What amazes me on a continual basis is the ability of some posters here to technically analize (without documentation or source code) the implementation of a wild a$$ rumor of a technology that people spend years on studying and refute such "_implementation_" due to technical difficulties, physics, the phases of the moon or what ever!
You know who you are... Please wait until you get through some university courses and a couple of actual projects before you engage in technical analysis that would be well beyond your depth under the best of circumstances. UFB!
okay this thread gets all over the place, so im not sure where to begin.
scott or brisby, have you ever assembled a cluster, im going to assume you'll answer no for the sake of argument. ive done it several times, using existing tools (pooch and linda on OSX, a few beowulf clusters as well) i also work on an sgi machine at work, but that's irrelevant. if i was going to set up a cluster again, it would be on a mac hands down. mac clustering is totally plug an play, unlike beowulf clusters that are extremely finicky. beowulf clusters have lots problems associated with them, which would warrant its own thread. clustering on Windoze, forget it, only if Gates came down to the office to admin. the network himself.
arguments made by any one about setting up a mac cluster being really time consuming are totally baseless. adding new machines to an existing cluster takes minutes, minutes not hours or days as is the case with a beowulf cluster. now if apple introduced a way to set up a cluster by checking a box then that time would be measured in seconds.
is there a market for mac clustering, HELLS yeah. if you don't think so, then you sir are a fool of exasperating magnitude. just because you don't have a use for it in your work place, dosn't mean that i don't or that allot of other might. oh and making statements about REAL CLUSTERS, any cluster that cranks out 1.5 teraflops ( <a href="http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html"; target="_blank">http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html</a>; ) IS a real cluster. mac clusters are quietly becoming all the rage in the sci/edu market, because they are cheaper to run and maintain. when your done clustering them you give 'em over to the secretaries, and drop the new hardware into the cluster.
is this technology for consumers, no way. if that's the point your trying to make, then shut up, that point should be obvious to any one that knows what clustering is for. if some one wanted to set up a cluster in their home for no particular reason, then by all means cluster away.
is this allen cat legit, i don't care if his next post is about apple fielding a quantom computer in 2004, it makes for good brain candy. nuff said.
is this a legit concept for apple to integrate into OSX server, yeah and its probably just left over zilla anyway.
<strong>okay this thread gets all over the place, so im not sure where to begin.
scott or brisby, have you ever assembled a cluster, im going to assume you'll answer no for the sake of argument. ive done it several times, using existing tools (pooch and linda on OSX, a few beowulf clusters as well) i also work on an sgi machine at work, but that's irrelevant. if i was going to set up a cluster again, it would be on a mac hands down. mac clustering is totally plug an play, unlike beowulf clusters that are extremely finicky. beowulf clusters have lots problems associated with them, which would warrant its own thread. clustering on Windoze, forget it, only if Gates came down to the office to admin. the network himself.
arguments made by any one about setting up a mac cluster being really time consuming are totally baseless. adding new machines to an existing cluster takes minutes, minutes not hours or days as is the case with a beowulf cluster. now if apple introduced a way to set up a cluster by checking a box then that time would be measured in seconds.
is there a market for mac clustering, HELLS yeah. if you don't think so, then you sir are a fool of exasperating magnitude. just because you don't have a use for it in your work place, dosn't mean that i don't or that allot of other might. oh and making statements about REAL CLUSTERS, any cluster that cranks out 1.5 teraflops ( <a href="http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html"; target="_blank">http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html</a>; ) IS a real cluster. mac clusters are quietly becoming all the rage in the sci/edu market, because they are cheaper to run and maintain. when your done clustering them you give 'em over to the secretaries, and drop the new hardware into the cluster.
is this technology for consumers, no way. if that's the point your trying to make, then shut up, that point should be obvious to any one that knows what clustering is for. if some one wanted to set up a cluster in their home for no particular reason, then by all means cluster away.
is this allen cat legit, i don't care if his next post is about apple fielding a quantom computer in 2004, it makes for good brain candy. nuff said.
is this a legit concept for apple to integrate into OSX server, yeah and its probably just left over zilla anyway.</strong><hr></blockquote>
Goodness. No comment neccesary. Your name says it all...
