." because of the identical cores. Maybe, honestly both of these CPUs are pretty limited next to a PC's CPU. But regardless, the PS3 has a lot more power there to tap, and won't be maxed out as fast...and plus, presumably as gamers we care about the final results, not necessarily how fast developers can get the most out of the hardware.
And also, regarding the asymetric thing, both Intel and AMD have shown interest in moving towards that, with specialized cores for different uses. So that's something that's going to come sooner or later for developers regardless.

Chapter 3": Next up we get confused ranting about GPUs. First he compares it to a completely different architecture. Apparently not realizing you can't just count shaders. Shaders are just execution units-basically specialized CPU cores, and just like CPUs, you can't just count the number of CPUs there and declare a winner. Heck, the Geforce 8800GTX "only" has 128 shaders, versus 320 on ATi's new part-and of course the 8800 is easily more powerful. Most of ATi's shaders on their new part aren't nearly as advanced as those in the 8800-they can't do as much work. (Plus he's once again randomly throwing out theoretical performance numbers, ignoring what we actually care about-GAME PERFORMANCE.)

Next we get a flat out lie-360's Xenos is "far more advanced...much more performance!", and throws out more random numbers (which he's uncritical of-of course-despite just attacking RSX's numbers as theoretical). Then he goes on about Xenos' "2 advanced effects it does for free!"...one of which actually isn't correct (HDR), and the other of which is only partially true (anti aliasing).

Xenos is actually comprised of two pieces-the main GPU has been estimated at around 200 million transistors-that's INCLUDING the system's memory controller. It's at least 50% less than RSX, which should tell you something right there (same with Cell versus Xenon). That "free" effect is because the second chip contains hard wired anti aliasing effects. It's not really free though, because using it eats up memory bandwidth, which is at a premium. Turn it on? A frame can't fit entirely in that 10MB cache. Turn it up to 4x? It has to hit main system RAM even harder.
Then he claims that embedded RAM is "revolutionary". Well, no, actually it's not. The Playstation 2 had 4MB, the Gamecube 3MB, and the Playstation 3 has it's own separate video RAM, like a PC. As soon as you fill up that 10MB buffer, you're hitting main RAM again-and a single frame with AA turned on already spills in to main RAM as I recall.
He then goes on about the embedded RAM's memory bandwidth-which as I just said, helps alleviate strain on main system RAM, but doesn't even fit a single frame with any AA turned on.

There are some very bizarre situations you could find, where Xenos would outperform RSX...if you somehow got it to process more vertex than pixel data. But the reality is, that's not how today's code is written and developers have no reason to write strange code like that for the Playstation 3, when they know exactly what kind of hardware they're writing for. Real world, RSX just plain does more operations per second (it also quite likely has double the number of ROPs as Xenos does, though that info has never been publicly revealed to me knowledge, but Nvidia's 7900 series GPU RSX is based on has 16 ROPs versus 8 on Xenos). (He also actually says RSX has more vertex shaders than it really does, but wouldn't comprehend that Xenos shaders DO LESS WORK than RSX's). Yeah, a unified shader architecture helps prevent hardware from sitting unused...but that's ignoring three huge issues-Xenos can't just arbitrarily devote any shaders to pixel or vertex data, it can only split it's shaders into three groups, each doing either one or the other. Second, developers KNOW WHAT HARDWARE is in the box, and will obviously code to minimize wasted time. Third, RSX just plain does more operations (and probably has double the ROPs)-so even if it actually were less efficient, it's got more brute force to make up for it.

Then he makes the completely insane assertion that Xenos would be better than RSX at 1080p. HUH!?! His "proof" is that "inferior GPUs always show their weaknesses at higher resolutions...". Well, that's nice, but RSX is only "inferior" because you said so.

In reality, RSX has over double the bandwith of Xenos, excluding that 10MB cache (I'll get to that...). Guess what higher resolutions need...that's right, MORE BANDWIDTH! Gee, RSX wins there. But Xenos does have that 10MB cache that helps alleviate strain on main system RAM-but since it can't fit an entire frame in the cache, it still ends up worse on bandwidth. And guess what happens when you turn up the resolution? EVEN LESS OF THE FRAME fits in the cache! That means that proportionally Xenos has even LESS bandwidth available (already less than RSX) at 1080p than at 720p!
But that's not all. Like I said before, RSX probably has double the ROPs of Xenos-meaning in addition to having close to double the memory bandwidth, it can output double the pixels every second. What do higher resolutions need besides more memory bandwidth? Oh, that's right MORE PIXELS.

