Why does Sony create such wacky architectures?

...

First you still assume he basically single-handendly designed both HW and OS and API for PS2, PSX and PS3
No, Kutaragi dictates what he wants built, poor Sony goons deliver what their boss asked, no questions asked. That's the way things work in organizations like Sony.

You can list all the exams he took in his university life and all the experience he might have accumulated in all these years ( he is not 20-25 )?
I don't remember Kutaragi claiming to code any game in his life, not even a tetris or blackjack. Beside, he went to school 35 years ago, when he was playing with transister circuits and never even touched a computer...

Now you say "he knows how to maximize any circuitry"... well but you also said that the whole PS2 HW ( both EE and GS ) is architecturally flawed and arguably a good design over-all in terms of elegance.
Kutaragi probably is a guru in circuit optimization. But I doubt he studied anything about systems architecture, and it shows on his works..

Also, going by your logic, Shigeru Miyamoto is to be held responsible of every small graphical glitch/flaw present in SM64...
Miyamoto is not a hardware man like Kutaragi, he is a coder. A completely different analogy.

The architecture makes much more sense than the Saturn architecture
I will insist that the opposite is true.

ideally is a direct extension of the PSX HW which people loved...
PSX2 resembles nothing like PSX1.

If I gave you a nice ARM7 CPU ( like the GBA CPU ) and I told you "now build me an OS with Windows like complexity, GUI, multi-tasking, etc..." and I gave you a compiler with BAREBONE libraries...
And I would attempt to port Linux package to CPU of your choice instead of writing from scratch instead...

you are basically calling Toshiba and IBM's as not having any balls and saying in the matter...
Well, I am pretty sure IBM people are pretty comfortable with something like CELL; the rest of world isn't.

Are you really telling me that Sony, IBM and Toshiba are not capable of providing an HLSL, a good OS and API support?
Something like CELL requires a totally different way of thinking to properly utilize, which doesn't come easy.
 
Panajev2001a said:
Again, you are declairing PS3 doomed/piece of crap 2 years before it hits the market,

Doomed and POC are 2 different problems. The ps3 is the least candidate of the next generation to be doomed, since it will be compatible with the ps1.

POC depends on the excution of the project.
 
Miyamoto is not a hardware man like Kutaragi, he is a coder. A completely different analogy.
Considering that he's a designer and not a coder, his analogy is very much correct.

I will insist that the opposite is true.
Are you guys talking about the same Saturn here? The one that had second processor tacked on in the last moment so that it can at least attempt to compete with PSX?
 
Re: ...

Deadmeat - can we get a ban for stupidity?

DeadmeatGA said:
No, Kutaragi dictates what he wants built, poor Sony goons deliver what their boss asked, no questions asked. That's the way things work in organizations like Sony.

He's one of many visionaries. He's an incredibaly farsighted and intellectual fellow whose ideology of the future is spot on when compared to that proposed by other visionaries such as Michio Kaku - perhaps you should do some background research...

I don't remember Kutaragi claiming to code any game in his life, not even a tetris or blackjack. Beside, he went to school 35 years ago, when he was playing with transister circuits and never even touched a computer...

When Gary Tarolli or David Kirk last go to school? Age means nothing - only a biased POS like you would raise it as a factor.

Kutaragi probably is a guru in circuit optimization. But I doubt he studied anything about systems architecture, and it shows on his works..

:rolleyes:

PSX2 resembles nothing like PSX1.

As stated, the PS2 is nothing but the PSX on steroids. The GS is basically the (PSX) * 16, the EE is an evolutionary increase. as previously stated.

And I would attempt to port Linux package to CPU of your choice instead of writing from scratch instead...

OMG! Sony and Mishushita are working together on a Linux based OS for home consumer devices!! And catch this, Okomoto - SCE's head of R&D IIRC - gave a presentation that showed the Cell OS to be based of Linux!!

Whoa...



Why are you here? I've seen you before, your a piece of shit who goes around instigating trouble. Please remove yourself before you hurt this board even more that it's already suffered lately. Thanks.
 
I couldn't say whether the PS2 should be considered exotic or not, but its design certainly didn't yeild impressive performance. PS2 graphics are a step back from even Dreamcast graphics, and that is inexcusable considering the PS2 is 50% more expensive and 15 months newer.