How about seeing an Apple-hosted and maintained cluster serving up the latest in online video games...?!?
Now that would be a non-scientific/educational/corporate usage of clustering, and would be a great boon to Apple for PR, and to Mac users for a solid source for online gaming...
And I would bet that an Apple-hosted and maintained rendering cluster could be rented out via the Internet for use by the new wave of digital artists and content creators... (https://apple.xRender.com/login.html)
Imagine using a cost-effective option for rendering out Maya scenes, just login and run the render on Apple's clusters...
OR! Imagine if Pixar licensed RenderMan to Apple, and Maya users could send RIBs out across the net for processing...
Just thinking differently...
<img src="graemlins/smokin.gif" border="0" alt="[Chilling]" /> Maya for Mac OS X <img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />
[quote]BTW, I totally agree with scott_h_phd. At, MY university it is the same way.
Folks who NEED clusters, BUY clusters.
(and for programmers sake....)
Folks who NEED clusters, and can't afford clusters, BUY a newspaper and start hunting for another line of work. []/quote]
You know, I'm getting fed up with hearing this shiznit about "the pros" at your Universities using "real" clusters. I work for a large space physics scientific computing center here at CU, and having heard the potential for this technology, we've come to a general conclusion that you guys are idiots, or don't actually work in this field.
This technology would bring enormous power to server machines -- and that means Xserve. It also brings enormous power to the workstations as well. With this technology built in to the kernel, we no longer would have to deal with the pain and delays of recompiling our code using multi-processor compilation tools like mpicc, just as an example. That does add a lot of overhead to our already strapped time, and we have an enormous code base which is almost always changing. Imagine if we can run our stable compiled product on 32 dedicated nodes of Xserves running this technology.
That is seriously good news for scientific computing. Why don't you guys go research the disadvantages with standard UNIX clustering (and by the way, especially when it comes to speed, maintenance and support, Xserve is the way to go these days) and look at the innovation in front of your eyes?
How about seeing an Apple-hosted and maintained cluster serving up the latest in online video games...?!?
Now that would be a non-scientific/educational/corporate usage of clustering, and would be a great boon to Apple for PR, and to Mac users for a solid source for online gaming...
And I would bet that an Apple-hosted and maintained rendering cluster could be rented out via the Internet for use by the new wave of digital artists and content creators... (https://apple.xRender.com/login.html)
</strong><hr></blockquote>
Unfortunately video games are very sensitive to latency, and that is one thing you tend to get lots of in a cluster environment.
And while I agree this could be a real boon to scientific computing on Macs, I'm not convinced that it will work on existing code as-is. I personally think there is more to be gained by defining a new or modified development model which is friendlier to clustering and distributed computing than the standard C/C++/Fortran/etc development model. It could also be hardware independent. The current model is just too tied to the hardware and a low-level, and doesn't account well enough for new features of the hardware (SIMD, cache controls etc). The current threading model relies on finely grained shared memory and, VMS notwithstanding, that simply cannot be lightning fast unless you are very careful not to touch the same data pages with multiple threads.
This project "Wolf" thing sounds an awful lot like<a href="http://www.csm.ornl.gov/pvm/pvm_home.html"; target="_blank">PVM</a>(Parallel Virtual Machine) or .<a href="http://www-unix.mcs.anl.gov/mpi/"; target="_blank">MPI</a>(Message Passing Interface). Both of these software packages have been around on the Unix side for years.
perhaps we might see blade servers from apple in the near future. imagine a dp-G4 with 2 gig's of ram and a 100 gig hard drive on a 6" blade. you could have a 4RU chassis that held 10 or more, now thats performance per sq. foot. apple could market one hell of a blade server, and i would buy that over a bunch of xserves. call it x-blade, set up a render farm on that baby, or map high energy plasma dynamics or whatever floats yer boat.
food for thought, allot of market analysts think blades are the wave of the future and all that. i could see this technology becoming very popular indeed.
(snip)Xserve is the way to go these days) and look at the innovation in front of your eyes?
sym</strong><hr></blockquote>
Am not a feline. And of course I ramble.
In previous posts you may see that I AGREE about Xserves being the way to go. This technology could make them much more valuable.
