If these new machines do anything to address the bandwidth issues across MPX then the rumoured high-end quad processor machine becomes plausible... and I'd expect to see that before I'd expect to see a completely new chip next week.</strong><hr></blockquote>
Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too. hmmm...
<strong>Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.</strong><hr></blockquote>
I hadn't even been thinking a single dual core chip, actually... just a really expensive high-end model with 4 "7470s" (or whatever they are actually called).
[quote]<strong>
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too.
</strong><hr></blockquote>
The big apps like Photoshop, Final Cut Pro, etc can probably spawn a thread per processor, thus leveraging a quad machine quite effectively. Anything that chops up a single large computation (so that a thread handles each part) will do well. Apps that have multiple different tasks will be limited by the number of tasks that they need done, and most of these will have been designed with just a dual processor in mind. Moving into the future, however, I think developers will need to think in terms of having more and more processors.
<strong> [quote]
hmmm...
</strong><hr></blockquote>
You're doing that a lot, it is really quite distracting.
Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too. hmmm...</strong><hr></blockquote>
Apple's been chanting at people for a couple of years now about the virtue of threads etc. I mean, Apple's high end offering has been a Dual for how long now? Anything that _really_ needs to go fast should be well on its way to utilizing the available processors.
And getting use out of 4 CPUs when serious thought has been applied to the 2 CPU version... shouldn't be too hard, should it? Make anything intensive an independent thread and let the OS deal with it gets a long way.
I think it's pretty clear that the niche Apple's been buying up companies in could use all the cycles it can get for instance
The one clear drawback to a Quad (over and above price even) is the completely inadequate MPX bus. Another 2 CPUs would just cut the useful memory bandwidth effectively in half again. Either a new chip from Mot or one from IBM would seem to address this via either RapidIO or Hypertransport.... I'm not sure a Quad wouldn't still be memory-throughput starved though.
It is already in current shipping Macs.</strong><hr></blockquote>
I was talking about how some board members inteperated your (somtimes) vague posts.
That is, that the Apollo would see a pipeline extension and greater execution units, like the 7400 ->7450.
The Apollo, as it turned out, was an SOI 7450 (primarily).
I'm not taking a swipe at you, just satirizing the board members who exagurate the message, mix metaphores and shot the messenger when their mac dreams don't materialize.
Also are you REALLY trying to compare the technology used in SAHARA with this latest CPU?!?! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> </strong><hr></blockquote>
Perhaps someone has already responded to this, but I haven't read further yet. Anyway, I think they were comparing the circumstances, not the actual technology.
Wrong. ILM switched from SGI to Intel/Linux machines in their render farms. They did not switch from Apple. Their reasons for switching has nothing to do with Apple.
*snip!*</strong><hr></blockquote>
Any why did it have nothing to do with Apple? I think the point here is that they didn't "switch" to Apple. I seriously question whether they were ever even in contention. Think that bugs Steve? Think it bugs the PowerMac product line managers? I do. That would have been a huge win both monetarily and in PR. The fact that Apple can't play with the big boys yet merely demonstrates how badly they do need a processor like this new one.
Personally, if it isn't available now, six months will leave Apple heaving a big "Whew!" having managed to stay competative, just. Twelve months will leave Apple in the dust. Again.
Any why did it have nothing to do with Apple? I think the point here is that they didn't "switch" to Apple. I seriously question whether they were ever even in contention. Think that bugs Steve? Think it bugs the PowerMac product line managers? I do. That would have been a huge win both monetarily and in PR. The fact that Apple can't play with the big boys yet merely demonstrates how badly they do need a processor like this new one.
Personally, if it isn't available now, six months will leave Apple heaving a big "Whew!" having managed to stay competative, just. Twelve months will leave Apple in the dust. Again. </strong><hr></blockquote>
OS X still can't compete with Linux as a system for performance in some areas yet even ignoring the chips. That will improve with 10.2 but certainly they have work still to do.
As for the chips themselves you don't know anything about them. You can't say what will happen when they are actually released or where Apple will stand with respect to performance.
OS X still can't compete with Linux as a system for performance in some areas yet even ignoring the chips. That will improve with 10.2 but certainly they have work still to do.</strong><hr></blockquote>
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).
