These figures you have quoted for the 970s other than the 1.8 are highly optimistic. You will not get linear scaling unless, as I said, this processor is extra-ordinary (ie, not like all others). SPECfp2000 doesn't get benefit from MP systems. Look at this page and tell me which systems get 2x perf for 2x CPUs:
The multi-CPU P4's & Xeons have heat issues & cost through the nose. (Yes, there are things more expensive than Macs.)
</strong><hr></blockquote>
While P4's and Xeons are hotter than G4s (and 970s) they still ship and work well. And for similarly configured systems x86 systems are the same or lower price. (Even name brand systems.)
[quote]<strong>
That's not the main thing though.
</strong><hr></blockquote>
You missed my point - the comparison of the x86 and 970 was the main thing. Please re-read my post.
[quote]<strong>
All you need to do is read the line "2xG4" and the line "2x 970"
5.57x as fast. Trample-all-overness.
</strong><hr></blockquote>
Sure I agree the 970 will trample the G4 - that's not what I was talking about. If you wan't to argue you are going to have to change the subject.
[quote]<strong>
And the thing that's been the big hurdle historically at Apple to Quads is the _bus_. With the bus fixed, a Quad-ppc can easily undercut a Quad-Xeon in price... Which means the '4x 970' line might be interesting.
Having the bus fixed should also be a HUGE benefit to AltiVec, which isn't reflected (at all) in the Spec numbers. No Mac-types were too concerned when 'pc Hz' was around 1.3-1.5x Mac Hz - AV quite effective for what it does.</strong><hr></blockquote>
So MP P4s and Xeons are hot and expensive but Quad-PPC 970s are going to be cool and cheap?
I agree that the new bus will make quads easier and VMX/AltiVec scream. Again, if you wan't to argue...
I was making the point that people should have a realistic expectation of the performance of the 2.5 GHz 970. It will be competetive (with x86) but has yet to prove its trample-all-over-ness (vs x86). Personally I *hope* it does but I think it is a wait and see issue, particularly considering Intel are doing some serious improvements to the Prescott core.
The Pentium took about 90 days to manufacture, around the clock, 24 hours a day production time, from start to finish . In previous posts here and elsewhere, the time to manufacture cpu's has been mentioned to be about 60 days, from start to finish.
Does any one know if this has been significantly reduced?</strong><hr></blockquote>
When we order micros (from Motorola, for example), it takes them 8-12 weeks to get them to our production lines. (These are wimpy 16-bit micros, by the way.) What takes time is for them to put our program into ROM.
Beats me what they do with Pentiums. It's already designed-- they just have to pack sand, er, so to speak.
Well, I meant two separate fairly surprising items... at least based on current conventional wisdom re: the PPC970's and Apple's new machines.</strong><hr></blockquote>
...or maybe you're trying to tell me that you'd rather not answer, and I'm just not getting the hint. <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
If the 970 was indeed partially developed for Apple (the proof is when it shows up on apple.com), then what was IBM's incentive to make this processor? I doubt AIX and POWER-line Linux worksation/server sales are hurting right now, especially with the 1.45Ghz POWER 4+ recently made available.
IBM is a business, and recently quite a successful one. Apple, on the other hand, is seen by many investors as being in a precarious position in the market - dependent on an influx of switchers to survive - either from Windows or from their own OS 9. To make a significant new processor designed to specs which put it in competition not with Sun Blade 1000s but with Dell P4 machines indicates that either (1) IBM thinks they can sell a lot of Linux and AIX on a new architecture with no professional apps available, or (2) it was designed for Apple, and perhaps IBM is intending on not just selling Linux on a new architecture, but an existing architecture with existing users and apps - OS X.
Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
So we all go out and get one, right? But then what'll happen to these boards? No more G5 talk, everyone is happy, duals that satisfy even the most demanding 3d'er and gamer. We might actually have to do something with these machines!
