What would you do if you were in charge of Xbox 360 design?
- Thread starter shaderguy
- Start date
Titanio said:
Unfortunately, the Xbox was/is the victim of being the leader in a market where the lowest common denominator often wins out due to the financial model. Even if Xbox was the market leader (heh, they have a long way before they sniff that territory!), I doubt that would have changed much. In a market with 120M console units, the smart money for developers is to hit as many machines as possible. That means the most powerful machine gets under utlized. It stinks, but I think that has frequently been the case.
No doubts there. Europe frequently gets the short end of the stick. I mean, you still don't even have the PSP and Sony is getting all upset at the grey market :? I do not envy EU gamers. You guys put up with a lot of poor treatment. Obviously government standards would help
:cough::HD TV::cough: but for being the 2nd biggest market you general you guys do sure get the short end of the stick and get treated like the step child no one cares about. At least you get BF2 a few days early!I am going to play devils advocate here
Xbox 360, in my opinion, even with another year of development, would not be much better.
In general, significant leaps in performance result from process changes. This year we have seen the gradual shift from 110nm/130nm process to the 90nm process. We also saw Intel get burned in that transition. 65nm looks like it may hit the end of 2005 in small quantities from the market leaders but will be very limited in quantity. It is unknown what problems will creep up with 65nm parts. We have yet to even get our first 90nm GPU and it does not look like NV will have a flagship GPU on the 90nm process until 2006.
So MS is on the cutting edge right now. Note how both NV and ATI had GPUs come in at ~300M transistors at 90nm. One the balance of Die Size/Transistor/Frequency there is a fine balance. On the outside edge of that balance is Heat and Power Consumptions.
Basically, waiting another 12 months wont do much for MS. They are not going to get more transistors into their GPU or their CPU on the 90nm process--at least not significantly. You look at the "refreshes" of CPUs and GPUs and you are usually looking at 5-15% jump in performance.
That ain't gonna cut into the paper lead the PS3 has in FLOPs performance.
But lets say for the sake of arguement that MS could get 65nm at the end of 2006. And lets say they are able to build above and beyond on the cutting edge GPU design (it is already a design that is a good year or two away from the PC market). And lets say, amazingly, MS gets the same yields on first run 65nm chips as they are going to get at 90nm (wont ever happen). So lets say in transistor count and frequency bump their CPU gets close to the PS3 CPU in FLOPs. And their GPU ends up being a 350M logic transistors at 600MHz. And since we are playing with numbers lets say they go with 512MB of 256bit memory for ~46GB/s of system bandwidth.
Ok, so MS has a small lead in performance now. Still not ahead in all areas, but ahead overall. (Of course this is all fictional and assumes 65nm would be ready on the GPU and CPU and in good enough quantities).
So now what?
1. They are in the same position as last gen. Negative #1. They ship last. Negative #2. The extra power is under utilized. OUCH.
2. The thing is a freaking money pit! With a 300M transistor CPU and a 450M transistor GPU and 256bit memory the thing is going to be really expensive. The whole slogan about, "You pay a 95% premium for the last 5% of performance" deal. Then there will be heat... oh boy will there be heat. Lots of it.
Oh, and that beautiful plan of considating chips at the 65nm and 45nm process shrinks? Forget it. This system would be doomed to be expensive for the entire lifecycle.
I never see MS going back to that model. It has never worked for anyone in the history of game machines. MS took some tough blows to get into the market, and MS can absorb those. But they are not crazy intentionally.
It all goes back to the question of: Why have more potential power, that extra 5-25%, if it never gets used? Why not just make an effecient machine that is EASIER to develop for? Works for Nintendo. Worked for the PS1. The quality of the software is not dependant soly on the theoretical peak performance of any system. A well design, balanced system based on a holistic approach is better imo.
I think Xbox 360 and PS3 are both good holistic approaches coming from different angle and philosophies.
