PDA

View Full Version : Re: Intel talking to Nintendo, Microsoft, Sony about TFLOP chipsfor NEXT generation game systems after being locked out of current-gen


Bill Davidsen
December 21st 07, 10:24 PM
AirRaid wrote:
> Chip company has spoken to hardware manufacturers, touting new
> processors
>
> Intel, the company responsible for the majority of processors used in
> desktop PCs, has spoken to Sony, Microsoft and Nintendo about what
> chips might power the next round of consoles.
>
>
> http://www.gamesindustry.biz/content_page.php?aid=31534
>
>
> "Intel, the company who produce countless components and processors
> that hide away in millions of PCs across the world, has begun talks
> already with Microsoft, Nintendo and Sony about providing parts for
> the next-generation of consoles. That's right folks; you simply can't
> use next-gen to describe the 360 anymore, as we're already onto the
> Xbox 720, or whatever it will be called.
>
> When the 360 and it's rivals were built, Intel lost out to IBM, who
> secured console-based work with all the main three manufacturers. This
> time, Intel have got in early, according to Business Week. Speaking to
> the site, a company rep also rather enthusiastically predicted 'hands
> free' controllers that will see you using hand movements to play
> games, though we doubt the humble button will become redundant quite
> so readily.
>
> The good news is that if Intel do get the work, we will be able to
> enjoy a games machine running on chips that can perform an
> unimaginable 1 trillion calculations a second"
>
> http://xboxer.tv/2007/12/next_nextgen_already_under_dis.html
>
>
> "Intel has talked to console video game makers about using chips that
> can perform in excess of 1 trillion calculations per second
> (BusinessWeek.com, 2/12/07) in future products that use cameras to
> track body motion to control the action, instead of using buttons or
> joysticks. "We imagine some future generation of [Nintendo's] Wii
> won't have hand controllers," says Justin Rattner, Intel's chief
> technology officer. "You just set up the cameras around the room and
> wave your hand like you're playing tennis." Intel missed out on
> supplying chips for the current generation of game systems, and is
> trying to gain a foothold there."
>
>
> http://www.businessweek.com/technology/content/dec2007/tc20071212_550604.htm

I personally doubt that talking first has anything to do with who gets
the contracts. The vendors will balance cost, performance, and features,
and make a decision based on profit.

That's not a bad thing, but performance is generally not an issue now,
other than people writing brute force solutions because they're gamers,
not programmers, and believe that hardware should make crappy code look
good anyway.

--
Bill Davidsen
He was a full-time professional cat, not some moonlighting
ferret or weasel. He knew about these things.

Terje Mathisen
December 22nd 07, 05:19 PM
Bill Davidsen wrote:
>
> I personally doubt that talking first has anything to do with who gets
> the contracts. The vendors will balance cost, performance, and features,
> and make a decision based on profit.

This is almost certainly correct.
>
> That's not a bad thing, but performance is generally not an issue now,
> other than people writing brute force solutions because they're gamers,
> not programmers, and believe that hardware should make crappy code look
> good anyway.

This otoh is totally bogus:

Games programming is probably the only existing source of new
programmers who actually care about performance, care to an extent where
even 25% speedups are a big deal.

Yes, there are a lot of programmers even inside successful games
publishers who don't know/care about what makes a program fast, but
those same companies probably employ up to half of all the current
world-class performance programmers.

Terje
--
- >
"almost all programming can be viewed as an exercise in caching"

Nick Maclaren
December 22nd 07, 05:29 PM
In article >,
Terje Mathisen > writes:
|> Bill Davidsen wrote:
|> >
|> > That's not a bad thing, but performance is generally not an issue now,
|> > other than people writing brute force solutions because they're gamers,
|> > not programmers, and believe that hardware should make crappy code look
|> > good anyway.
|>
|> This otoh is totally bogus:

Oh, come now! It's not TOTALLY bogus - just largely so.

|> Games programming is probably the only existing source of new
|> programmers who actually care about performance, care to an extent where
|> even 25% speedups are a big deal.

Nope. HPC provides some, too. Not a lot, but a few.

And most of the people Bill Davidsen were talking about (i.e. who write
crappy code and believe that it is the hardware's business to make it
run fast) are in neither gaming nor HPC. In those areas, people learn
better - in other areas, they can get away with being idiotic, sloppy
and just plain crazy.

E.g. you can explain why, as time goes by, GUIs get slower at popping
up new windows and even responding to mouse clicks - but most people
(even most GUI programmers) can't.


Regards,
Nick Maclaren.

Terje Mathisen
December 22nd 07, 07:39 PM
Nick Maclaren wrote:
> In article >,
> Terje Mathisen > writes:
> |> Bill Davidsen wrote:
> |> >
> |> > That's not a bad thing, but performance is generally not an issue now,
> |> > other than people writing brute force solutions because they're gamers,
> |> > not programmers, and believe that hardware should make crappy code look
> |> > good anyway.
> |>
> |> This otoh is totally bogus:
>
> Oh, come now! It's not TOTALLY bogus - just largely so.
>
> |> Games programming is probably the only existing source of new
> |> programmers who actually care about performance, care to an extent where
> |> even 25% speedups are a big deal.
>
> Nope. HPC provides some, too. Not a lot, but a few.

Indeed.

There are probably at least an order of magnitude less HPC programmers
than (performance) games programmers, but still significant,
particularly due to having thought a lot about clusters vs SMP,
single-core vs dual/quad/many-core etc.

Terje
--
- >
"almost all programming can be viewed as an exercise in caching"