I'm curious as to how Apple will use the power of the 970s and 980s in the future. They brought vidoe editing to the masses to take advantage of the G4. What next?
<strong>Nevyn, you almost completely ignored my post.</strong><hr></blockquote>
Sorry
[quote]Originally posted by MartianMatt:
<strong>These figures you have quoted for the 970s other than the 1.8 are highly optimistic. </strong><hr></blockquote>
Granted. But all of them got the same optimistic treatment -> not totally off into left field. And the 970's bus might be increasing in there - not sure. Multi-CPUs do nothing for Spec - but they _DO_ have dramatic effects on real code -> "effective SpecFP". They'll have to come up with something else soon anyway - hyper-threading seems like MP to the program.
[quote]Originally posted by MartianMatt:
<strong>So MP P4s and Xeons are hot and expensive but Quad-PPC 970s are going to be cool and cheap?</strong><hr></blockquote>
Cool_er_ and Cheap_er_.
Some Xeon MP _chips_ cost more than some PowerMacs. They have a larger die-size, and appearently are more complex to arrange into an MP config. Shrug. The heat dissipation stats are scary.
[quote]Originally posted by MartianMatt:
<strong>It will be competetive (with x86) but has yet to prove its trample-all-over-ness (vs x86).</strong><hr></blockquote>
Ok.
What I was trying to say is this:
The G4 when introduced had a (very brief) reign at the top. It hasn't been until this last year/year and a half that Mac people have really started hurting speedwise. (Ignoring _price_) Part of the reason it hasn't hurt so much as P4's Hz climbed like a rocket is Altivec - which is somewhat crippled in current G4/G4+ chips simply by the bus.
The 970 should _demolish_ the G4+. The bus improvements, extra FPU... That's what the Spec stats _show_. You are right, those stats do not show the 970 crushing the P4....
But anything that demolishes the G4+ _that_ bad should be better off that 'competitve with x86'. Spec didn't show the G4 doing as well against the Pentiums as the photoshop tests/RC5/other altivecable tests showed either.
At the very least, I'd say that a custom written demo utilizing the 970's features should be disturbing. The P4's version wouldn't have access to 64-bit calcs, which we'd be sure to use lots of, the P4's version couldn't keep all the computation units filled (since they can't all be used simultaneously), and AV with a lot more bus to play with....
From what I gather, Bochs has been basically a one-man operation. How much of its sluggishness is just due to inefficient coding (no offense intended to Kevin Lawton, if you're out there)? I would imagine Apple engineers could massage his code to a fare-thee-well...
As I understand it, Bochs is written in C++. There's no way that's going to run at decent speed without a clean rewrite.
Bochs is so slow it's ridiculous. I can try to launch an old DOS game, and it doesn't even launch before I eventually get fed up and kill the damn thing. Sure, the 970 will be a lot faster than my G4/450, but not that much faster...
<strong>Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.</strong><hr></blockquote>
I think you are missing the real incentive for IBM. The POWER line of CPUs are the brains of mid to high-end servers. Time is showing a huge market for low-end (blade) style servers - this lite version of the POWER CPUs are for that end of the spectrum.
IBM has a huge incentive to stay competitive in their back yard. The 970 will allow them to span the server spectrum from high to low. Apple might have put a chunk of cash into the development of this chip (?) - but IBM sure as hell is reaping some benefits from it as well.
I think you are missing the real incentive for IBM. The POWER line of CPUs are the brains of mid to high-end servers. Time is showing a huge market for low-end (blade) style servers - this lite version of the POWER CPUs are for that end of the spectrum.
IBM has a huge incentive to stay competitive in their back yard. The 970 will allow them to span the server spectrum from high to low. Apple might have put a chunk of cash into the development of this chip (?) - but IBM sure as hell is reaping some benefits from it as well.</strong><hr></blockquote>
One, I didn't say that they won't benefit; I just opined that those benefits wouldn't be enough to make IBM go for this expensive proposition.
