G5 Hurting Apple more and more
I think most can agree that IBM is hurting Apple more and more with the delay of faster G5's and with quanity of G5's. The shortage of G5's is forcing Apple to delay products and back order. The delay in development that was promised (3Ghz) is also hurting. They would have gotten so many more orders for a dual 3Ghz at the price of the 2.5. I just think IBM is hurting Apple bad.
Comments
Both Apple & IBM are pushing as hard as they can, but even the 2.5 G5 is very leading edge technology - as evidenced by the liquid cooling. Like the introduction of OS X, the move to IBM is not an "overnight" job. There is significant work being done at all levels, from pushing up the speed of the G5 to 65 nm processing to dual cores. It's a major effort and we're just starting to see the benefits.
It's also important to understand that it is not only IBM working on the leading edge. The 2.5 G5 PM was ready for the 30" display - except for the graphics card.
Originally posted by kenaustus
It's also important to understand that it is not only IBM working on the leading edge. The 2.5 G5 PM was ready for the 30" display - except for the graphics card.
... which NVIDIA was trying to fab at IBM's Fishkill plant.
The move to another ISA would be so jarring for so many reasons that PowerPC would really have to hit a massive crisis before Apple considered it. This isn't a massive crisis.
It's not necessarily even a crisis. If demand for the iMac is as wild as AI recently claimed, the G5 shortage could simply be a matter of Apple not ordering enough of them. The article corroborates that theory by saying that the introduction of the eMac G5 has been pushed back only a quarter or so—not enough to cope with a serious yield issue, but more than enough to buy a lot more production capacity at Fishkill and ramp production up to a new level. In other words, this could be like the hard drive shortage that strangled production of the iPod mini, not like the G4 speed dump that plagued the original PowerMac G4.
Or, it could be like the speed dump. Or, heck, it could be both. We just don't know for sure. IBM has been crowing about the fact that they're finally getting good yields, though, so between that, the AI article, and a hunch, I think the problem is demand outstripping (contracted) supply, not IBM failing to supply what they promised.
damn, and summer is so far away... lots more B&M to come
it's a matter of time. how much is your time worth to you?
if it takes me less time to finish a project because i have a faster machine then it's worth alot and that makes it worth getting a faster machine.
the problem is there are no faster machines on the apple front and what's more frustrating is i can't budget for future purchases or budget time as affectively if apple had a consistent upgrade cycle for their machines.
it took a year to go from 2 ghz to 2.5 ghz and no one knows when or if the next upgrade cycle will come.
chung lee
Originally posted by Altivec_2.0
I think most can agree that IBM is hurting Apple more and more with the delay of faster G5's and with quanity of G5's. The shortage of G5's is forcing Apple to delay products and back order. The delay in development that was promised (3Ghz) is also hurting. They would have gotten so many more orders for a dual 3Ghz at the price of the 2.5. I just think IBM is hurting Apple bad.
This time the competetion isn't really moving either.
Originally posted by chunglee
dirk,
it's a matter of time. how much is your time worth to you?
if it takes me less time to finish a project because i have a faster machine then it's worth alot and that makes it worth getting a faster machine.
This line of reasoning is getting pretty worn out (from being invoked so often as a justification). Unless you are running jobs back to back, 8 hrs a day, with zero setup, a 20% speed boost isn't going to add up to jack in the amount of time it frees up for you. More likely, all that potential is flushed down the drain as the computer is waiting (over billions of clock cycles) for you, the user, to figure out which button to press or how to swoop the mouse for your next operation. If you do happen to be running offline processes back to back, and time is truly money for you, then you should be figuring out which renderfarm-like setup you want to do your job on rather than what workstation. To assert that a 2.5 Ghz G5 isn't "fast enough" for individual use simply defies reason.
Originally posted by Randycat99
This line of reasoning is getting pretty worn out (from being invoked so often as a justification). Unless you are running jobs back to back, 8 hrs a day, with zero setup, a 20% speed boost isn't going to add up to jack in the amount of time it frees up for you. More likely, all that potential is flushed down the drain as the computer is waiting (over billions of clock cycles) for you, the user, to figure out which button to press or how to swoop the mouse for your next operation. If you do happen to be running offline processes back to back, and time is truly money for you, then you should be figuring out which renderfarm-like setup you want to do your job on rather than what workstation. To assert that a 2.5 Ghz G5 isn't "fast enough" for individual use simply defies reason.
Bingo!