[quote]Actually, as long as you are wishing for data why not wish for the Spec numbers <hr></blockquote>
How about those Spec numbers posted on The Register a while back? (OK, they were a bit dodgy, as they're higher than the Power4 and scale superlinearly with core frequency... (but that could just be experimental variance)).
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).</strong><hr></blockquote>
It competes well with Windows and MacOS on the desktop!
It's all true ATM, but Linux is the only viable direct competitor to Microsoft in the desktop space. No other business plan can compete with a monopoly.
Of course, Linux doesn't have a business plan, it has a developer user base and the GPL, surpassing any possible business plan.
I think Linux is definatly on its way to surpassing Windows in terms of ease-of-set-up.
Mandrake (8.2), for example, won't work (easily) with my Studio Display. I've hit a bug dual booting it off the same hard drive as Mac OS X. KDE sucks. However, once I had set it up with a CRT monitor, and 2 hard disks (too lazy to fix the bug myself), it was and is a joy to use! A commercial system with similar functionality would have cost $1000s, yet the Mandrake PPC CD Set cost me a mere $75 Australian all up!
If these 2 bugs were eliminated, Linux would have been easier for me than Windows. Mandrake Install is far easier and more powerful than Windows XP! As each year goes by, Linux becomes a greater potential and realised force in the desktop market.
And while I would like to see Apple with a >10% market-share, it's not like another monopoly would be a good thing. Linux can never be a monopoly due to the fact that it is community developed, with miriad packages packages in to hundreds of targeted distributions, I can't think of a better OS for the <90%.
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).</strong><hr></blockquote>
Don't take what I said wrong I love OS X as a desktop system and I believe it has a very promising life ahead of it.
They do need to keep working on those areas where it still lags though and I would be surprised if they aren't working on them already.
<strong>It's all true ATM, but Linux is the only viable direct competitor to Microsoft in the desktop space. No other business plan can compete with a monopoly.</strong><hr></blockquote>
I'd really love to get into debating this -- but I think it'd take the thread too far off of topic. Suffice it to say that Linux is FREE, yet no one wants to use it on their desktop. Literally no one (statistically speaking) -- that speaks volumes, IMHO.
It either needs to improve radically (which seems unlikely, given the Linux crowd's propensity to get into heated debates over which commandline text editor is best, which of the hundred window managers is best, which toolkit is best, etc.), or they need to go beyond free and start _paying_ people to use it.
[quote] I'd really love to get into debating this -- but I think it'd take the thread too far off of topic. Suffice it to say that Linux is FREE, yet no one wants to use it on their desktop. Literally no one (statistically speaking) -- that speaks volumes, IMHO. <hr></blockquote>
How's Linux free? You still have to go to a store like CompUSA and pay for whaver version, just like Mac OS or Windows. Explain to me how Linux is quote on quote "free".
<strong>How's Linux free? You still have to go to a store like CompUSA and pay for whaver version, just like Mac OS or Windows. Explain to me how Linux is quote on quote "free".</strong><hr></blockquote>
<strong>How's Linux free? You still have to go to a store like CompUSA and pay for whaver version, just like Mac OS or Windows. Explain to me how Linux is quote on quote "free".</strong><hr></blockquote>
You can download it without directly paying anything. -> Monetary cost = 0 -> FREE. You can either download the source and compile it yourself, or some places you can download binaries.
However, this all assumes that time is not free, which is just not the case. So I do basically agree that it's not completely free.
When you have a render farm and can simply duplicate the disk 42 times (or however many), that's a tremendous advantage over paying $$$ and installing software with serial number checks etc.
With Linux, it is clear that duplicating the OS across multiple disks this way is perfectly ok.
With Darwin, it's pretty much the same, although not as widely done.
With Mac OS X, it's clearly _not_ ok, although Apple is not aggressive about really preventing it.
With Doze, not only is it clearly not ok, there's a lot of security/authorization checking going on, and there's a watchdog group funded by MS which is reputedly very irritating to interact with even if you do have the correct licenses etc.
Comments
<strong>
If these new machines do anything to address the bandwidth issues across MPX then the rumoured high-end quad processor machine becomes plausible... and I'd expect to see that before I'd expect to see a completely new chip next week.</strong><hr></blockquote>
Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too. hmmm...