Two, IBM is committing all it's resources to Linux, and for running a blade server, Linux on Intel works fine. In fact, the Linux/970 solution has a major disadvantage of having just about no available software and only running two distributions (SuSE 8.0 and RedHat 7.1, which by RedHat's own mouth is obsolete). RH may have a release of 8.0/8.1 in the wings, but for now Linux/POWER is immature.
As I understand it, Bochs is written in C++. There's no way that's going to run at decent speed without a clean rewrite.
Bochs is so slow it's ridiculous. I can try to launch an old DOS game, and it doesn't even launch before I eventually get fed up and kill the damn thing. Sure, the 970 will be a lot faster than my G4/450, but not that much faster...</strong><hr></blockquote>
Yup, I see that now after nosing around the Bochs site for a while. Obviously, Bochs is not the answer, and the columnists recommending it didn't know what they were talking about ( :eek: I'm shocked! How could that be???? )
Ah, well. Maybe Apple is writing their own, or perhaps 'tis not to be...
If the 970 was indeed partially developed for Apple (the proof is when it shows up on apple.com), then what was IBM's incentive to make this processor? I doubt AIX and POWER-line Linux worksation/server sales are hurting right now, especially with the 1.45Ghz POWER 4+ recently made available.
IBM is a business, and recently quite a successful one. Apple, on the other hand, is seen by many investors as being in a precarious position in the market - dependent on an influx of switchers to survive - either from Windows or from their own OS 9. To make a significant new processor designed to specs which put it in competition not with Sun Blade 1000s but with Dell P4 machines indicates that either (1) IBM thinks they can sell a lot of Linux and AIX on a new architecture with no professional apps available, or (2) it was designed for Apple, and perhaps IBM is intending on not just selling Linux on a new architecture, but an existing architecture with existing users and apps - OS X.
Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
IBM has had their own adgenda with the PowerPC since the beginning of the alliance. I'm not positive, but I would bet one of the main things that is driving their long term strategy is to free themselves from as much outside suppliers as possable, and the PowerPC allows them to do this, It also might bring in more vendors building PowerPC computers than currently exist, which means even more revenue for IBM. IBM looses potential profit (or a market cost advantage over the competition) with every Pentium based PC they sell, they are giving that profit to Intel for their processors and Microsoft for licensing fees.
No more Intel, no more Microsoft. They are again in charge of their destiny. If they have a client that can help bankroll the processor development by buying sufficient quantities (Apple), it is but icing on the cake.
i love reading posts by people who are like 800 times smarter than me....
g</strong><hr></blockquote>
I agree, where is programmer? He has some great insights. And Moki has some great posts but I wish he would spill the beans and say what he expects.
I was wondering how many changes/updates are required to the OS kernel to support the 970 and its companion chip. I do not design computers so I was wondering how long it would take to develop the companion chip and the OS X updates that will make the whole thing run. I have to believe it is more work than going from the 7455 to the 7457, but how much more I haven't a clue.
<strong>everyone sure has gotten their expectations up. There has been no announcement that apple will even use the 970 or anything else for that matter. THey could do something completely arbitruary. Lets get back to earth. Mac users this is Houston... do you read me?</strong><hr></blockquote>
Yah, you could be correct . . . , but Apple would be throwing away alot of developement $$ (ie: since Feb 2000, do the math). Let's just say that an unrefutable source says that if the 970 isn't for Apple, then why were they involved in the chip's developement since it's inception (read 02/2000).
And "Rickag", I tend to agree with "Rhumgod", because the 970 production process testing was half way completed in early November 2002 on 200mm and 300mm wafers.
I was wondering how many changes/updates are required to the OS kernel to support the 970 and its companion chip.</strong><hr></blockquote>
That depends if Apple wants to support the 970 in both 32 and 64 bit modes. If they are only interested in the 32bit mode of operations/address space. Probably not that much.
There is a lot more work to do on the hardware side.