As for being idiots or not working in the field? Neither are true, (but I am no expert at clustering). I do work around the clusters every day though. And I also manage the 3D G4 Lab in the Art Building. I still don't find anyone able to support that this Kernel Level code changing would actually improve things for every application. This technology is to sophisticated to make part of all of Apple's product line. Would you share your processor over a 56k dialup line? Would a network of eMacs really whoop down on a nice Sun Cluster? I really don't think that any consumer level implementation of this would be practicle or efficient. The arguments are about bandwidth, networking know how, job distribution, and much much more. Not to mention the notion of 'sharing' is detestible in a University environment.
Labs and such may exist where collective computing might help, but Apple isn't going to change the way we work (again). This is a Xserver/Selective Application software.
This is no personal argument and I hope you can keep this to the facts. Try addressing the other issue here. Mainly Presley....
Okay, an Optical Mouse, (like we have now) but which has a center rocker switch which senses a 'tilt left' as a left click, and a 'tilt right' as a right click. I pose a major problem with this. I'd wager a huge percentage of users who have been screaming for a 2 button mouse have been doing so for gaming purposes. How handy will it be to be in a game rocking a mouse left in right? To jump and then rapidly fire you'd have to quickly go from one extreme to another rather then simply a 'choice application' of your finger to the appropriate mouse button. I think at 'tilting mouse' would be a hinderance and I personally will stick with a LogicTech optical if such a tilting mouse appears.
So lets finish this thread up. I believe (pending the validity of these fake rumors) that the Wolf upgrade to software will be for OS X Server as oppposed to standard OS X. If it is applied to OS X which we all use it will only work for a select 'few' programs (ie, iShake). These programs will be sold to professionals and basically keep the confusion out of the consumer market's hair. I also believe that presley will have a huge ammount of criticism after it comes out and that a large ammount of people who wished for a 2 button mac mouse will be dissappointed and return to a 1 button mouse which won't accidentally click wrong if you rock your hand incorrectly.
oh yeah the mouse, well ive never really needed more than one button to be honest. indeed brisby, the two button mouse seems a little unlikely. given apple stessing ease of use in OSX, the only needing one button thing. it could be some option, to make "switchers" feel at home. i can see the ads now, "bring in your hated little box, well upgrade you to the two button mouse". honestly i hope not, the current ads make me "switch" channels when they come on.
As for being idiots or not working in the field? Neither are true, (but I am no expert at clustering). I do work around the clusters every day though. And I also manage the 3D G4 Lab in the Art Building. I still don't find anyone able to support that this Kernel Level code changing would actually improve things for every application. This technology is to sophisticated to make part of all of Apple's product line. Would you share your processor over a 56k dialup line? Would a network of eMacs really whoop down on a nice Sun Cluster? I really don't think that any consumer level implementation of this would be practicle or efficient. The arguments are about bandwidth, networking know how, job distribution, and much much more. Not to mention the notion of 'sharing' is detestible in a University environment.
Labs and such may exist where collective computing might help, but Apple isn't going to change the way we work (again). This is a Xserver/Selective Application software.
</strong><hr></blockquote>
Hmmm, you seem to suggest that any such WOLF would only really be useful in a few, select environments. That might at first be true.
In my lab we have two 32 x 1.7GHz Linux clusters running our CFD code. Currently, it is not possible to justify an Apple solution due to the poor performance/cost ratio. Trust me, it was one of my projects to evaluate.
However. In the office there are another 50 or so assorted very high performance computers (including two Macs...) mostly belonging to individuals and functioning to provide pre- and post-processing, visualisation, and all the other mundane things such as text editing etc. Now if we were able to harness the mostly wasted cycles of those other 50 computers with an all Mac/OS X solution, the entire economics of the situation would be flipped.
Scientific computing is big. It is high-profile and it is high-margins. This alone is reasonable incentive for Apple.
Now as to how this would spread outside of scientific computing, I'll leave that up to you to speculate. In the meantime, I can't help but realise that the chaps in the biology department have over 1000 G4s mostly sitting idle...
I´ll try and check back here in a few days if I can. Currently in Europe at an internet cafe. This thread has become big and I currently don´t have the time to read all posters now. A meeting soon. But I will read them all and reply.