<strong>Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.</strong><hr></blockquote>
I hadn't even been thinking a single dual core chip, actually... just a really expensive high-end model with 4 "7470s" (or whatever they are actually called).
[quote]<strong>
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too.
</strong><hr></blockquote>
The big apps like Photoshop, Final Cut Pro, etc can probably spawn a thread per processor, thus leveraging a quad machine quite effectively. Anything that chops up a single large computation (so that a thread handles each part) will do well. Apps that have multiple different tasks will be limited by the number of tasks that they need done, and most of these will have been designed with just a dual processor in mind. Moving into the future, however, I think developers will need to think in terms of having more and more processors.
<strong> [quote]
hmmm...
</strong><hr></blockquote>
You're doing that a lot, it is really quite distracting.
<strong>
Yeah, a dual core might make sense -- then there would be no such thing as a single processor PowerMac, and the possibility of a quad processor one would be very real.
I don't know, though -- the diminishing returns on a desktop computer for a quad processor machine make it sound a little unlikely to me. Most people would rather have a specific task run faster than have a specific task bog down their entire machine less.
Not to mention that the only programs that would benefit from a quad processor machine directly would be those that are multiple processor aware -- usually high end stuff... though Apple has purchased some rather high end software companies recently, too. hmmm...</strong><hr></blockquote>
Apple's been chanting at people for a couple of years now about the virtue of threads etc. I mean, Apple's high end offering has been a Dual for how long now? Anything that _really_ needs to go fast should be well on its way to utilizing the available processors.
And getting use out of 4 CPUs when serious thought has been applied to the 2 CPU version... shouldn't be too hard, should it? Make anything intensive an independent thread and let the OS deal with it gets a long way.
I think it's pretty clear that the niche Apple's been buying up companies in could use all the cycles it can get for instance
The one clear drawback to a Quad (over and above price even) is the completely inadequate MPX bus. Another 2 CPUs would just cut the useful memory bandwidth effectively in half again. Either a new chip from Mot or one from IBM would seem to address this via either RapidIO or Hypertransport.... I'm not sure a Quad wouldn't still be memory-throughput starved though.
<strong>
It is already in current shipping Macs.</strong><hr></blockquote>
I was talking about how some board members inteperated your (somtimes) vague posts.
That is, that the Apollo would see a pipeline extension and greater execution units, like the 7400 ->7450.
The Apollo, as it turned out, was an SOI 7450 (primarily).
I'm not taking a swipe at you, just satirizing the board members who exagurate the message, mix metaphores and shot the messenger when their mac dreams don't materialize.
Barto
<strong>
Also are you REALLY trying to compare the technology used in SAHARA with this latest CPU?!?! <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> </strong><hr></blockquote>
Perhaps someone has already responded to this, but I haven't read further yet. Anyway, I think they were comparing the circumstances, not the actual technology.
Just me though.
<strong>
Wrong. ILM switched from SGI to Intel/Linux machines in their render farms. They did not switch from Apple. Their reasons for switching has nothing to do with Apple.
*snip!*</strong><hr></blockquote>
Any why did it have nothing to do with Apple? I think the point here is that they didn't "switch" to Apple. I seriously question whether they were ever even in contention. Think that bugs Steve? Think it bugs the PowerMac product line managers? I do. That would have been a huge win both monetarily and in PR. The fact that Apple can't play with the big boys yet merely demonstrates how badly they do need a processor like this new one.
Personally, if it isn't available now, six months will leave Apple heaving a big "Whew!" having managed to stay competative, just. Twelve months will leave Apple in the dust. Again.
<strong>
Any why did it have nothing to do with Apple? I think the point here is that they didn't "switch" to Apple. I seriously question whether they were ever even in contention. Think that bugs Steve? Think it bugs the PowerMac product line managers? I do. That would have been a huge win both monetarily and in PR. The fact that Apple can't play with the big boys yet merely demonstrates how badly they do need a processor like this new one.
Personally, if it isn't available now, six months will leave Apple heaving a big "Whew!" having managed to stay competative, just. Twelve months will leave Apple in the dust. Again.
OS X still can't compete with Linux as a system for performance in some areas yet even ignoring the chips. That will improve with 10.2 but certainly they have work still to do.
As for the chips themselves you don't know anything about them. You can't say what will happen when they are actually released or where Apple will stand with respect to performance.