- The 970 is quad pumped giving an effective bus of 1.8 Ghz! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> Yes, a bus faster than any G4...
- IBM licenses OSX for some markets. This breaks the long time business objection of Apple being single source.
- The 970 is used across the board. Prices are set more by the box and features than by processor.
- The towers are dual with a quad on top or in a rack.
- The 12" powerbook replaces the iBook. A version with a polycarbonate case is called the eBook.
- Apple engineers a graphics card that has a GPU from ATI or nVidia but also has a G4ish chip: one integer, no fpu, one altivec and a pile of L2. Quartz moves entirely to the graphics board. Now you know what Racer was doing. Code name: Ludicrous Drive.
- The UI becomes totally vector based and thus scalable for different pixel densities.
- Apple and IBM intro high end monitors that are 200ppi.
- Apple intros a headless, affordable, iBriq for the hacker/server/renderfarm crowds.
- Apple engineers a memory controller with a large "L3ish" built in. Memory controller controls it completely. In this way it can optimise it for many processors at once while minimising bus snooping. The "L3" is large enough to preclude the need for external RAM in some of the more modest machines...like the iBriq.
- Bandwidth is so high every machine can channel Steve's RDF in real time. Steve turns into a big glowing ball on stage and ascends to become the first Star Child. Stanley Kubrick rolls in his grave.
<strong>Come on, they just released new ones. That delay is only on the dual 1.4, which is a new chip. The other models say "same day."</strong><hr></blockquote>
Just checked again, and now the dual is 7-10 days and you're right the others are at one day. (I swear they said 3-4 weeks last time I checked...)
- The 970 is quad pumped giving an effective bus of 1.8 Ghz! Yes, a bus faster than any G4...
- IBM licenses OSX for some markets. This breaks the long time business objection of Apple being single source.
- The 970 is used across the board. Prices are set more by the box and features than by processor.
- The towers are dual with a quad on top or in a rack.
- The 12" powerbook replaces the iBook. A version with a polycarbonate case is called the eBook.
- Apple engineers a graphics card that has a GPU from ATI or nVidia but also has a G4ish chip: one integer, no fpu, one altivec and a pile of L2. Quartz moves entirely to the graphics board. Now you know what Racer was doing. Code name: Ludicrous Drive.
- The UI becomes totally vector based and thus scalable for different pixel densities.
- Apple and IBM intro high end monitors that are 200ppi.
- Apple intros a headless, affordable, iBriq for the hacker/server/renderfarm crowds.
- Apple engineers a memory controller with a large "L3ish" built in. Memory controller controls it completely. In this way it can optimise it for many processors at once while minimising bus snooping. The "L3" is large enough to preclude the need for external RAM in some of the more modest machines...like the iBriq.
- Bandwidth is so high every machine can channel Steve's RDF in real time. Steve turns into a big glowing ball on stage and ascends to become the first Star Child. Stanley Kubrick rolls in his grave.
Comments
[quote]<strong>
[code]
1x G4+ 1.0GHz 187
1x G4+ 1.4GHz 261
2x G4+ 1.4GHz 523
1x P4 3.4GHz 1249
1x 970 1.8GHz 1051
1x 970 2.5GHz 1459
2x 970 2.5GHz 2918
4x 970 2.5GHz 5839
</pre><hr></blockquote>
</strong><hr></blockquote>
These figures you have quoted for the 970s other than the 1.8 are highly optimistic. You will not get linear scaling unless, as I said, this processor is extra-ordinary (ie, not like all others). SPECfp2000 doesn't get benefit from MP systems. Look at this page and tell me which systems get 2x perf for 2x CPUs:
<a href="http://www.spec.org/cpu2000/results/rfp2000.html"; target="_blank">http://www.spec.org/cpu2000/results/rfp2000.html</a>;
[QUOTE]<strong>
The multi-CPU P4's & Xeons have heat issues & cost through the nose. (Yes, there are things more expensive than Macs.)
