home *** CD-ROM | disk | FTP | other *** search
- Path: news.vol.it!news
- From: bizzetti@mbox.vol.it (Fabio Bizzetti)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: AB3D II beats Quake....
- Date: 31 Mar 1996 17:17:09 GMT
- Organization: Video On Line
- Distribution: world
- Message-ID: <6772.6664T1007T1458@mbox.vol.it>
- References: <nuijen.4o7e@grafix.xs4all.nl>
- NNTP-Posting-Host: molcl3.vol.it
- X-Newsreader: THOR 2.22 (Amiga;TCP/IP)
-
-
- >In a message of 23 Mar 96 Fabio Bizzetti wrote to All:
-
- >< A very very long message about the PowerPC chip versus 68060HP-PA >
-
- >Ehmm.. Fabio...Hi :)
-
- Hi ;)
-
- >I never heard of the HP-PA chip, what kind of chip is it ?
-
- It's one of the fastest RISC's, with MultiMedia core.
- It means it's even faster for MultiMedia stuffs.
-
- >What computer uses them ??
-
- Multimedia/Graphics Workstations.
-
- > I mean, the powerpc chip is used by some
- >Clones and offcourze by the Powermacs.. But the HP-PA chip ??
-
- Multimedia/Graphics Workstations v2.0 ;)
-
- >I never even heard of it !!
-
- You missed something very interesting ;)
-
- > Unless you're talking here about the
- >regulare 060's, and in that case. I think you're nuts !!!
-
- Nope, officially there will not be a 68070, so we must migrate.
- The 68030 (low-end) /68060 (high-end) was only to make this migration realistic,
- quick and easy, and to mantain a 2nd CPU for all that timing tasks that can be
- done only with parallel processing, wheter or not you've 200000 MIPS, like
- "hardware emulation of the OS" for games/demos. ;)
-
- >But then I'm no expert. But I guess the
- >PowerPC is more widely used and I think that it would therefor be
- >a lot cheaper and easier for AT to get these chips.
-
- The future of PPC is much less sure than the one of i.e. 80x86 :(
- Anyway, IMO PPC has been the best choice (due to the commericial benefits)
- *only* if we re-gain what we've lost not using the (non commercially optimal)
- but much more powerful HP-PA.
-
- > I mean that
- >was the problem with the 060 and the 040, there weren't many
- >that buyers, so not many were produced.
-
- Hmm.. I really think that 040 has been sold, and that people asks for 060 lotsa
- times more than Motorola ever expected, thus they had big problems to satisfy
- the request of 060's.
-
- Not exactly an insuccess for these CPU's, maybe for the poor Motorola. :(
-
- >Another thing.. I've read that end '96 or somewhere in '97 the PowerPC
- >chip can already reach speeds of 600Mhz.
-
- Before PPC does it, practically all the others will have reached it, including
- 80x86. It has nothing to do with PPC design, it has to do with technology.
-
- > I think the PowerPc chip is simply more supported.
-
- Agreed. But we need also more power, adding the the PPC what it can't do
- proporly (read: quickly).
- Also 80686 will have a "MultiMedia core", in-chip.
-
- As always, IBM will give us tomorrow the technology that nearly everyone had
- yesterday..
-
- > It doesn't matter if one chip is better. Look at
- >the Intel versus Motorola. I mean the 68000 was a better CPU then the
- >Clone chips these days.. But look now.. A pentium is about 4x times
- >faster then the 68060 (don't know if I'm correct here, just guessing)
-
- due to more advanced technology (read: Mhz)
-
- what about taking the best from both world? This is custom design.
-
- >, but then again the Pentiums are badly designed.
-
- bigtime, but they have lotsa Mhz..
-
- > Did you know that
- >if you don't give the CPU something to do, it's performance will go down ??
-
- ??? O:)
- =) well...yeah!
-
- I got what you mean though.. ehmm.... :)
-
- >Until you reset offcourse..
