I think the original Motorola roadmap listed the G5 using 0.10µ process. When it was uptdated, October/November? last year, the process was changed to 0.13µ.
Please note that I said,"I think", I offer no proof, maybe someone could?
Also, note the current roadmap lists a 0.15µ process for the G4, and everyone is predicting a 0.13µ process for the next generation of G4, Why, I don't know???
>Unfortunately the eCORE itself is not that highly customizable -- the system-on-a-chip is.
Sorry, yeah that's what I meant actually. I wanted to say that chips based off this eCORE can be quite different and that by looking at the 8540 you don't necessairly know what a Desktop "G5" might look like.
I am just a little concerned that while speak about 266 or 400 DDR on a .13 or .1µm Chip while Intel works on a 533 FSB and .09µm successor of the Xeon...
With the Xserve being out for no more than a few weeks, do you think one can expect important (that is, beyond Xserve's specifications) Powermac revisions the next month? Is there some interference in the specifications between the two product lines?
With the Xserve being out for no more than a few weeks, do you think one can expect important (that is, beyond Xserve's specifications) Powermac revisions the next month? Is there some interference in the specifications between the two product lines?</strong><hr></blockquote>
The Xserve already has a significantly different structure to the PowerMacs so I wouldn't expect them to grow along parallel lines.
<strong>I am just a little concerned that while speak about 266 or 400 DDR on a .13 or .1µm Chip while Intel works on a 533 FSB and .09µm successor of the Xeon...</strong><hr></blockquote>
Keep in mind, however, that Intel talks publicly about what its plans are for the next 5 years whereas with Apple-related product we're lucky to know 5 minutes in advance.
Intel is likely going to retain the process lead, but the feature size measurement is not the be all and end all of a process (i.e. yields, copper, wafer size, SOI, etc etc).
As for the FSB Intel is building, sure it sounds impressive to have it run at 533MHz... but that doesn't tell us how efficient its, how wide it is, what constraints it puts on system design, how expensive it is to implement, etc. I'm a believer in NUMA (i.e. on-chip memory controllers), so I don't care how fast Intel tries to make its FSB -- I think its a bad choice on their part. Time will tell, if I'm right (and AMD, Moto, etc).
I started a new thread with this, sorry for that. So now I post it here!
We got really excited over the new case, but if I remember correctly The Register (and the Inquirer) were running stories about a 2.4 ghz G5 (7500) and 1.5 ghz 7470 G4 being sampled by Motorola already and targeted at a summer (1.5) and Q1 next year (2.4) release.
I am just wondering to not have heared none of these lately, not even about Motorola starting to use HiP7 or anything like that (which surely would have been mentioned in a press release or two).
Could it be that Apple will not release more rthan a CPU upgrade in August opting to leave out the stopgap solution and going to the "G5" first thing in Q1 2003? Or is there any proof that The Register is full of BS on the PPC front? Is a mhz jump at more than 266 mhz even realistic?
I am reading The Register for around three years now and I can not really recall them going after pumped-up rumors to get hits. Most of the stuff they post is true (or turns out to be). Halfway. It just strikes me as odd that 7 months ago the rumor channels were full of "blazingly fast" test machines and "2.4 ghz G5s" whereas now people seem fine with a 1.2 ghz machines. Which I actually wouldnt mind either..
Check the Polish info on the G5+ above. Don't know if it's real, but I found it interesting. Due Quarter 4 2002 or Quarter 1 2003. Hope it's real and it's early.
Check the Polish info on the G5+ above. Don't know if it's real, but I found it interesting.</strong><hr></blockquote>
Duh, I totally missed that. That comes from skipping top post and reading only what Programmer writes.. <img src="graemlins/embarrassed.gif" border="0" alt="[Embarrassed]" />
Comments
Please note that I said,"I think", I offer no proof, maybe someone could?
Also, note the current roadmap lists a 0.15µ process for the G4, and everyone is predicting a 0.13µ process for the next generation of G4, Why, I don't know???
