Yeh, here I've been really thinking the G5 will be 6 - 9 months away, at least.
Then I noticed that the Motorola roadmap shows the G5 to be 32 bit & 64 bit parts. Basically, a G4 with a new pipeline plus Rapid I/O and HiP7 manufacturting process. Don't get me wrong, still a mean feat, but not like a 64 bit processor.
Does any one know if Archintosh has released inside type information in the past? Is this unusual for them? Seems they are more of a news format.
<strong>I think we are all being led around by the nose and we our having our legs pulled.
If we keep believing in these rumors, this MWSF will be the absolute biggest disappointment yet!</strong><hr></blockquote>
Yeah. Bigger than the Paris disappointment last October. No wait-- they cancelled it.
Bigger than the MWNY last July.
I'm a believer in the Law of Averages, so I think we're due. Whether it's a G5 or not this January remains to be seen, but there are still several Rabbits that Steve has to pull out of his Reality-Distorted Top Hat, like a new iMac.
Even if a so-called "G5" is ready by January, Steve may wait until Tokyo or Seybold if the iMac is introduced in San Francisco. (I like how I'm using "Steve" as if he's my next door neighbor.)
Remember also that Naked Rats said that the upcoming "G5" is not the true successor to the throne, but rather a G4 made up to look like a G5. Kind of like the Man in the Iron Mask, but without Leonardo DiCaprio.
According to some research done by an AI poster ages ago (almost a year!), Raycer (who was bought by Apple) were basically working on a z-buffer accelerator on steroids. Hardware z-sorting.
In 3D, there are 3 (actually, 4, but lets not go there) axises (typo on purpose): x, ie: horizontal, y, ie: vertical, and z, ie : depth.
Z sorting goes like this: You are rendering a house in OpenGL.
Since you are facing the house dead on, it is a waste of processing power to render a) the sides of the house, which are not visible, b) the back of the house, same reason, and everything behind the house, for the same reason.
But let's say there are windows in the house. They are semi-translucent, and allow you to see what is behind them, perhaps a wall w/ a painting on it.
Graphics calculations like these are all z-buffer calcs. You take the wall and painting, 'multiply' them by the translucency of the window, maybe overlay a little fog for atmospherics, and render this to the screen in realtime.
A z-buffer calc. accelerator, ie a hardware chip on the G5 motherboard, plows through this, offloading this from the cpu, which can focus on other tasks.
So WTF does this have to do w/ accelerating Aqua?
Translucency in Aqua. All inactive windows, menus, etc, all overlapping, all the time. Throw in a spiffy Mac OS X xterm into the mix, to maximize the eye candy and increase the jealousy in your PC-lamer friends.
All those layers of translucency, overlapping each other, requires some serious reverse-order z-buffer calculating. Take the desktop pic, add a few overlapped inactive windows, and stick an xterm on top of that, and you have 5 or six levels of partially visible stuff. If there is a chip that a) figures out what isn't visible, and throws that data out of graphics memory, and b) figures out all the multiplied layers, and what the layers overlapped should look like as a 2d surface (ie your screen), you have a very fast, very snappy GUI.
Hope this helps.
Maya ain't coming to OSX because of 400 mhz G4s, boys. They're here for the G5's coming in 2002.
[quote]My current knowledge of OpenGL leads me to think a quartzlike engine could theoretically be made, but in a VERY Rube Goldberg like manner that would still see much of the code run in the CPU. The main issue here is how the pipeline is laid out and how the acceleration engines choose to implement portions of it for the speed gain.<hr></blockquote>
Excellent post, AirSluf. My knowledge of OpenGL is about 5 years old, so I defer.
Law of Averages? Maybe Law of Averages according to a chronic gambler!
Btw... that's a "clever" name...
Derek</strong><hr></blockquote>
Okay-- The Law of Really, Really Big Numbers Made by an Infinite Number of Red Necks Shooting an Infinite Number of Bullets at Street Signs to Come up with Shakespeare's Hamlet. In Braille.
BTW: Technically, it's Scheißkopf if this word existed in German. I'll have to check with the (German) girlfriend.