Anyhow, the above was a best case scenario. Trying to hit 65nm would be a nightmare. You are talking a 1 year turn around from the first 90nm GPU to the first 65nm GPU. Consoles are $300 home entertainment devices. I think sometimes we are expecting too much, too soon. This gen is really spoiling us.
Top of the line GPUs? Never in my dreams would a console GPU be comparable to the top end GPUs on the consumer market. I never thought they would have as much memory as a standard PC.
Anyhow, I do not see 2006 offering MS the time to make a big leap over what they have now. And any leap they did make would not benefit them.
Maybe in EU. I know in the US that things have slowed down--and games just look darn ugly compared to the PC and what the next gen offers. I barely touch my GCN anymore. I am ready for new features and better technology integration. I am ready for online to be standard in all games.
And looking at the games coming out this year for the PS/Xbox/GCN make me yawn. Some great titles, but not enough. Things are definately slowing down.
I guess in the end I do not seem them offering anything much better (if better at all) by waiting a year. So I would rather have the same thing 12mo early than 12mo later.
MS going early is better for them financially as most would agree. But I also think it is good for the industry. I think competition is good. I also think spicing up that last "slow year" is good for us hard core gamers. I am an early adopter. I played the PS2 when it first came out. ditto Xbox. Got a GCN because it worked with my VGA monitor.
I have been playing these things for like 5 years :? And the lack of power has really shown itself in the lack of AI, physics, and other areas. I am ready for consoles to pass PCs again
Of course you make some good points, I just don't think waiting a year is good for MS, or would benefit them. And as an anxious gamer, I do want to see what this next gen has to offer. Call me greedy!
wco81 said:
As I noted above, 2006 does not look to offer much hope in the manufacturing process. A lot of the industry is tied to the manufacturing processes. That is why we usually see new designs on smaller processes. New features take up a TON of realestate. So does more power. You can make a chip better over time on a process, but not a lot. And you are always fighting heat, power consumption, and yields. We do need to be realistic for a $300 device.
This is why Sony is on the 90nm process. We are not seeing Sony with a 2 CELL 500M transistor CPU.
The GPU market gives some clues though. Every 2 years we tend to get ~100% jump in performance at the top end on high end features. In the 1 year refresh in the middle we usually see a 10-20% jump in performance. There are exceptions (like when an IHV gets very aggressive for an expensive flagship part that sells in small volume), but overall chip makers are tied to a degree to the technologies they use.
I was surprised to hear both of the Xbox 360's main chips are on the 90nm process. MS and Sony are just hitting 90nm just right to hit the mainstream consumer market. If 65nm was realistic I think Sony would be shooting for it, but even they are not.
Just my opinion of course.
DuckThor Evil
Legend
Titanio said:
Well going to 65nm might be possible, but just maybe. I'm thinking that the machine would probably be about 2x more powerful, and when that is on the screen, it's not really that much, anyways MS can't afford Sony to have 6 months headstart this time, although it's possible or maybe even likely that Sony would also launch later if MS launches in nov 2006.
pc999 said:Welcome.
But do you really think that 4-6 + mouths would be enough to do that, especially if they already do have prototypes since some time ago?
Probably a solution like, already sugested, of adding a PPU (not by me) would creat a really beast and the transsistor count would not go up that much as in your sugestion.
Just my opinion, others are better than me to this.
At first thanks for the welcome.
About the specs,this is the kind of design i would have liked M.S to use as i think that it would give them the edge over Sony.
Off course,i could be wrong as i am not an expert of hardware.
Now,if the question is to know how they could have improved their current
design with additionnal months then i suggest:
*More EDRAM to avoid tiling and enable 4*fsaa at 1080i resolution
*An additionnal C.P.U (more likely a low power /cheap Intel or transmetta C.P.U ) which will act as the main C.P.U and deal with the A.I.Physics,animation and any fp intensive task would be let to the XeCPU.
Moreover this C.P.U could be used for a more efficient bacward compatibility as it would execute all the XCPU code without need for emulation.
The XeCPU could then be used to help the GPU for NV2A emulation.
Huh. I guess there'll be no "C2" then...