</strong><hr></blockquote>
While P4's and Xeons are hotter than G4s (and 970s) they still ship and work well. And for similarly configured systems x86 systems are the same or lower price. (Even name brand systems.)
[quote]<strong>
That's not the main thing though.
</strong><hr></blockquote>
You missed my point - the comparison of the x86 and 970 was the main thing. Please re-read my post.
[quote]<strong>
All you need to do is read the line "2xG4" and the line "2x 970"
5.57x as fast. Trample-all-overness.
</strong><hr></blockquote>
Sure I agree the 970 will trample the G4 - that's not what I was talking about. If you wan't to argue you are going to have to change the subject.
[quote]<strong>
And the thing that's been the big hurdle historically at Apple to Quads is the _bus_. With the bus fixed, a Quad-ppc can easily undercut a Quad-Xeon in price... Which means the '4x 970' line might be interesting.
Having the bus fixed should also be a HUGE benefit to AltiVec, which isn't reflected (at all) in the Spec numbers. No Mac-types were too concerned when 'pc Hz' was around 1.3-1.5x Mac Hz - AV quite effective for what it does.</strong><hr></blockquote>
So MP P4s and Xeons are hot and expensive but Quad-PPC 970s are going to be cool and cheap?
I agree that the new bus will make quads easier and VMX/AltiVec scream. Again, if you wan't to argue...
I was making the point that people should have a realistic expectation of the performance of the 2.5 GHz 970. It will be competetive (with x86) but has yet to prove its trample-all-over-ness (vs x86). Personally I *hope* it does but I think it is a wait and see issue, particularly considering Intel are doing some serious improvements to the Prescott core.
MM
[ 02-28-2003: Message edited by: MartianMatt ]</p>
<strong>moki, do you agree to the above being conventional wisdom about apple and the 970, or am i missunderstanding your post?
I did not have sexual relations with that woman...
err... what did you say?
<strong>
Ok, I'm asking for help.
The Pentium took about 90 days to manufacture, around the clock, 24 hours a day production time, from start to finish . In previous posts here and elsewhere, the time to manufacture cpu's has been mentioned to be about 60 days, from start to finish.
Does any one know if this has been significantly reduced?</strong><hr></blockquote>
When we order micros (from Motorola, for example), it takes them 8-12 weeks to get them to our production lines. (These are wimpy 16-bit micros, by the way.) What takes time is for them to put our program into ROM.
Beats me what they do with Pentiums. It's already designed-- they just have to pack sand, er, so to speak.
<strong>
Well, I meant two separate fairly surprising items... at least based on current conventional wisdom re: the PPC970's and Apple's new machines.</strong><hr></blockquote>
...or maybe you're trying to tell me that you'd rather not answer, and I'm just not getting the hint. <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
<strong>
there may be more than one bag of sand, too... two more that I can think of actually... I posted it here as well:
<a href="http://www.ambrosiasw.com/webboard/Forum64/HTML/001106.html"; target="_blank">http://www.ambrosiasw.com/webboard/Forum64/HTML/001106.html</a>;
[ 02-27-2003: Message edited by: moki ]</strong><hr></blockquote>
Another thought on new suprises:
If the 970 was indeed partially developed for Apple (the proof is when it shows up on apple.com), then what was IBM's incentive to make this processor? I doubt AIX and POWER-line Linux worksation/server sales are hurting right now, especially with the 1.45Ghz POWER 4+ recently made available.
IBM is a business, and recently quite a successful one. Apple, on the other hand, is seen by many investors as being in a precarious position in the market - dependent on an influx of switchers to survive - either from Windows or from their own OS 9. To make a significant new processor designed to specs which put it in competition not with Sun Blade 1000s but with Dell P4 machines indicates that either (1) IBM thinks they can sell a lot of Linux and AIX on a new architecture with no professional apps available, or (2) it was designed for Apple, and perhaps IBM is intending on not just selling Linux on a new architecture, but an existing architecture with existing users and apps - OS X.
Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
[ 02-28-2003: Message edited by: Anonymous Karma ]</p>
I'm curious as to how Apple will use the power of the 970s and 980s in the future. They brought vidoe editing to the masses to take advantage of the G4. What next?
<strong>Nevyn, you almost completely ignored my post.</strong><hr></blockquote>
Sorry
[quote]Originally posted by MartianMatt:
<strong>These figures you have quoted for the 970s other than the 1.8 are highly optimistic. </strong><hr></blockquote>
Granted. But all of them got the same optimistic treatment -> not totally off into left field. And the 970's bus might be increasing in there - not sure. Multi-CPUs do nothing for Spec - but they _DO_ have dramatic effects on real code -> "effective SpecFP". They'll have to come up with something else soon anyway - hyper-threading seems like MP to the program.
[quote]Originally posted by MartianMatt:
<strong>So MP P4s and Xeons are hot and expensive but Quad-PPC 970s are going to be cool and cheap?</strong><hr></blockquote>
Cool_er_ and Cheap_er_.
Some Xeon MP _chips_ cost more than some PowerMacs. They have a larger die-size, and appearently are more complex to arrange into an MP config. Shrug. The heat dissipation stats are scary.
[quote]Originally posted by MartianMatt:
<strong>It will be competetive (with x86) but has yet to prove its trample-all-over-ness (vs x86).</strong><hr></blockquote>
Ok.
What I was trying to say is this:
The G4 when introduced had a (very brief) reign at the top. It hasn't been until this last year/year and a half that Mac people have really started hurting speedwise. (Ignoring _price_) Part of the reason it hasn't hurt so much as P4's Hz climbed like a rocket is Altivec - which is somewhat crippled in current G4/G4+ chips simply by the bus.
The 970 should _demolish_ the G4+. The bus improvements, extra FPU... That's what the Spec stats _show_. You are right, those stats do not show the 970 crushing the P4....
But anything that demolishes the G4+ _that_ bad should be better off that 'competitve with x86'. Spec didn't show the G4 doing as well against the Pentiums as the photoshop tests/RC5/other altivecable tests showed either.
At the very least, I'd say that a custom written demo utilizing the 970's features should be disturbing. The P4's version wouldn't have access to 64-bit calcs, which we'd be sure to use lots of, the P4's version couldn't keep all the computation units filled (since they can't all be used simultaneously), and AV with a lot more bus to play with....
and where is programmer??
i love reading posts by people who are like 800 times smarter than me....
g
<strong>
From what I gather, Bochs has been basically a one-man operation. How much of its sluggishness is just due to inefficient coding (no offense intended to Kevin Lawton, if you're out there)? I would imagine Apple engineers could massage his code to a fare-thee-well...
[ 02-28-2003: Message edited by: TJM ]</strong><hr></blockquote>
As I understand it, Bochs is written in C++. There's no way that's going to run at decent speed without a clean rewrite.
Bochs is so slow it's ridiculous. I can try to launch an old DOS game, and it doesn't even launch before I eventually get fed up and kill the damn thing. Sure, the 970 will be a lot faster than my G4/450, but not that much faster...
<strong>Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.</strong><hr></blockquote>
I think you are missing the real incentive for IBM. The POWER line of CPUs are the brains of mid to high-end servers. Time is showing a huge market for low-end (blade) style servers - this lite version of the POWER CPUs are for that end of the spectrum.
IBM has a huge incentive to stay competitive in their back yard. The 970 will allow them to span the server spectrum from high to low. Apple might have put a chunk of cash into the development of this chip (?) - but IBM sure as hell is reaping some benefits from it as well.
<strong>
I think you are missing the real incentive for IBM. The POWER line of CPUs are the brains of mid to high-end servers. Time is showing a huge market for low-end (blade) style servers - this lite version of the POWER CPUs are for that end of the spectrum.