Check out <a href="http://www.thinksecret.com."; target="_blank">www.thinksecret.com.</a> What to make of this and please revist my "What Future Computer Will You Buy" thread.....will you purchase one if the posted specs come about? Will the name "G5" make a difference?
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around.
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around. </strong><hr></blockquote>
Because there is other evidence.
1) No Official Denials
2) Past statements his Steveness regarding the timely closing of the MHZ gap
3) MOT's own release timetable ...Q1.
4) Many of us believe Apollo is for PB G4 and possibly iMacs. Sahara G3's may be for ibook's.
5) MOT has conceded work is progressing well on the G5.
No one knows. Although ThinkSecret's record has been pretty good to my knowledge. This does hamper the case for the G5, but does not close it.
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around. </strong><hr></blockquote>
the fact that the G4 still has another scheduled revision due means SHIT. it really does. By using that logic we would be using a G3 750cx now just because there were more revisions left in the G3.
Apple will go to the G5 whenever its ready. If that is before or at the same time as the 7460 you can bet your money that Apple will go with the G5 over the 7460
Comments
<a href="http://www.architosh.com/news/2001-11/2001a-1130-appleg5.phtml"; target="_blank">http://www.architosh.com/news/2001-11/2001a-1130-appleg5.phtml</a>;
Definitely worth the read.
[ 11-30-2001: Message edited by: rickag ]</p>
[ 11-30-2001: Message edited by: Nitzer ]</p>
If a third of that is true, it'll be great.
If the clustering rumor comes true, I called it!
Then I noticed that the Motorola roadmap shows the G5 to be 32 bit & 64 bit parts. Basically, a G4 with a new pipeline plus Rapid I/O and HiP7 manufacturting process. Don't get me wrong, still a mean feat, but not like a 64 bit processor.
Does any one know if Archintosh has released inside type information in the past? Is this unusual for them? Seems they are more of a news format.
Considering that they deal with computationally intensive software in a PC-dominated field, I can understand their excitement, though.
---PPC Roadmap shows Q1 release
---No official denials
---published rumors that quite frankly, seem credible.
---job's supposed statement that the MHZ gap would be closed near the end of the year 2001
[ 12-01-2001: Message edited by: SDW2001 ]</p>
If we keep believing in these rumors, this MWSF will be the absolute biggest disappointment yet!
Forgive my ignorance of 2d vs 3d programming/chipset building . . .
Are there large differences between creating/designing a 2d graphics chip vs a 3d graphics chip?
It seems kinda of silly to buy a 3d company to accelerate 2d functions in Quartz. It sounds kinda like retrofitting.
I realize that 2d and 3d probably share a lot of functions or processes or whatnot - but how similar - or dissilimar are they?
Is this even a viable solution?
<strong>I think we are all being led around by the nose and we our having our legs pulled.
If we keep believing in these rumors, this MWSF will be the absolute biggest disappointment yet!</strong><hr></blockquote>
Yeah. Bigger than the Paris disappointment last October. No wait-- they cancelled it.
Bigger than the MWNY last July.
I'm a believer in the Law of Averages, so I think we're due. Whether it's a G5 or not this January remains to be seen, but there are still several Rabbits that Steve has to pull out of his Reality-Distorted Top Hat, like a new iMac.
Even if a so-called "G5" is ready by January, Steve may wait until Tokyo or Seybold if the iMac is introduced in San Francisco. (I like how I'm using "Steve" as if he's my next door neighbor.)
Remember also that Naked Rats said that the upcoming "G5" is not the true successor to the throne, but rather a G4 made up to look like a G5. Kind of like the Man in the Iron Mask, but without Leonardo DiCaprio.
~e
<strong>
I'm a believer in the Law of Averages, so I think we're due.
</strong><hr></blockquote>
Law of Averages? Maybe Law of Averages according to a chronic gambler!
Btw... that's a "clever" name...
Derek
In 3D, there are 3 (actually, 4, but lets not go there) axises (typo on purpose): x, ie: horizontal, y, ie: vertical, and z, ie : depth.
Z sorting goes like this: You are rendering a house in OpenGL.