With the advent of Dreamcast, I thought suffering through sub-480p graphics was a thing of the Saturn/PSX/N64-era past. Unfortunately, Sony's PS2 brought back the ugliness of aggravating flicker from interlaced output and the annoying lower detail and letterboxed borders of 448 res in many titles (even in the big-name games like FFX, MGS2, ICO, SH2, J&D, etc). Coupling that along with its incompatibility to superior displays like VGA and HDTV in all but a few games, I'm simply not able to display PS2 games as attractively as I can with all my Dreamcast and Xbox titles, and many GameCube games.

The marginal improvement to poly counts and lighting the PS2 gives over the Dreamcast doesn't even come close to offsetting its profound downgrade to detail clarity and image stability. In a comparison between the two, my eyes immediately pick up on the more vibrant, sharper, and clearer display in which my Dreamcast games run - not on some subtle increases in sophistication the PS2 can add to some graphic areas, things that take a backseat anyway to more influential things like art direction and programming polish. Smart programming, modelling, and design can help compensate for things like geometry and lighting, but the sharper and more vibrant display from progressive scan and consequently VGA/HDTV is a global effect that only comes from better output.
 
Whats the advantage of creating these costly exotic architectures when similar performance can be had by using "off the shelf" components?

Actually, sooner or later, the "off the shelf" components will use exotic architectures anyway. What's wrong about being early ?
 
Vince....

Deadmeat - can we get a ban for stupidity?
Stupidity is in the eyes of beholders.

He's one of many visionaries.
The problem with most "visionaries" is that they never got proper training in fields they represent, Bill Gates(A Havard freshman dropout), Kutaragi Ken(No formal training in computer engineering or development), Steve Jobs(What's his credential anyway???), etc while the real visionaries don't gain any fame for their work.

He's an incredibaly farsighted and intellectual fellow whose ideology of the future is spot on
Having seen the vision of something far more impacting than CELL, I am not impressed. I swear I toyed with almost every computing architecture published until I hit on something that offered an ultimate scalability without comprimising codability(Call it the Borg of Computers). Kutaragi san is where I was at back in 1999 and I pity him, I too was once thinking something similar to CELL but soon realized it wasn't workable and moved on.

As stated, the PS2 is nothing but the PSX on steroids.
So you don't know anything. I repeat once again, PSX2 resembles nothing of PSX1, other than the embedded SOC for backward compatibility. It works on a completely different processing principle.

OMG! Sony and Mishushita are working together on a Linux based OS for home consumer devices!!
But I would not port Linux to CELL; Linux's kernel is too large to fit into CELL PU's cache and accessing the kernel image off-chip would be bandwidth taxing. PSX2 EE performance got killed by memory latency, and CELL aggregates the problem even more with larger objects.

And catch this, Okomoto - SCE's head of R&D IIRC - gave a presentation that showed the Cell OS to be based of Linux!!
More the reason that PSX3 won't be worth its R&D budget. CELL would clearly benefit from a custom microkernel than Linux.

Why are you here?
There are two sides to every coin.

your a piece of shit who goes around instigating trouble.
I have never launched a personal offense to you yet you offend me this way. I believe you owe me an apology.
 
Lazy8, I don't mean to argue but what do you base this off of? I'll admit that when the PS2 was initially released it couldn't hold a candle to the Dreamcast in terms of overall graphics. But this situation has changed for the most part, I'm even starting to see PS2 games with textures that look better than even the best Dreamcast games graphics wise. Like I said before, it takes time for developers to come to grips with hardware and be able to leverage it in a way that's truly impressive. People called me crazy when I said graphics on Xbox games initially didn't impress me, even SEGA ones. Everyone here knows I'm the biggest SEGA supporter you will see, but I'm not a blind follower that can't see reality in a situation. Games on the Xbox are only now starting to surprise me now that devs are finally learning how to use it and optimize for it unlike treating it like a PC that you need to develop for the minimum spec. Halo didn't impress me in the slightest, simply for the fact that I knew a PC version would eventually be coming out and I could play it at 1280*1024. Now PDO impresses me, but most games on Xbox so far don't impress me graphically.

Maybe Xbox games have better graphics than PS2 games on a regular basis yeah, but many PS2 games are starting to impress me graphics wise as of late. Games like ZOE2 and many other Konami games have caught my attention. The thing that impresses me the most is from a technical standpoint and in the art. Look at what Konami is getting out of the PS2, and that's all on 4MB of graphics memory. True that it is eDRAM, but still that's some really impressive graphics that I thought I'd never see. Game developers making PS2 games have now learned the ins and outs of the system and what it can and can't do. The artistic stuff that impresses me is the fact that certain devs are flexing the strengths of the PS2 hardware. For me, it all comes down to thinking how far a system can go in terms of graphics. I never thought the PS2 would be at the level it is today and I'm now eager to see how blown away I am by the architecture a year from now.