<strong>
OS X still can't compete with Linux as a system for performance in some areas yet even ignoring the chips. That will improve with 10.2 but certainly they have work still to do.</strong><hr></blockquote>
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).
How about those Spec numbers posted on The Register a while back? (OK, they were a bit dodgy, as they're higher than the Power4 and scale superlinearly with core frequency... (but that could just be experimental variance)).
<strong>
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).</strong><hr></blockquote>
It competes well with Windows and MacOS on the desktop!
2.0 and 7.0, that is.
Of course, Linux doesn't have a business plan, it has a developer user base and the GPL, surpassing any possible business plan.
I think Linux is definatly on its way to surpassing Windows in terms of ease-of-set-up.
Mandrake (8.2), for example, won't work (easily) with my Studio Display. I've hit a bug dual booting it off the same hard drive as Mac OS X. KDE sucks. However, once I had set it up with a CRT monitor, and 2 hard disks (too lazy to fix the bug myself), it was and is a joy to use! A commercial system with similar functionality would have cost $1000s, yet the Mandrake PPC CD Set cost me a mere $75 Australian all up!
If these 2 bugs were eliminated, Linux would have been easier for me than Windows. Mandrake Install is far easier and more powerful than Windows XP! As each year goes by, Linux becomes a greater potential and realised force in the desktop market.
And while I would like to see Apple with a >10% market-share, it's not like another monopoly would be a good thing. Linux can never be a monopoly due to the fact that it is community developed, with miriad packages packages in to hundreds of targeted distributions, I can't think of a better OS for the <90%.
Barto
Please refer to <a href="http://www.silicon.com"; target="_blank">www.silicon.com</a> if this link does'nt work for cookie-related reasons.
<strong>
sure, but Linux can't compete with _anything_ in the desktop OS arena -- which IBM claims this chip is geared towards (as well as low-end servers).</strong><hr></blockquote>
Don't take what I said wrong I love OS X as a desktop system and I believe it has a very promising life ahead of it.
They do need to keep working on those areas where it still lags though and I would be surprised if they aren't working on them already.
<strong>It's all true ATM, but Linux is the only viable direct competitor to Microsoft in the desktop space. No other business plan can compete with a monopoly.</strong><hr></blockquote>
I'd really love to get into debating this -- but I think it'd take the thread too far off of topic. Suffice it to say that Linux is FREE, yet no one wants to use it on their desktop. Literally no one (statistically speaking) -- that speaks volumes, IMHO.
It either needs to improve radically (which seems unlikely, given the Linux crowd's propensity to get into heated debates over which commandline text editor is best, which of the hundred window managers is best, which toolkit is best, etc.), or they need to go beyond free and start _paying_ people to use it.
Maybe start another topic and we can debate this?
::segue into PPC processor debate::
[ 08-12-2002: Message edited by: moki ]</p>
very well said.
ps. linux people don't buy software either.
<strong>How's Linux free? You still have to go to a store like CompUSA and pay for whaver version, just like Mac OS or Windows. Explain to me how Linux is quote on quote "free".</strong><hr></blockquote>
Download
<strong>
Download</strong><hr></blockquote>
Hey then everything is free
(Yes I know what you mean)
<strong>How's Linux free? You still have to go to a store like CompUSA and pay for whaver version, just like Mac OS or Windows. Explain to me how Linux is quote on quote "free".</strong><hr></blockquote>
You can download it without directly paying anything. -> Monetary cost = 0 -> FREE. You can either download the source and compile it yourself, or some places you can download binaries.
However, this all assumes that time is not free, which is just not the case. So I do basically agree that it's not completely free.
When you have a render farm and can simply duplicate the disk 42 times (or however many), that's a tremendous advantage over paying $$$ and installing software with serial number checks etc.
With Linux, it is clear that duplicating the OS across multiple disks this way is perfectly ok.
With Darwin, it's pretty much the same, although not as widely done.
With Mac OS X, it's clearly _not_ ok, although Apple is not aggressive about really preventing it.
With Doze, not only is it clearly not ok, there's a lot of security/authorization checking going on, and there's a watchdog group funded by MS which is reputedly very irritating to interact with even if you do have the correct licenses etc.
Linux is free only if your time is free.