Very well put. Most people are concerned about performance at the wrong level in their overall process.
i went from a 933 quicksilver to a 2 ghz dual g5.
what i do is batch process large photoshop files in the thousands of images, i capture video, i render video, i edit video, i author dvd's. i have added additional g5's as my needs increased.
my production significantly improved from switching machines. would i see as significant of an increase in production if i switched to a 2.5 ghz machine. obviously not as much but what about 3ghz or 3.5 or 4 ghz etc....
additionally, the sooner we get faster machines the quicker prices drop for previous machines making them more affordable.
if i had access to a render farm i would use it in a heart beat. btw, how much is a render farm?
if you don't think you will be more productive with a faster machine i have no intention of trying to change your mind. i was giving a testimonial of my personal experience to a question of why one would want or need a faster machine.
the funny thing is you contradicted yourself in your own reply. you are suggesting i won't save time by using a faster machine but then you suggest i should be using a render farm. that i don't get.
you are also more then welcome to come to my studio and batch process 30,000 40mb images from new york fashion week.
you can use my g4 or g5, your choice.
as always,
chung lee
Originally posted by chunglee
randycat,
i went from a 933 quicksilver to a 2 ghz dual g5.
what i do is batch process large photoshop files in the thousands of images, i capture video, i render video, i edit video, i author dvd's. i have added additional g5's as my needs increased.
Then you are already exploiting a "farm" sort of strategy (if you are still keeping these older computers around for some kind of work). Hence your need for brute clockspeed is decoupled. Perhaps you may find even greater economies of scale by considering many more cheaper G4 era machines to funnel your long queue of offline work, rather than fewer, more expensive G5 machines? Logic suggests there will be a sweet spot in there somewhere, and it is less likely that it lies over the "fewer, more expensive machines" range. By simply hoping for faster machines, you are simply begging to ride an even bloodier, bleeding edge. It's that simple.
Quote:
my production significantly improved from switching machines. would i see as significant of an increase in production if i switched to a 2.5 ghz machine. obviously not as much but what about 3ghz or 3.5 or 4 ghz etc....
Economies of scale top out the farther you push against that semi-firm ceiling. Also consider that other components of the computer (motherboard, memory, system bus, network interface, HD throughput, L2 cache size impact, etc.) will not be scaling up to the same extent of mere clockspeed. So some kinds of work instances may become prematurely constrained unrepresented by simple clockspeed increases.
Quote:
additionally, the sooner we get faster machines the quicker prices drop for previous machines making them more affordable.
True, but you seem to be missing out on the largest bargains in older machines by chasing after these "slow", superceded G5 models.
Quote:
if i had access to a render farm i would use it in a heart beat. btw, how much is a render farm?
If you really have that much offline work on queue as you claim, a "farm" of some sort should be what you are looking into if you are serious about optimizing your $'s to CPU time.
Quote:
if you don't think you will be more productive with a faster machine i have no intention of trying to change your mind. i was giving a testimonial of my personal experience to a question of why one would want or need a faster machine.
Sounds like your experience is not exactly indicative in profile (in terms of workflow) of Apple customers at large, yet you would justify a particular product grade for an entire userbase based on your profile? Hmmm...
Quote:
the funny thing is you contradicted yourself in your own reply. you are suggesting i won't save time by using a faster machine but then you suggest i should be using a render farm. that i don't get.
...not a contradiction- simply a key distinction that your real solution lies in multiplicity, not clockrate. Additionally, there was an extra qualification given there that you did not cite, but you seem to fit that qualification. As such, you are a different sort of customer than the mainstream that Apple tries to serve.
Quote:
you are also more then welcome to come to my studio and batch process 30,000 40mb images from new york fashion week.
Shouldn't need to if you really are batch processing. You just got to figure out where and how much to nest your "farm" and how many high speed machines to set aside for personal use by your staff (more so to make'em happy, rather than to save x amount of manhours that have any real existence in practice).
Quote:
you can use my g4 or g5, your choice.
Maybe your setup work practices simply need revision? Then you wouldn't be looking at it in terms of what one machine your work gets done on.
I can understand there is a certain sense of grandeur in attacking with brute force, but that is not always the answer to getting the most bang for your resources, fwiw.
Yeah, it would be nice to be at 3ghz and blowing the competition away. But the laws of physics seem to have other ideas.
that was a long reply and i even understood some of it. somebody should hire you to be systems administrator.
i think what i was replying to was your generalization and dismissiveness in your original reply.
you don't think faster machines increase productiveness (saving time etc..) but i beg to differ. there are numerous other costs and logisitcal reasons you didn't mention.
software licensing, maintenance costs, space, networking etc...
there is an equilibrium point and it's different for everyone.
we can agree to disagree. these are my personal views and my views for my situation only.
as always,
chung lee
p.s.
not trying to get into an argument with you and i appreciate your meticulous reply.
Originally posted by chunglee
randycat,
additionally, the sooner we get faster machines the quicker prices drop for previous machines making them more affordable.
I really don't think this is true anymore. The CPU is only one piece of the price, and it's not cheaper for IBM to make slower chips than faster ones. They will simply stop making the slower chips.
The cost for the FSB, hard drives, pci slots, modern graphics cards all add up to a much larger percentage of the machines price. That's why you can get a fast P4 machine from eMachines for so little. The CPU is fast, but all the other parts are crap and so is the speed. Meanwhile, a decent machine costs a thousand more, has the same processor exactly, but runs circles around the eMachine computer.