Sorry, yeah that's what I meant actually. I wanted to say that chips based off this eCORE can be quite different and that by looking at the 8540 you don't necessairly know what a Desktop "G5" might look like.
I am just a little concerned that while speak about 266 or 400 DDR on a .13 or .1µm Chip while Intel works on a 533 FSB and .09µm successor of the Xeon...
<strong>Motorola's G5 is coming!
Check this <a href="http://www.heise.de/newsticker/data/hes-22.07.02-000/"; target="_blank">Link</a>
(German)</strong><hr></blockquote>
With the Xserve being out for no more than a few weeks, do you think one can expect important (that is, beyond Xserve's specifications) Powermac revisions the next month? Is there some interference in the specifications between the two product lines?
<strong>
With the Xserve being out for no more than a few weeks, do you think one can expect important (that is, beyond Xserve's specifications) Powermac revisions the next month? Is there some interference in the specifications between the two product lines?</strong><hr></blockquote>
The Xserve already has a significantly different structure to the PowerMacs so I wouldn't expect them to grow along parallel lines.
<strong>I am just a little concerned that while speak about 266 or 400 DDR on a .13 or .1µm Chip while Intel works on a 533 FSB and .09µm successor of the Xeon...</strong><hr></blockquote>
Keep in mind, however, that Intel talks publicly about what its plans are for the next 5 years whereas with Apple-related product we're lucky to know 5 minutes in advance.
Intel is likely going to retain the process lead, but the feature size measurement is not the be all and end all of a process (i.e. yields, copper, wafer size, SOI, etc etc).
As for the FSB Intel is building, sure it sounds impressive to have it run at 533MHz... but that doesn't tell us how efficient its, how wide it is, what constraints it puts on system design, how expensive it is to implement, etc. I'm a believer in NUMA (i.e. on-chip memory controllers), so I don't care how fast Intel tries to make its FSB -- I think its a bad choice on their part. Time will tell, if I'm right (and AMD, Moto, etc).
We got really excited over the new case, but if I remember correctly The Register (and the Inquirer) were running stories about a 2.4 ghz G5 (7500) and 1.5 ghz 7470 G4 being sampled by Motorola already and targeted at a summer (1.5) and Q1 next year (2.4) release.
I am just wondering to not have heared none of these lately, not even about Motorola starting to use HiP7 or anything like that (which surely would have been mentioned in a press release or two).
Could it be that Apple will not release more rthan a CPU upgrade in August opting to leave out the stopgap solution and going to the "G5" first thing in Q1 2003? Or is there any proof that The Register is full of BS on the PPC front? Is a mhz jump at more than 266 mhz even realistic?
I am reading The Register for around three years now and I can not really recall them going after pumped-up rumors to get hits. Most of the stuff they post is true (or turns out to be). Halfway. It just strikes me as odd that 7 months ago the rumor channels were full of "blazingly fast" test machines and "2.4 ghz G5s" whereas now people seem fine with a 1.2 ghz machines. Which I actually wouldnt mind either..
Check the Polish info on the G5+ above. Don't know if it's real, but I found it interesting. Due Quarter 4 2002 or Quarter 1 2003. Hope it's real and it's early.
<strong>
Compare the e500 manual to the AltiVec PEM:
<a href="http://www.altivec.org/tech_specifications/altivec_pem.pdf"; target="_blank">http://www.altivec.org/tech_specifications/altivec_pem.pdf</a>;
The e500 SIMD instructions ARE NOT AltiVec.</strong><hr></blockquote>
Ouch! You got me!
I didn't read the tech paper, I saw SIMD unit, and assumed it was the Altivec...
Ass U Me
<strong>xype,
Check the Polish info on the G5+ above. Don't know if it's real, but I found it interesting.</strong><hr></blockquote>
Duh, I totally missed that. That comes from skipping top post and reading only what Programmer writes.. <img src="graemlins/embarrassed.gif" border="0" alt="[Embarrassed]" />
[ 07-24-2002: Message edited by: Matsu ]</p>