The reason SJ was so sure that Apple would be able to offer 3GHz machines this summer is because he probably had a dual 8GHz G5 sitting at home as he said it.
Thus, he knew that 3GHz was do-able - Apple would just have to bide their time whilst IBM figured out how to mass produce them reliably.
Apple probably has stockpiles of the forthcoming parts already - it's just waiting for the right time to make it's move.
Am I correct in saying that when Apple announces a speed bump to an existing line, the new machine are usually available immediately? Isn't usually the new product announcements that we have to wait for?
Yah I'm sure they do have 3ghz around some where. I seriously doubt jobs would have been as bold to say that 3ghz would be ready by EOSUMMER'04 if they weren't going already... I believe he did say they were going. Someone correct me on that one if I'm wrong.
The biggest thing... I have a gut feeling about apple by looking at their past. They like to milk products without updating them as long as possible. Look at the currenty product lines for every computer by apple. Especially the PowerMacs. The powermacs are down in sales over 60% from the first quarter they were released... approaching 8 months since shipment of first duals to consumers (not VT).
If apple releases anything below 3ghz in the next update... I wouldn't hold my breath for 3ghz very long. Apple will want to milk that announcement also. Thats why the more time that goes on...the further we get to 3ghz being announced.
Has anyone seen apple (not before jobs came back) update products less then 6 months apart between announcements?
I don't think the 3 Ghz Powermac G5 will be out this summer. Jobs may want it out in the summer but it all depends on IBM. I think in the Fall. Maybe a 2.5 Ghz G5 upgrade this summer to let the customers be satisfied intill 3 Ghz in the fall.
There are a lot of parts that go into a new system, I wouldn't necessarily lay the blame for delays at IBM's feet. They claim that their success rates on their new processes is significantly better than the initial success rates of their design customers on the same processes. That means that if Apple's new system controller is going 90 nm at the same time as the 970FX, it is much more likely that Apple design is having troubles than IBM's.
Reaching 3 GHz on the 90 nm process will be a stretch for the 970FX, but it might be able to make it. If the yields at that speed are poor I wouldn't be surprised to see Apple price the top end machine out of this world, with the intention of selling most machines in the mid 2 GHz range. If this is the plan then they could probably ship any day between now and WWDC. I wouldn't expect them to go completely without updates for a whole year, and sales are probably dropping because nobody else does either and everybody is waiting for the shoe to drop.
The other possibility is that Jobs was confident because IBM had a new processor design sampling last year at WWDC and it has architectural changes and is on the 90 nm process, which together make it a sure thing to hit 3 GHz. In that case I'd expect to first see a 970FX update into the mid 2 GHz range, followed by the new chip before end of summer. The first update is going to have to happen right away now, or else the releases will be awfully close together.
This is a fairly major transition (although not the huge leap that happened last WWDC), so a bit of engineering slippage is to be expected. For Jobs to say what he did last year, however, implies a very high degree of confidence and he wouldn't have said end-of-summer unless they thought it was achievable (in some form) well before then. They know all about schedule slippage and would have padded their estimates by quite a bit.
That means that if Apple's new system controller is going 90 nm at the same time as the 970FX, it is much more likely that Apple design is having troubles than IBM's.
Unless. of course, IBM is sharing their design tools with Apple which is likely since it appears that IBM is rebranding Apple's northbridge for use in at least one of their 970 blade implementations.
Quote:
The other possibility is that Jobs was confident because IBM had a new processor design sampling last year at WWDC and it has architectural changes and is on the 90 nm process, which together make it a sure thing to hit 3 GHz.
This seems possible if not, at the very least, more likely than before given IBM's end-of-month press conference they have scheduled to talk about the future of the PowerPC ISA.
Unless. of course, IBM is sharing their design tools with Apple which is likely since it appears that IBM is rebranding Apple's northbridge for use in at least one of their 970 blade implementations.
This seems possible if not, at the very least, more likely than before given IBM's end-of-month press conference they have scheduled to talk about the future of the PowerPC ISA.
IIRC the IBM "success rate" figures were for people using their design tools. If those tools are like most other design tools, there is a certain learning curve to their usage.
There is a lot of PPC activity going on right now. Probably more than ever before. It'll be interesting to see how much IBM goes public on...
