Gabe Nevell Next gen Console Video Interview

I'm not asking for your personal opinion.

Those games are insanely popular.

 

Yes, but popular games =! good games.

 

Bingo. Xbox 1 may have given PC devs a bit of a sweet tooth in this regard. The "culture shock" for lack of a better term, is to be expected.

 

CS was a mod which was made by Valve, its their project.

 

In that interview he admits he doesn't know much about multithreaded programming and then he begins to rant about him not asking for that and bla bla.. He sounds like an old man scared of new developments.

 

dudes gunna tear his hair out when pentium 5 arrives

 

I always thought the original CS wasn't made by valve, and that they then bought it to use it in HL2.

 

You are sort of correct. It wasn't originally made by valve but it was purchased and turned into an official mod.

From here.

 

I dunno, but I thought he said something like this... Sony made the PS3 architecture this way so that the game/engine code written to make use of the SPEs is far too hard to port to any other platform. Sony also wants to build the image that they'll have monopoly with this gen again, so that all developers will take the PS3 as their primary platform. This would result in developers with limited budgets to be unable to port their games to other platforms, which is exactly what Sony wants; and it will also strengthen their monopoly as well. At least as I got it...

 

Do you Laa-Yosh believe that? I hate to hear this guy cry all the time. Make some games. Speaking your mind is one thing, but shouldn't he give us some positives of what's he going to do to tackle the problem. Epic has. Why can't he?

 

I think it would be wrong to characterise the design choices as being motivated (or motivated solely) by a desire to limit cross-platform porting, or the success of that. If they really wanted to obfuscate PS3 development relative to other systems, they could have done a lot more to do that. But yeah, it is a natural consequence of the design that ports may be more difficult, and I'm sure Sony are not unhappy about that. And assuming a strong position in the market, they can get away with it.

But you know, it's not like porting has been a walk in the park before. Pentium code doesn't port very readily to a EE I imagine either, or vice versa.

 

And doesn't Gabe know that IBM and Toshiba had something to do with making the CELL too? Is he telling me that the CELL is only going to be used for the PS3? He doesn't know that STI has already sold it to the Mercury company? It's also going to be in TVs too Gabe!

 

oh so so! sorry, i though i didnt had to quote what Gabe said about code difficulty's to port over to other platforms. And neither was my intention to. What i wrote in the 2nd paragraph was part of a personal opinion about sony's decision regarding Cell. Sorry if i wrote it so that it could misslead some ppl interpreting it as if it was some comment about gabe's interview.

 

What is a "gaming chip" exactly? Thats the first time Ive ever seen that term used. Im pretty sure it isnt x86 in all of its legacy. Thats what PC developers are certainly accustomed to, but its definitely not specifically aimed towards gaming application. And it sure as hell isnt the Xcpu, which last I checked is a multi-core processor and will hardly be a cakewalk to develop for in its own right. So whats a gaming chip?

 

*looks at you all weird like and says....

Its whatever you want it to be O.O

It is what it is, trying to say that Sony is trying to gear it towards something seems to be the current misconception. That seems to be Sonys use for the CELL, the question is, what will IBM and Toshiba use it for? Also..as mckmas8808 said, its already going to be used for a company named Mercury...which from what i've seen work with advance imaging and things of that nature.

 

More here.

Yep more than the PS3. I think the CELL chip will still be great for videogames too though.

Link

http://www.businessweek.com/technology/content/aug2005/tc20050825_1931_tc024.htm

 

I would just point out that Windows Update has had the ability to offer driver updates for years, and has in the past occasionally offered video driver updates.

But the problem has been that video drivers have never been completely stable. One version would break some apps, and the next would fix those, but break some others. In fact Carmack was bitching about that in his recent keynote. This is the big reason Windows Update rarely if ever offers any driver updates.

Imagine the nightmare you'd have if you ran Half-Life 3, it autoupdated your video drivers, and broke Doom 4. Then you ran Doom 4, it autoupdated your video drivers, and broke Battlefield 3. Then you ran Battlefield 3, it autoupdated your video drivers, and your machine crashes now all the time on boot.

One of the big changes in Longhorn is the new driver model (LDDM), which moves a bunch of the hairy bits of the 3D driver out of kernel mode and adds virtual memory, GPU virtualizing and scheduling, and other stuff which should radically improve stability, reduce driver complexity, and as native WGF hardware comes, much better performance.

If anything, making the driver model better and easier for 3D IHVs to write stable and fast drivers will do much more for reducing customer support issues than automatic patching would, IMHO.

 

Thank you, thats exactly what I was trying to get at. There's no such thing as a "gaming chip" per se. A chip gets used for certain applications or it doesn't. We know for sure Cell is being used to develop games, so that makes it as much of a "gaming chip" as anything else. But silly people with even sillier agendas would like to portray the Cell design as Sony's way of "bullying" game developers for their own subversive means. Its ridiculous nonsense.

 

It's a kind of bunk argument though isn't it? For what Cell could make porting from PS3 harder, it makes porting 'to' PS3 10x worse, so it would be more like Sony shooting their own foot if this was their primary intent.
That said I'm more under impression Sony is trying to software abstract everything to hell even if to the point of neutering the hardware, so we could well be facing years of fighting their driver interfaces, not the hardware.
And I don't really mean that as a good thing.

Which is a great reason it shouldn't have taken MS 10 years to come up with an improved driver model.