While I actually do agree with his assertion that this generation should stick to 720p, there's no question that RSX is better equipped to handle it. That's why quite a few games actually do use 1080p. The 360 has an option in it's menu to set the output to 1080p, but all that's actually doing is taking the game's 720p output and running it through a scaler* to get to 1080p-the game isn't actually running at 1080p. In fact not all 360 games even hit 720p. Several have used 1024x720 (like Project Gotham Racing) and had the scaler up that to the full 1280x720.

The reality is, several first generation Playstation 3 games were already running at 1080p while their second generation 360 counterparts were at 720p. An example is Marvel: Ultimate Alliance. At 1080p, the frame rate on the PS3 version wasn't supposed to be as stable, but it had better special effects, and was pushing 2.25x the resolution of the 360 version.

I've suspected that as PS3 development matures, some developers may opt to run the PS3 version of the game at 1080p, since it can basically do that higher resolution for "free" (since the game will be developed for the lowest common denominator-ie the 360-although I suspect most consumers won't understand that the 360's 1080p menu setting isn't actually 1080p). Overall though, I don't think RSX is really powerful enough to push that resolution, not without massively sacrificing other graphical elements. I do think 720p will be the norm this generation-and should be.

Next he baselessly claims that Xenos is the "first step to the future..." blah blah, and RSX is "far behind". Wow, so being more powerful is far behind. Okay.

(*This hardware scaler is one area I think Microsoft did a great job. Basically it means developers don't have to do much to set up various resolutions like 720p and 1080i, the scaler just takes the 720p output, and scales it to whatever resolution the user selects. The Playstation 3, like every other platform, the developer has to actually code for the different resolutions. That's caused an issue for early games that didn't support 1080i, as some people own early HDTVs that only support that resolution.)

OS: Next he claims the PS3's OS takes up more RAM and CPU power. No source for this one of course. The claim that the OS takes up 32MB of video RAM is especially strange. Doing what, exactly? Why would an OS store part of itself in video RAM? His 360 figures naturally don't jive with what I've seen on tech sites either.

And even if this were true, since the PS3 doesn't have as much overhead for networking stuff, I have a hard time believing Sony can't get it's memory usage lower, and regardless, PS3 has more CPU power to start with even if it was wasting more (and again, the assertion that the OS is running on an SPE is just silly).

Blu-Ray: Let's see how he manages to lie about this. Ah, PS3 can only have 256MB of textures at most! Oh, except that's a lie. RSX can use the entire 512MB available for textures, same as the 360.

Next he goes on about various games having long play times on DVD. Well duh, there are NES and Gameboy games with super long play times.

My personal thought is as a GAME MACHINE Sony would have been better off releasing PS3 without Blu Ray. They could have launched a year earlier, at probably $400 or less, but with the same better hardware (plus the better reliability, larger installed base, etc.). The 360 couldn't have stood up against that.

But that said, there's no question Blu Ray is an advantage. Blue Dragon on the 360 already uses three discs, and it won't be the last. And creativity is going to be stifled somewhat by a lack of storage space, as developers will have to shoot for 9GB max for most projects. His assertion that it doesn't benefit the consumer is a flat out lie.

Then he sums it up, claiming the 360 won three of his four categories (even though it didn't win any, here in the real world).

And of course his idiotic comparison failed to mention the Playstation 3's standard hard drive. 360 games have to be coded without the assumption that a drive won't be present. That gives PS3 developers the chance to use virtual memory among other benefits, effectively giving the PS3 more RAM and faster load times. (Oh, and that hard drive is even a normal off the shelf drive, easily user replaceable.)

And of course it fails to mention two huge benefits-the PS3 runs dramatically quieter than the 360. It's a night and day difference, despite the PS3 being more powerful/using larger chips. So far it's failure rate is beneath the radar-no mass failures at all, versus a 33%+ failure rate according to analysts on the 360.

But oh no, fanboi here thinks the 360 WINZ DUDEZ!! !