Instead of IBM clones... Apple clones... only to IBM... wouldn't that be ironic. I know I know... clones don't work with less then 5% market share... but it was an ironic thought
Yah it is amazing how popular ppc is getting. My assembly teacher is considering doing a ppc / 680x0 class. I really think it would be well needed...
Comments
Originally posted by Messiah
The reason SJ was so sure that Apple would be able to offer 3GHz machines this summer is because he probably had a dual 8GHz G5 sitting at home as he said it.
Thus, he knew that 3GHz was do-able - Apple would just have to bide their time whilst IBM figured out how to mass produce them reliably.
Apple probably has stockpiles of the forthcoming parts already - it's just waiting for the right time to make it's move.
Am I correct in saying that when Apple announces a speed bump to an existing line, the new machine are usually available immediately? Isn't usually the new product announcements that we have to wait for?
Yah I'm sure they do have 3ghz around some where. I seriously doubt jobs would have been as bold to say that 3ghz would be ready by EOSUMMER'04 if they weren't going already... I believe he did say they were going. Someone correct me on that one if I'm wrong.
The biggest thing... I have a gut feeling about apple by looking at their past. They like to milk products without updating them as long as possible. Look at the currenty product lines for every computer by apple. Especially the PowerMacs. The powermacs are down in sales over 60% from the first quarter they were released... approaching 8 months since shipment of first duals to consumers (not VT).
If apple releases anything below 3ghz in the next update... I wouldn't hold my breath for 3ghz very long. Apple will want to milk that announcement also. Thats why the more time that goes on...the further we get to 3ghz being announced.
Has anyone seen apple (not before jobs came back) update products less then 6 months apart between announcements?
Reaching 3 GHz on the 90 nm process will be a stretch for the 970FX, but it might be able to make it. If the yields at that speed are poor I wouldn't be surprised to see Apple price the top end machine out of this world, with the intention of selling most machines in the mid 2 GHz range. If this is the plan then they could probably ship any day between now and WWDC. I wouldn't expect them to go completely without updates for a whole year, and sales are probably dropping because nobody else does either and everybody is waiting for the shoe to drop.
The other possibility is that Jobs was confident because IBM had a new processor design sampling last year at WWDC and it has architectural changes and is on the 90 nm process, which together make it a sure thing to hit 3 GHz. In that case I'd expect to first see a 970FX update into the mid 2 GHz range, followed by the new chip before end of summer. The first update is going to have to happen right away now, or else the releases will be awfully close together.
This is a fairly major transition (although not the huge leap that happened last WWDC), so a bit of engineering slippage is to be expected. For Jobs to say what he did last year, however, implies a very high degree of confidence and he wouldn't have said end-of-summer unless they thought it was achievable (in some form) well before then. They know all about schedule slippage and would have padded their estimates by quite a bit.
Originally posted by Programmer
That means that if Apple's new system controller is going 90 nm at the same time as the 970FX, it is much more likely that Apple design is having troubles than IBM's.
Unless. of course, IBM is sharing their design tools with Apple which is likely since it appears that IBM is rebranding Apple's northbridge for use in at least one of their 970 blade implementations.
The other possibility is that Jobs was confident because IBM had a new processor design sampling last year at WWDC and it has architectural changes and is on the 90 nm process, which together make it a sure thing to hit 3 GHz.
This seems possible if not, at the very least, more likely than before given IBM's end-of-month press conference they have scheduled to talk about the future of the PowerPC ISA.
Originally posted by Tomb of the Unknown
Unless. of course, IBM is sharing their design tools with Apple which is likely since it appears that IBM is rebranding Apple's northbridge for use in at least one of their 970 blade implementations.
This seems possible if not, at the very least, more likely than before given IBM's end-of-month press conference they have scheduled to talk about the future of the PowerPC ISA.
IIRC the IBM "success rate" figures were for people using their design tools. If those tools are like most other design tools, there is a certain learning curve to their usage.
There is a lot of PPC activity going on right now. Probably more than ever before. It'll be interesting to see how much IBM goes public on...
Instead of IBM clones... Apple clones... only to IBM... wouldn't that be ironic. I know I know... clones don't work with less then 5% market share... but it was an ironic thought
Yah it is amazing how popular ppc is getting. My assembly teacher is considering doing a ppc / 680x0 class. I really think it would be well needed...