I will give you some history on what I know to all. The information that has been provided to me comes from the ¨ones¨ in the know. I do believe some of the info has been beefed up, but I´ve tried to post only what I know. I´ve filtered together information to give you a picture of the technology. Whether these rumors are true or not, I believe they are 100% REAL, based on the ´ones´ that provided it to me.
The most interesting codename that has been tossed around is ´David´s Stone´ This one is big, some think its the big one. I think that´s why it´s been coined with other projects. Its very frustrating for me not to have all the information to put together but I´m doing the best I can.
There a many many many many projects going on at Apple. Trust me on this, there is a road map for the future and its being executed when the technology is right to implement. Timing is the key factor for products. To Fran, there is a PDA like device coming. I have some details but not enough where I feel its good to post yet.
Could this information be disinformation? I thought at times it could be. Honestly, after the Nexus show when Glove didn´t come out, I thought maybe I was Apple´s vehicle to throw wrinkles out on the boards. But I don´t think this is the case.
Presley has NO Buttons. I just been informed its an optional upgrade mouse (pro market)
Off to a meeting. And by the way, my profession is the business of information.
[quote]my profession is the business of information<hr></blockquote>
alanmcjones, if you're really working for "the business of information" it shouldn't be that hard to get ya... <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" />
Comments
<strong>So if this clustering technology is true, does that mean you guys will go away?
</strong><hr></blockquote>
You should be so lucky.
I don't work on Macs actually, but I want a new one at home to do cool stuff with.
<strong>
You should be so lucky.
I don't work on Macs actually, but I want a new one at home to do cool stuff with.</strong><hr></blockquote>
What platform do you work on, out of curiosity? What kind of programming do you do?
<strong>
What platform do you work on, out of curiosity? What kind of programming do you do?</strong><hr></blockquote>
He does the 'naked kind'. (if you don't know, don't ask)
You know who you are... Please wait until you get through some university courses and a couple of actual projects before you engage in technical analysis that would be well beyond your depth under the best of circumstances. UFB!
<strong>iCode, go ahead and name names!</strong><hr></blockquote>
<img src="graemlins/oyvey.gif" border="0" alt="[No]" />
iCode
scott or brisby, have you ever assembled a cluster, im going to assume you'll answer no for the sake of argument. ive done it several times, using existing tools (pooch and linda on OSX, a few beowulf clusters as well) i also work on an sgi machine at work, but that's irrelevant. if i was going to set up a cluster again, it would be on a mac hands down. mac clustering is totally plug an play, unlike beowulf clusters that are extremely finicky. beowulf clusters have lots problems associated with them, which would warrant its own thread. clustering on Windoze, forget it, only if Gates came down to the office to admin. the network himself.
arguments made by any one about setting up a mac cluster being really time consuming are totally baseless. adding new machines to an existing cluster takes minutes, minutes not hours or days as is the case with a beowulf cluster. now if apple introduced a way to set up a cluster by checking a box then that time would be measured in seconds.
is there a market for mac clustering, HELLS yeah. if you don't think so, then you sir are a fool of exasperating magnitude. just because you don't have a use for it in your work place, dosn't mean that i don't or that allot of other might. oh and making statements about REAL CLUSTERS, any cluster that cranks out 1.5 teraflops ( <a href="http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html"; target="_blank">http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html</a>; ) IS a real cluster. mac clusters are quietly becoming all the rage in the sci/edu market, because they are cheaper to run and maintain. when your done clustering them you give 'em over to the secretaries, and drop the new hardware into the cluster.
is this technology for consumers, no way. if that's the point your trying to make, then shut up, that point should be obvious to any one that knows what clustering is for. if some one wanted to set up a cluster in their home for no particular reason, then by all means cluster away.
is this allen cat legit, i don't care if his next post is about apple fielding a quantom computer in 2004, it makes for good brain candy. nuff said.
is this a legit concept for apple to integrate into OSX server, yeah and its probably just left over zilla anyway.
<strong>What kind of programming do you do?</strong><hr></blockquote>
Computer programming.