To answer the topic question: I'd make it more visually appealing. It looks just like, well, a box. Should be cooler somehow, maybe being able to change the face plates or whatever along these lines, high-tech look and many, many colours to choose from or combine.
EDIT:
PS3 looks just as bad, Revolution is the only one with nice design.
EDIT:
PS3 looks just as bad, Revolution is the only one with nice design.
Acert93 said:No doubts there. Europe frequently gets the short end of the stick. I mean, you still don't even have the PSP and Sony is getting all upset at the grey market :? I do not envy EU gamers. You guys put up with a lot of poor treatment. Obviously government standards would help
I disagree, Europe is not the 2nd biggest market. Europe is a bunch of mini markets,that requires significant cost for localization and time for each of these cultures. Europe is by far the hardest place to launch anything because of the many SKUs that are required. It's not as profitable as the NA or Japan and therefor not given priority. People act like Europe gets that shaft when in fact its just logic and good business.
croc hunter2
Newcomer
My take,
MS are trying to get a jump on sales next gen and this is a risk.
Risk factors as i see them
*short lifespan annoys some of the userbase (Euro)
*lose the power advantage they had this gen
*possibility people will hang on until the PS3 appears.
*$60 games (rumoured) may alienate the userbase even if the other 2 go with that pricing
*devs dev for Sony and MS get ports making the possible power diff more pronounced
Possible gains
*get the first of the next gen hype and therefore sales
*all so powerful hard to tell the difference between games therfore lose nothing on that front
*get to set prices
Things that are unlikely
*Any sort of presence in Japan
MS are trying to get a jump on sales next gen and this is a risk.
Risk factors as i see them
*short lifespan annoys some of the userbase (Euro)
*lose the power advantage they had this gen
*possibility people will hang on until the PS3 appears.
*$60 games (rumoured) may alienate the userbase even if the other 2 go with that pricing
*devs dev for Sony and MS get ports making the possible power diff more pronounced
Possible gains
*get the first of the next gen hype and therefore sales
*all so powerful hard to tell the difference between games therfore lose nothing on that front
*get to set prices
Things that are unlikely
*Any sort of presence in Japan
Logic and good business to charge this area some 50% markup on virtually everything?!
And despite that...How peculiar, that charging people more is less profitable. :?
Just get the fuckers to learn bloody english for god's sake!!! Enough with this dubbing and all!! Next step, we all get to learn japanese and play the original titles the way they're meant to be played.
I don't believe you! Localization means a few translations of documents and game (or menu) text - not a mammoth task. Can't see logistics being a problem either with EU offices, EU open market etc. Any links to how much more things cost to release to EU then elsewhere?_xxx_ said:
Shifty Geezer said:
I don't believe you've ever managed a business. Localization costs could mean alot of things.
Possibly including, and not limited to.
1. Administration costs to keep an office open for a particular country. (for returns and overseeing localization. )
2. Hire translators.
3. Redo voice/audio in game and cutscenes. )either outsourced or in-house studio)
4. Hire local actors. (sometimes union)
4. In-house programming to fix anything broken during localization. (debugging)
5. In-house testing / QA (Most Japanese can't read/understand dutch)
5. Unique packaging creation.
6. Law Dept. for local tax and corporation paperwork.
7. Im sure Im forgetting 10 more things.
Of course you could outsource any or all of these but then you're costs go up dramatically, hence the game price goes up. dramatically.
I think most people haven't managed an international business!Pozer said:
Normally there's an EU office, same as a US and Japanese office
Which'll be what, a few thousand?
same again
...more stuff...
At about $75 a game, you're suggesting it costs $25 more per disc to distribute in the EU? Yes there's more paperwork, but no more than a handful of office staff can cope with. Nothing to explain the needs for $millions more expenditure than selling to the US or Japan. Plus what about everything else being more expensive when they don't need to pay for actors, programmers etc.? TVs for example. Translate the manual and menus and that's it (plus decoding hardware).
Similar threads