IBM has a huge incentive to stay competitive in their back yard. The 970 will allow them to span the server spectrum from high to low. Apple might have put a chunk of cash into the development of this chip (?) - but IBM sure as hell is reaping some benefits from it as well.</strong><hr></blockquote>
One, I didn't say that they won't benefit; I just opined that those benefits wouldn't be enough to make IBM go for this expensive proposition.
Two, IBM is committing all it's resources to Linux, and for running a blade server, Linux on Intel works fine. In fact, the Linux/970 solution has a major disadvantage of having just about no available software and only running two distributions (SuSE 8.0 and RedHat 7.1, which by RedHat's own mouth is obsolete). RH may have a release of 8.0/8.1 in the wings, but for now Linux/POWER is immature.
<strong>
As I understand it, Bochs is written in C++. There's no way that's going to run at decent speed without a clean rewrite.
Bochs is so slow it's ridiculous. I can try to launch an old DOS game, and it doesn't even launch before I eventually get fed up and kill the damn thing. Sure, the 970 will be a lot faster than my G4/450, but not that much faster...</strong><hr></blockquote>
Yup, I see that now after nosing around the Bochs site for a while. Obviously, Bochs is not the answer, and the columnists recommending it didn't know what they were talking about ( :eek: I'm shocked! How could that be????
Ah, well. Maybe Apple is writing their own, or perhaps 'tis not to be...
<strong>
Another thought on new suprises:
If the 970 was indeed partially developed for Apple (the proof is when it shows up on apple.com), then what was IBM's incentive to make this processor? I doubt AIX and POWER-line Linux worksation/server sales are hurting right now, especially with the 1.45Ghz POWER 4+ recently made available.
IBM is a business, and recently quite a successful one. Apple, on the other hand, is seen by many investors as being in a precarious position in the market - dependent on an influx of switchers to survive - either from Windows or from their own OS 9. To make a significant new processor designed to specs which put it in competition not with Sun Blade 1000s but with Dell P4 machines indicates that either (1) IBM thinks they can sell a lot of Linux and AIX on a new architecture with no professional apps available, or (2) it was designed for Apple, and perhaps IBM is intending on not just selling Linux on a new architecture, but an existing architecture with existing users and apps - OS X.
Basically, what I'm saying is - if IBM put the money into designing this for Apple (not that they wouldn't get any use out if it themselves, but I don't think they would have made it this ambitious otherwise) then there's some kickback for them in the deal. Whether that's IBM workstations running OS X or something completely different I don't know.
[ 02-28-2003: Message edited by: Anonymous Karma ]</strong><hr></blockquote>
IBM has had their own adgenda with the PowerPC since the beginning of the alliance. I'm not positive, but I would bet one of the main things that is driving their long term strategy is to free themselves from as much outside suppliers as possable, and the PowerPC allows them to do this, It also might bring in more vendors building PowerPC computers than currently exist, which means even more revenue for IBM. IBM looses potential profit (or a market cost advantage over the competition) with every Pentium based PC they sell, they are giving that profit to Intel for their processors and Microsoft for licensing fees.
No more Intel, no more Microsoft. They are again in charge of their destiny. If they have a client that can help bankroll the processor development by buying sufficient quantities (Apple), it is but icing on the cake.
[ 02-28-2003: Message edited by: *l++ ]</p>
<strong>i want more posts from moki....
and where is programmer??
i love reading posts by people who are like 800 times smarter than me....
g</strong><hr></blockquote>
I agree, where is programmer? He has some great insights. And Moki has some great posts but I wish he would spill the beans and say what he expects.
I was wondering how many changes/updates are required to the OS kernel to support the 970 and its companion chip. I do not design computers so I was wondering how long it would take to develop the companion chip and the OS X updates that will make the whole thing run. I have to believe it is more work than going from the 7455 to the 7457, but how much more I haven't a clue.