Since you are facing the house dead on, it is a waste of processing power to render a) the sides of the house, which are not visible, b) the back of the house, same reason, and everything behind the house, for the same reason.
But let's say there are windows in the house. They are semi-translucent, and allow you to see what is behind them, perhaps a wall w/ a painting on it.
Graphics calculations like these are all z-buffer calcs. You take the wall and painting, 'multiply' them by the translucency of the window, maybe overlay a little fog for atmospherics, and render this to the screen in realtime.
A z-buffer calc. accelerator, ie a hardware chip on the G5 motherboard, plows through this, offloading this from the cpu, which can focus on other tasks.
So WTF does this have to do w/ accelerating Aqua?
Translucency in Aqua. All inactive windows, menus, etc, all overlapping, all the time. Throw in a spiffy Mac OS X xterm into the mix, to maximize the eye candy and increase the jealousy in your PC-lamer friends.
All those layers of translucency, overlapping each other, requires some serious reverse-order z-buffer calculating. Take the desktop pic, add a few overlapped inactive windows, and stick an xterm on top of that, and you have 5 or six levels of partially visible stuff. If there is a chip that a) figures out what isn't visible, and throws that data out of graphics memory, and b) figures out all the multiplied layers, and what the layers overlapped should look like as a 2d surface (ie your screen), you have a very fast, very snappy GUI.
Hope this helps.
Maya ain't coming to OSX because of 400 mhz G4s, boys. They're here for the G5's coming in 2002.
[ 12-02-2001: Message edited by: stimuli ]</p>
[quote]My current knowledge of OpenGL leads me to think a quartzlike engine could theoretically be made, but in a VERY Rube Goldberg like manner that would still see much of the code run in the CPU. The main issue here is how the pipeline is laid out and how the acceleration engines choose to implement portions of it for the speed gain.<hr></blockquote>
Excellent post, AirSluf.
<strong>
Law of Averages? Maybe Law of Averages according to a chronic gambler!
Btw... that's a "clever" name...
Derek</strong><hr></blockquote>
Okay-- The Law of Really, Really Big Numbers Made by an Infinite Number of Red Necks Shooting an Infinite Number of Bullets at Street Signs to Come up with Shakespeare's Hamlet. In Braille.
BTW: Technically, it's Scheißkopf if this word existed in German. I'll have to check with the (German) girlfriend.
~e
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around.
<strong>Regarding ThinkSecret:
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around.
Because there is other evidence.
1) No Official Denials
2) Past statements his Steveness regarding the timely closing of the MHZ gap
3) MOT's own release timetable ...Q1.
4) Many of us believe Apollo is for PB G4 and possibly iMacs. Sahara G3's may be for ibook's.
5) MOT has conceded work is progressing well on the G5.
No one knows. Although ThinkSecret's record has been pretty good to my knowledge. This does hamper the case for the G5, but does not close it.
[ 12-02-2001: Message edited by: SDW2001 ]</p>
<strong>
Because there is other evidence.
2) Past statements his Steveness regarding the timely closing of the MHZ gap
5) MOT has conceded work is progressing well on the G5. ]</strong><hr></blockquote>
If Apollo appears at 1.4Ghz at MWSF, that would constitute 'closing the gap', because we would go from being 1Ghz behind to 500+Mhz behind the P4.
Motorola did say the G5 is progressing, but they also said, don't get your hopes up so soon (paraphrase).
MOSR has 0% credibility with me. TheRegister I will have to wait and see how this pans out.
Peace.
<strong>Regarding ThinkSecret:
I've been saying it all along in my signature. Why do we insist on believing in pie-in-the-sky G5 processors, when it was made clear a year ago that Apollo was the next G4 upgrade in the time-line.
I believe the 'G5' source om MOSR and TheRegister is a Wintel jerk jerking us around.
the fact that the G4 still has another scheduled revision due means SHIT. it really does. By using that logic we would be using a G3 750cx now just because there were more revisions left in the G3.
Apple will go to the G5 whenever its ready. If that is before or at the same time as the 7460 you can bet your money that Apple will go with the G5 over the 7460