the "G5" isn't going to be multiple cores on one die,
but it will be derived from IBMs multicore POWER4
The chip in question is of course the PowerPC 970 (that's PDF of the microprocessor forum presentation on the 970)
In short, take a power4, lop off core #2, reduce the amount of L2 cache, add an altivec execution unit, change the bus interface and make it on a smaller (.13 rather than.18) process, and eh voila, PowerPC 970
hrmm, is it possible they've utilised the altivec hardware for page rendering?, would explain why I've seen posts on usenet from people claiming it's slower than IE on their B&W G3s
also, WRT the smooth scrolling, it occured to me that they might render the entire page and hold it in a backbuffer, just moving up and down something that's been prerendered should easily be faster than rendering it 'in realtime!'
Get back to me when Lian Li make a case as nice as the Coolermaster ATC-710
The guts of that one are the same as the Antec/Chieftec towers BTW... complete with (metal) clip-in drivebays for 3.5" devices and rails for the 5.25" bays.
I can SEE 20 frames per second.. individually, that's below the point my eye needs to create the illusion of motion. (incidentally, I frequently notice shuddering on film)
That's not "speed" in computer performance terms, that's frequency, in clock signal terms.
If you made a clock that ticked every quarter second, and the tick corresponded with one quarter of a second of motion, it would be ticking with four times the frequency of a normal clock, yet the SPEED of the clock would be identical.
The Radeon 8500 is ~10% behind the Ti4200 in benchmarks, which is >~10% faster than a Ti500
Hence Ti4200 > Radeon 8500 > Ti500
No "proof" needed, this is all known and documented data.
The initial 8500 drivers were dodgy though and they stopped it performing to it's best potential, so if you're thinking back to reviews when it was released you'd remember Ti500's outpacing it somewhat.
um.. Sega had stopped making Dreamcasts when the PS2 was released?
Additionally, I'll take a game at 1600x1200/32bit with 4x FSAA and ansiotropic filtering cranked up at 85FPS over 1600x1200/32bit with no FSAA ar ansio.
ATM I can do the latter with my Ti4200, not the former, if I had an R9700Pro, I could do the former.
No, the number of execution pipes, the kind of instructions they have to deal with and how well the rest of the system can keep the processor fed with data all have an impact on speed, clock frequency alone is a terrible indicator of performance because.. because well, it doesn't indicate performance.
For a nice extreme example of this, compare an IBM POWER4+ @ 1.45Ghz (austensibly a PowerPC chip) with the Pentium 4 Northwood @ 3.06Ghz
Notice that the POWER4+ beats the unholy crap out of the Pentium 4 even though it's clock frequency is below half?
Slashdot Mirror
the "G5" isn't going to be multiple cores on one die,
.18) process, and eh voila, PowerPC 970
but it will be derived from IBMs multicore POWER4
The chip in question is of course the PowerPC 970 (that's PDF of the microprocessor forum presentation on the 970)
In short, take a power4, lop off core #2, reduce the amount of L2 cache, add an altivec execution unit, change the bus interface and make it on a smaller (.13 rather than
To be fair, the noise from the MDD Powermacs is a tad nasty if you want to use it in a production (audio/video) environment...
:)
that said, they ain't got NOTHIN' on a 7000rpm Delta
USB was Intels doing wasn't it?
:)
regardless, it was apple that popularized it..
notice how all the circa 1998 USB devices are.. well, imac coloured
As an Englishman myself, I have this to say.
:p
It's our language and we can butcher it any way we see fit.
So there!
IBM have a grasp on what should be in a laptop processor, it's called a 750FX, buy an iBook if you want one :)
hrmm, is it possible they've utilised the altivec hardware for page rendering?, would explain why I've seen posts on usenet from people claiming it's slower than IE on their B&W G3s
:)
also, WRT the smooth scrolling, it occured to me that they might render the entire page and hold it in a backbuffer, just moving up and down something that's been prerendered should easily be faster than rendering it 'in realtime!'
No facts, just idle musings
Thing is, the 12" Powerbook is quite clearly using a variation of the 12" iBook case design....
It's because Microsofts implementations of an idea tend to be halfassed, whilst Apple, usually, do a much slicker job of it.
Get back to me when Lian Li make a case as nice as the Coolermaster ATC-710
The guts of that one are the same as the Antec/Chieftec towers BTW... complete with (metal) clip-in drivebays for 3.5" devices and rails for the 5.25" bays.
If you must troll, at least do it in an educated way.. the firewire disk enclosures would be limited the same way if you connected them to a PC...
Wait till AFTER Ceta Alpha VI explodes, then the property prices on V will drop drastically.
Sure, on optimised Altivec code that'll fit in the L2 cache dataset and all, the G4 will do some serious performing.
:)
Try that same algorithm on a larger dataset though, and it'll smack into memory bandwidth limitations, hard.
OTOH, the POWER4 performs roughly on a par with the G4/Altivec RC5 cruncher using standard boring integer code, this bodes well for the PowerPC 970
It's our language and we can do whatever we like with it!
emphasis on "Chugs along" I feel....
...
I can SEE 20 frames per second.. individually, that's below the point my eye needs to create the illusion of motion. (incidentally, I frequently notice shuddering on film)
I stop noticing fps improvements at ~70fps
That's not "speed" in computer performance terms, that's frequency, in clock signal terms.
If you made a clock that ticked every quarter second, and the tick corresponded with one quarter of a second of motion, it would be ticking with four times the frequency of a normal clock, yet the SPEED of the clock would be identical.
You realise that a desktop P4 3.06Ghz can kick out around 100w of heat when going 100% on something... right?..
:)
Do you have ANY idea what that would do to your lap?
The Radeon 8500 is ~10% behind the Ti4200 in benchmarks, which is >~10% faster than a Ti500
Hence Ti4200 > Radeon 8500 > Ti500
No "proof" needed, this is all known and documented data.
The initial 8500 drivers were dodgy though and they stopped it performing to it's best potential, so if you're thinking back to reviews when it was released you'd remember Ti500's outpacing it somewhat.
You do know that the GFFX can't do anything much that the Radeon 9700 can't?.. right?...
um.. Sega had stopped making Dreamcasts when the PS2 was released?
Additionally, I'll take a game at 1600x1200/32bit with 4x FSAA and ansiotropic filtering cranked up at 85FPS over 1600x1200/32bit with no FSAA ar ansio.
ATM I can do the latter with my Ti4200, not the former, if I had an R9700Pro, I could do the former.
"1) ATI has wildly unstable drivers."
No moreso than Nvidia these days, unless you don't count nv4_disp.dll BSODS as "unstable" for some reason...
" 1 MB cache (512 KB)"
And the rest. the 1MB cache block in the Powerbook is L3, the 7455 has 256KB of L2 as well.
No, the number of execution pipes, the kind of instructions they have to deal with and how well the rest of the system can keep the processor fed with data all have an impact on speed, clock frequency alone is a terrible indicator of performance because.. because well, it doesn't indicate performance.
For a nice extreme example of this, compare an IBM POWER4+ @ 1.45Ghz (austensibly a PowerPC chip) with the Pentium 4 Northwood @ 3.06Ghz
Notice that the POWER4+ beats the unholy crap out of the Pentium 4 even though it's clock frequency is below half?
That's hardly 5GB for $15 ;)
Musicmatch jukebox is so bad I wouldn't use it if I were being paid..
I'll stick to Ephpod and my Win32 ipod thankyouverymuch