<strong>everyone sure has gotten their expectations up. There has been no announcement that apple will even use the 970 or anything else for that matter. THey could do something completely arbitruary. Lets get back to earth. Mac users this is Houston... do you read me?</strong><hr></blockquote>
Yah, you could be correct . . . , but Apple would be throwing away alot of developement $$ (ie: since Feb 2000, do the math). Let's just say that an unrefutable source says that if the 970 isn't for Apple, then why were they involved in the chip's developement since it's inception (read 02/2000).
And "Rickag", I tend to agree with "Rhumgod", because the 970 production process testing was half way completed in early November 2002 on 200mm and 300mm wafers.
<strong>
I was wondering how many changes/updates are required to the OS kernel to support the 970 and its companion chip.</strong><hr></blockquote>
That depends if Apple wants to support the 970 in both 32 and 64 bit modes. If they are only interested in the 32bit mode of operations/address space. Probably not that much.
There is a lot more work to do on the hardware side.
- The 970 is quad pumped giving an effective bus of 1.8 Ghz! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> Yes, a bus faster than any G4...
- IBM licenses OSX for some markets. This breaks the long time business objection of Apple being single source.
- The 970 is used across the board. Prices are set more by the box and features than by processor.
- The towers are dual with a quad on top or in a rack.
- The 12" powerbook replaces the iBook. A version with a polycarbonate case is called the eBook.
- Apple engineers a graphics card that has a GPU from ATI or nVidia but also has a G4ish chip: one integer, no fpu, one altivec and a pile of L2. Quartz moves entirely to the graphics board. Now you know what Racer was doing. Code name: Ludicrous Drive.
- The UI becomes totally vector based and thus scalable for different pixel densities.
- Apple and IBM intro high end monitors that are 200ppi.
- Apple intros a headless, affordable, iBriq for the hacker/server/renderfarm crowds.
- Apple engineers a memory controller with a large "L3ish" built in. Memory controller controls it completely. In this way it can optimise it for many processors at once while minimising bus snooping. The "L3" is large enough to preclude the need for external RAM in some of the more modest machines...like the iBriq.
- Bandwidth is so high every machine can channel Steve's RDF in real time. Steve turns into a big glowing ball on stage and ascends to become the first Star Child. Stanley Kubrick rolls in his grave.
[ 02-28-2003: Message edited by: David M ]</p>
<strong>Come on, they just released new ones. That delay is only on the dual 1.4, which is a new chip. The other models say "same day."</strong><hr></blockquote>
Just checked again, and now the dual is 7-10 days and you're right the others are at one day. (I swear they said 3-4 weeks last time I checked...)
<img src="embarrassed.gif" border="0">
- The 970 is quad pumped giving an effective bus of 1.8 Ghz! Yes, a bus faster than any G4...
- IBM licenses OSX for some markets. This breaks the long time business objection of Apple being single source.
- The 970 is used across the board. Prices are set more by the box and features than by processor.
- The towers are dual with a quad on top or in a rack.
- The 12" powerbook replaces the iBook. A version with a polycarbonate case is called the eBook.
- Apple engineers a graphics card that has a GPU from ATI or nVidia but also has a G4ish chip: one integer, no fpu, one altivec and a pile of L2. Quartz moves entirely to the graphics board. Now you know what Racer was doing. Code name: Ludicrous Drive.
- The UI becomes totally vector based and thus scalable for different pixel densities.
- Apple and IBM intro high end monitors that are 200ppi.
- Apple intros a headless, affordable, iBriq for the hacker/server/renderfarm crowds.
- Apple engineers a memory controller with a large "L3ish" built in. Memory controller controls it completely. In this way it can optimise it for many processors at once while minimising bus snooping. The "L3" is large enough to preclude the need for external RAM in some of the more modest machines...like the iBriq.
- Bandwidth is so high every machine can channel Steve's RDF in real time. Steve turns into a big glowing ball on stage and ascends to become the first Star Child. Stanley Kubrick rolls in his grave.
<hr></blockquote>
I gotta see that.
Lemon Bon Bon