Next year I expect the Xbox to impress me very much now that devs take advantage of a thing called bump mapping, fill rate, and geometry, and huge pretty textures.

As for Dreamcast, I've seen the best of what it can do and some of you on here should know that. What the best DC had isn't nearly impressive as the best of PS2 has to offer so far.
 
wazoo said:
You answer yourself to your critics. The xbox is more costly and came almost 2 years later, and still most games do look comparable, some may even look better on the ps2, and people are still blown away by ps2 games (Konami games for reference) 3 years after the japanese launch. To be fair, you would need to imagine what would be the ps2 power if it launched with the xbox or what would be the xbox if it launched in Mars 2000.

Back in 2000 I could have made the same sentence about DC and PS2 :

"The PS2 is more costly and came almost 2 years later, and still most games do look comparable, some may even look better on the DC, and people are still blown away by DC games (Sega games for reference) 3 years after the japanese launch. To be fair, you would need to imagine what would be the DC power if it launched with the PS2 or what would be the PS2 if it launched in November 98"

Funny isn't it ? it couldn't be more true :D
So I don't think that proves anything about the need of a tricky architecture because DC was an exemple of using of the shelves components in a standard architecture.
 
halo, doa3, wreckless, PGR, Splinter Cell, motoGP, Fever,... 5h343 there are so many games that look far better than anything on PS2 it's ridiculous to compare them.. It boggles my mind that one couldn't recognize this. Even hardcore PS2 fanboys usually concede it. PS2 has strong points, and there is an argument over which console is better, but it certainly is not in the graphics department.. have you been wearing your glasses lately?
 
Lazy, it´d be best if you limit your peotic skills to Shenmue. It is very questionable where exactly DC outdoes PS2, even in terms of textures PS2 is showing in several upcoming games textures that outdo what DC could handle.

There´s also games that DC would have a lot of trouble even trying to mimmick, games like GT3, MGS2 and FFX I doubt would be able to be reproduced perfectly, and it could be argued that those are merely first generation games. Going from SoA to FFX is a serious jump in visual quality. How about reproducing Jak&Daxter or the polygonal complexity of Ratchet & Clank?

DC showed the absolute best it could do in Shenmue 2, and it is pretty clear that its limitations are very apparent.
 
The NAOMI 2 technology was available the same time as the PS2's and the costs were VERY similiar. I've already explained this before here in this forum. SONY had money to burn. SEGA didn't. DC was already out when the N2 design was complete so SEGA couldn't just release a DC 1.5. They'd be killing themselves considering their financial situation. However since N2 is somewhat backwards compatible with DC, SEGA could've released N2 as a DC1.5 if they had the money which they didn't.

That said, looking at software development costs/difficulty, image quality, efficiency, etc., which hardware design was better? The PS2/System 246 or NAOMI 2?
 
I find that to be a mere excuse. Regardlesss of Sega´s situation, NAOMI 2 is not DC, it´s as simple as that, and PS2 is clearly the most powerfull one when comparing the best looking games on both plattforms.
 
Almasy said:
I find that to be a mere excuse. Regardlesss of Sega´s situation, NAOMI 2 is not DC, it´s as simple as that, and PS2 is clearly the most powerfull one when comparing the best looking games on both plattforms.

If you say so :rolleyes:

BTW this thread is related to hardware design choice and why SONY chose to design the PS2/System 246 the way they did at that time. There were other design choices at the time that were more efficient and arguably cheaper too. ;)

The DC came out 18 months before PS2 so of course the PS2 is going to be more powerful regardless whether or not it used "off the shelf" parts...
 
I swear I toyed with almost every computing architecture published until I hit on something that offered an ultimate scalability without comprimising codability(Call it the Borg of Computers). Kutaragi san is where I was at back in 1999 and I pity him, I too was once thinking something similar to CELL but soon realized it wasn't workable and moved on.
Oh, holy crap Deadmeat... Do you even realize how weak this sounds? Have you ever created or published any of your major accomplishments to back up those outlandish claims? Whenever I read something like this from you, I remember reading a letter in some crappy teen magazine where a guy pissed off at the whole world, claims that he would know how to better arrange governments, solve the world hunger and unemployment. Guess what the 'teen psychologist' answered him? He kindly explained that he should seek professional help (and he was dead serious about it)...
 