-
- phhheeew ;-)
-
- :)
-
- >The things about Custom-gfx-chips :
- >Why not use a SVGA ??
-
- because it's crap for anything else than a static WorkBench.
- I haven't seen cool games with 8088+SVGA, but I've seen them with 68000+OCS.
- so, would you prefer 80586+SVGA games, or the same CPU's MIPS+AAAA?
-
- > I mean they are cheap and updated very often !!!
-
- Watch the AgaEXTENDER: it's fully programmable and thus can *not* become
- obsolete, neither in 20 years. If they make AGA as *static design*, as soon
- as "bandwidth technology" :D evolves, you can "overclock" AGA of any number
- of times (having re-made them using the new technology, *but* with no need
- to redesign them because you used a *static design*). Quite perfect:
- compatible, cheap, upgradable, fully programmable ( so you dont need to make
- a new chipset when i.e. new video modes are required ).
-
- >If AT is designing it's own chips, it would mean AT needs about
- >30,40 very experienced gfx-chips designers and millions $$$
-
- Nope. Just 1 chip and some designers.
- They're re-designing AGA to make it consume less, and this is the good occasion
- to change its design into *static*, allowing cheap evolution following the
- philosophy above explained. It means no comprimizes: only huge advantages.
-
- > And we all know
- >that is not the situation right now. I mean a SVGA chip costs 10$ ? What
- >do think the Custom chips will cost if AT desgined them ??
-
- Extremely less in the performances/price ratio.
-
- > More then 10$,
- >thats for sure.
-
- Well, it's a paradox:
- The "10$ SVGA chips" are cheap, but they can only run a static WorkBench screen
- on low-end Amigas. So you need an high-end Amiga, where you can spend also 20$..
-
- ;D
-
- > And those chips would need to be upgraded every year to keep
- >up with the compitition.
-
- Nope:
- The design of AgaEXTENDER is fully programmable; meaning that *NO* updates will
- be necessary. As soon as technology evolves, you've only to make the same old
- chips using it, exploiting the new clock speeds, and you're done.
-
- If i.e. a day developers will need hybrid chunky+planar modes, both AGA and
- expecially SVGA will be obsolete in a moment, as AGA became because of chunky.
-
- AGA+AgaEXTENDER will still be there doing the job with 100% efficiency, being
- fully programmable.
-
- >I vagely remember you said something about a AGA-extension or something
- >alike.
-
- ..
-
- >I don't know what this project means, so won't say anything about it.
-
- You're cool pal :)))
- Wish there were many more like you. ;)
-
- >Only this..
-
- NO!!!! ;)
-
- > a extension is a usually a simply tempororarely (shit can't
- >remember how you spell it :( , did I spell it right ;) ??), that means where
- >still nowhere next year.
-
- The AgaEXTENDER aim is exactly the opposite, so dont worry.
-
- >Ps.. I'm not an expert at this areas, just saying what I'm thinking.
- >So please no flames...
-
- Nope, no flames mate.
-
- Ciao :)
-
-
- >-------=============<<<<<<<<<<<<<+>>>>>>>>>>>>>==============-------
- > _____ _ _ Projectleader project Mandor
- > / ____\ | \ | |
- > | | __ | \| | Usenet: nuijen@grafix.xs4all.nl
- > | |__\ \ | |\\ | FIDO : 2:286/407.46
- > \______/erbert|_| \_|uijen AMY : 39:153/201.46
- >-------=============<<<<<<<<<<<<<+>>>>>>>>>>>>>==============-------
- >-- Via Xenolink 1.985B5, XenolinkUUCP 1.1
-
-
-
-
- /-----------------------------------------------------------------------\
- | Fabio "Maverick" Bizzetti - bizzetti@mbox.vol.it - Maverick* at IRC |
- | The maker of "CyberMan" and "Virtual Karting" |
- | working on "VirtualRally" & "StarFighter" |
- \-----------------------------------------------------------------------/
-
-
-