How does someone delude themselves so much? And to what end? If you own a 360 and can't afford both systems, just enjoy your 360, no reason to make up random stuff.

wanna post that a fifth time?

by the way, i'm not an xbox fanboy so i'm not actually putting any effort into this....i just hate sony because the new system is overpriced crap on a stick.



i'm a nintendo fanboy and nintendo's games aren't about the raw power of the system as much as solid games that are fun...which they've succeeded in achieving with zelda, wii sports, super paper mario....metroid prime 3 is about to drop and destroy everything.


and sony stole the six axis from nintendo's idea for motion sensing.


not to mention getting rid of force-feedback.

The PS3 is the most high quality, traditional console out of all the systems. The force-feedback, in my opinion, was no big loss, and they're going to make a new controller with both motion sensing and vibration soon enough. We'll see some good games that utilise the motion sensing in the second generation.

Wii Sports and Super Paper Mario really don't compare to good shooting games, which are a minority for the Wii. It's really about choice- between the PS3 and Wii; do you want advanced games with tough A.I, which present a greater and more intense game experience in the traditional sense than previous generations, or games which don't last for very long, and wear you out? Are you really going to play traditional games on the Wii, which has inferior processing capabilities, or more conventional games on the other consoles? The Wii cannot possibly have games with as much depth and advanced physics as Oblivion, Resistance, or Grand Theft Auto IV, can it? The Wii has last gen hardware, but with an added way to play games. I don't find that appealing, I prefer playing actual next generation games on my PS3.

Solid games which are fun? Oblivion's much more fun than Wii Sports.

Here is the hardware specs on all of the next gen consoles.

XBOX 360- 3.2Ghz PowerPC Processor, 512MB Dual-Channel Ram
PS3- 2.5 Ghz IBM 1500 Processor, 384Mb SDRAM
Wii- 733Mhz ATI Hollywood Processor, 88MB PCI100 SDRAM

PowerPC, FTW!

That's bogus. The Playstation 3's Cell CPU also runs at 3.2 GHz, and has considerably more execution hardware than 360's Xenon. Xenon has 3 cores derived from the PowerPC architecture (but more privative than a modern PowerPC chip), and Cell has one almost identical core, and 7 SPE cores that are a somewhat different design.

Both systems have the exact same amount of RAM. 512MB.

Where are you coming up with this stuff?



You do of course realize that all three systems use PowerPC based CPUs, right?

By the way guys, the multiple posts last night weren't my fault. The forum was acting bizarre, and I can't delete them now.

And I'll say again to the "Wii concentrates on fun games" thing-that's entirely in your opinion. I'm MUCH more interested in the true next generation games coming out on PC, 360, and PS3 (and even stuff in that vein that was released on the X-Box 1) than anything that's been announced for the Wii. You've got every right to enjoy those games, but that doesn't mean they're "better" than the stuff hitting the other platforms. For me, the kind of non-linear, deep gameplay that's possible with more processing power is much for fun than the short, communal style mini-games that dominate on the Wii. I have zero interest in those, and not enough interest in the "big" games on the system to buy one right now. (Super Paper Mario is the only game on the system I'd consider paying full price for-I'd probably own it if it had been released on a system I owned.)

Wrong.




CPU-

Playstation 3- 256 MB XDR @ 3,2 GHz, 256 MB GDDR3 @ 700 MHz, GPU can access CPU memory.
X360- 512 MB GDDR3 @ 700 MHz shared between CPU & GPU, 10 MB EDRAM GPU frame buffer memory
Wii- 729 MHz PowerPC based IBM "Broadway"

GPU-

Playstation 3- 550 MHz RSX (based on NVIDIA G70 architecture)
X360- 500 MHz codenamed "Xenos" (ATI custom design)
Wii- 243 MHz ATI "Hollywood"

And it's kind of pointless even putting the Wii into these comparisons, since it's much worse off than the clock speeds would suggest. I mean at best it's roughly equal to the X-Box 1, about 50% better than a Gamecube. Developers largely don't even think it's that good (no hard drive, no programmable shaders, etc.)

And skafather84, calling the PS3 "overpriced crap" is just fanboism. That has no meaning, just random name calling based on nothing.

G4 Gamers Television had a special on the next gen consoles a couple of months ago.

Did you know that Sony and Nintendio were working on a 32-Bit CD console back in the
late 1980`s and early 1990`s. Nintendio decided to drop the whole thing because Sony
wanted too much of the share saying that the compact disc part of the project surpasses
the Programming part. Sony than stole nintendios prototype and came out w/ the Playstation.