Re: ...

No, Kutaragi dictates what he wants built, poor Sony goons deliver what their boss asked, no questions asked. That's the way things work in organizations like Sony.

Uhm weird, I thought that guys like Okamoto ( and in Cell case also the Toshiba and especially the IBM guys, what do they know ? ) had a tiny little bit of saying... but of course YOU know better...

plus, he could say "hey I want SMT in that chip", so ? He could say that, but the implementation of it and the compiler and tools side are a huge thing to complete and guess what ? I do not think that Kutaragi himself handled a large scale VHDL/Verilog program in the last 5-6 years ( or more )...

My analogy with Miyamoto stands...

Like the famous director Arturo Toscanini once showed, the orchestra if well trained can work on its own ;)

I don't remember Kutaragi claiming to code any game in his life, not even a tetris or blackjack. Beside, he went to school 35 years ago, when he was playing with transister circuits and never even touched a computer...

What the heck is your point ? What do you know of the programming experience he has ? What do you know of what he coded ?

Plus it is funny you say he never touched a computer... tough to run logic simulators without touching the PC... was he using Jedi tricks ?

Kutaragi probably is a guru in circuit optimization. But I doubt he studied anything about systems architecture, and it shows on his works...

I seriously do not believe I am reading that... so you pass with Comp. Engineering degrees without understanding system architecture concepts ? You tell me anyone can be any good at optimizing a circuit withou understanding the big picture ? Are you telling me he only studied as a simple close minded electric engineer who doesn't know about computer architecture and theory of computing ?

and it shows on his works... again "his"... HE DID NOT LAYED OUT THE WHOLE PSX and PS2 ARCHITECTURES BY HIMSELF... THERE WERE LARGE TEAMS WORKING ON THEM... competent teams judging from the results...

and who are you to go down so much on PS2 and PSX... are you angry at the idea that a architecture designed with 1999 technological limitations still manage to kick some butts as games like Z.O.E. 2 and SH 3 show ( in 2003 ) ?

Are ylou unable to cope with the idea that bad documentation and lack of efficient high level libraries were to blame for the steep learning curve for PS2 development ?

Miyamoto is not a hardware man like Kutaragi, he is a coder. A completely different analogy.

His involvement with Nintendo games since the N64 days has been in a much different perspective than programming, the analogy holds wether you want to open your eyes or not

I will insist that the opposite is true.

Ok, now the ball is on your court, explains us how the Saturn was for its time a much better architecture in terms of performance and design...

PSX2 resembles nothing like PSX1.

Do you even bother reading what I post ? Did those basic spec sheets for the two consoles make you think ?

You know what is funny ? YOU were the one that was sustaining that position regarding PSX's GPU and PS2's GS... backpedaling without getting noticed must be a nice sport...

And I would attempt to port Linux package to CPU of your choice instead of writing from scratch instead...

And I would tell you the CPU I gave you is too slow for a simple port job and understanding of the underlying architecture and new code are needed...

Wow though, you're getting back in shape, by the tiome serious PS3 specs come out you should be all ready to tackle "opponents' arguments dodging", selective memory and selective quoting challenges... Avoiding the issue is one of your best abilities...

Well, I am pretty sure IBM people are pretty comfortable with something like CELL; the rest of world isn't.

hey buster, IBM is one of the company who IS designing the Cell OS and the APIs and programming tools related to Cell...

IBM's taks is also to make the rest of the world NOT TO HAVE to care TOO MUCH about Cell inner working...

Something like CELL requires a totally different way of thinking to properly utilize, which doesn't come easy.

Yes it is not easy... still I do not know, but given the size, resources ( HW and SW R&D wise ) and commitment of the conglomerate Sony, IBM Toshiba seems to have all the cards to put out an at least decent performance on this task...

Of course we all know your position even before they started revelaning anything...

Having seen the vision of something far more impacting than CELL, I am not impressed.

YOU not impressed by a technology SCE and PS3 related... what a SHOCK, let me call the news... :rolleyes:

I swear I toyed with almost every computing architecture published until I hit on something that offered an ultimate scalability without comprimising codability(Call it the Borg of Computers). Kutaragi san is where I was at back in 1999 and I pity him, I too was once thinking something similar to CELL but soon realized it wasn't workable and moved on.

care to illuminate us and/or present to the relevant patent your emminency has published ?

And again, you say Kutaragi IS where you were in 1999... you gotta commit man, either Kutaragi is the father of Cell technology or he is not...

I understand that in order to twist reality better is helatier to jump from a position to the other, I just happen to question the logic behind it...

So you don't know anything.

I tried to tell Vince, he is fighting against a "genius"... :rolleyes:

I repeat once again, PSX2 resembles nothing of PSX1, other than the embedded SOC for backward compatibility. It works on a completely different processing principle.

Just because it address the weaknesses of the 32-64 bits generation of 3D processing HW and developed/pioneered new technology to address the concerns that the "128 bits generation" brought...

Data sets have been increasing faster than the complexity of operations to be executed on each object and thus more and more bandwith is needed because we are focusing on moving data, not holding it locally... efficient processing also called for e-DRAM and not the complete move to UMA (hence the hybrid UMA model which also was followed for the Gamecube)...

As I posted there are similarities in the HW, PSX HAD specialized 3D processing HW ( GTE ) and Video decoding HW ( MJPEG ) all working together with the main RISC core to speed up oveall performance...

PS2 expanded this multiple DEDICATED ( yet flexible ) processing elements on a single chip concept as the EE shows similarities with the PSX CPU ( RISC Core + GTE + MJPEG decoder )...

If PSX programmers were presented off the bat with bad documentation and no efficient high level libraries and heavvy enphasys of ASM coding, you would have heard many different stories on PSX's supposed ease of programming...

By the same token if Saturn programmers were instantly given an OS with SMP support, efficient multi-threading implementation and good fast libraries that hid most of the underlying HW to the programmers and a good compiler and you would have heard many less bad stories about Saturn crazy HW...

But I would not port Linux to CELL; Linux's kernel is too large to fit into CELL PU's cache and accessing the kernel image off-chip would be bandwidth taxing. PSX2 EE performance got killed by memory latency, and CELL aggregates the problem even more with larger objects.

Of course they are not simply porting the Linux kernel as-is... I expect with the BILLIONS IBM has spent in internal Linux R&D ( Sony in the last 3-4 years got quite a bit of Linux experience too... ) to give them a bit more than that ;) There are parts of the Linux kernel that do not have to be modified too much and there are others that will require resources and expertise to be proted efficiently... again I expect Sony, IBM and Toshiba to have what it takes...

Still each APU ( 8 APUs per PE ) would have 128 KB of Local RAM (SRAM) and accessing the 64 MB of e-DRAM through that fast and fat 1,024 bits pipe is not THAT impossible...
 
The NAOMI 2 technology was available the same time as the PS2's and the costs were VERY similiar....There were other design choices at the time that were more efficient and arguably cheaper too.

And you expect Sony to drop their PS2 designs just like that? Would a PS2.5 compete well against Xbox? 8)
Yes,PS2 has been underwhelming but it is still cool considering its age.

I've already explained this before here in this forum. SONY had money to burn. SEGA didn't. DC was already out when the N2 design was complete so SEGA couldn't just release a DC 1.5. They'd be killing themselves considering their financial situation. However since N2 is somewhat backwards compatible with DC, SEGA could've released N2 as a DC1.5 if they had the money which they didn't.

Even if they have the money, releasing a DC1.5 would be a quick bye bye to that money. ;) Hey, Sony could have slapped a PS2.5, by adding another EE+GS pair to compete with Xbox in 2001, but they are wise enough not to do that. :devilish:
 
Deadmeat said:
I repeat once again, PSX2 resembles nothing of PSX1,
That's not exactly true.
Sound chip is a duplicate PSX sound chip.
Both use a Mips CPU.
Both have a movie decoder and transform processor coupled to the cpu.
Finally, according to your own words, the PS2 rasterizer is just 16x PSX. :p
(I still disagree on that, but just thought I'd mention it).

Yes, it's true PS2 is UMA styled and PSX wasn't really, but you can't argue based solely on that that they have no similiarities.

BTW this thread is related to hardware design choice and why SONY chose to design the PS2/System 246 the way they did at that time.
I am not entirely sure about this, but as far as I understand, 246 was more Namco's "design" then anything (if you can call it a design considering they practically changed nothing from a stock PS2). Either way, seeing it's arcade hw I am still at a loss why they didn't use 128mb configuration, which would have been a considerable improvement. I-32 GS probably wouldn't have been too hard to add either...
 
Back
Top