<strong>okay this thread gets all over the place, so im not sure where to begin.
scott or brisby, have you ever assembled a cluster, im going to assume you'll answer no for the sake of argument. ive done it several times, using existing tools (pooch and linda on OSX, a few beowulf clusters as well) i also work on an sgi machine at work, but that's irrelevant. if i was going to set up a cluster again, it would be on a mac hands down. mac clustering is totally plug an play, unlike beowulf clusters that are extremely finicky. beowulf clusters have lots problems associated with them, which would warrant its own thread. clustering on Windoze, forget it, only if Gates came down to the office to admin. the network himself.
arguments made by any one about setting up a mac cluster being really time consuming are totally baseless. adding new machines to an existing cluster takes minutes, minutes not hours or days as is the case with a beowulf cluster. now if apple introduced a way to set up a cluster by checking a box then that time would be measured in seconds.
is there a market for mac clustering, HELLS yeah. if you don't think so, then you sir are a fool of exasperating magnitude. just because you don't have a use for it in your work place, dosn't mean that i don't or that allot of other might. oh and making statements about REAL CLUSTERS, any cluster that cranks out 1.5 teraflops ( <a href="http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html"; target="_blank">http://www.daugerresearch.com/fractaldemos/USCCluster/USCMacClusterBenchmark.html</a>; ) IS a real cluster. mac clusters are quietly becoming all the rage in the sci/edu market, because they are cheaper to run and maintain. when your done clustering them you give 'em over to the secretaries, and drop the new hardware into the cluster.
is this technology for consumers, no way. if that's the point your trying to make, then shut up, that point should be obvious to any one that knows what clustering is for. if some one wanted to set up a cluster in their home for no particular reason, then by all means cluster away.
is this allen cat legit, i don't care if his next post is about apple fielding a quantom computer in 2004, it makes for good brain candy. nuff said.
is this a legit concept for apple to integrate into OSX server, yeah and its probably just left over zilla anyway.</strong><hr></blockquote>
Goodness. No comment neccesary. Your name says it all...
[ 06-25-2002: Message edited by: Da sinister ]</p>
How about seeing an Apple-hosted and maintained cluster serving up the latest in online video games...?!?
Now that would be a non-scientific/educational/corporate usage of clustering, and would be a great boon to Apple for PR, and to Mac users for a solid source for online gaming...
And I would bet that an Apple-hosted and maintained rendering cluster could be rented out via the Internet for use by the new wave of digital artists and content creators... (https://apple.xRender.com/login.html)
Imagine using a cost-effective option for rendering out Maya scenes, just login and run the render on Apple's clusters...
OR! Imagine if Pixar licensed RenderMan to Apple, and Maya users could send RIBs out across the net for processing...
Just thinking differently...
<img src="graemlins/smokin.gif" border="0" alt="[Chilling]" /> Maya for Mac OS X <img src="graemlins/smokin.gif" border="0" alt="[Chilling]" />
[quote]BTW, I totally agree with scott_h_phd. At, MY university it is the same way.
Folks who NEED clusters, BUY clusters.
(and for programmers sake....)
Folks who NEED clusters, and can't afford clusters, BUY a newspaper and start hunting for another line of work. []/quote]
You know, I'm getting fed up with hearing this shiznit about "the pros" at your Universities using "real" clusters. I work for a large space physics scientific computing center here at CU, and having heard the potential for this technology, we've come to a general conclusion that you guys are idiots, or don't actually work in this field.
This technology would bring enormous power to server machines -- and that means Xserve. It also brings enormous power to the workstations as well. With this technology built in to the kernel, we no longer would have to deal with the pain and delays of recompiling our code using multi-processor compilation tools like mpicc, just as an example. That does add a lot of overhead to our already strapped time, and we have an enormous code base which is almost always changing. Imagine if we can run our stable compiled product on 32 dedicated nodes of Xserves running this technology.
That is seriously good news for scientific computing. Why don't you guys go research the disadvantages with standard UNIX clustering (and by the way, especially when it comes to speed, maintenance and support, Xserve is the way to go these days) and look at the innovation in front of your eyes?
sym
<strong>Alternative usages of clustering...
How about seeing an Apple-hosted and maintained cluster serving up the latest in online video games...?!?
Now that would be a non-scientific/educational/corporate usage of clustering, and would be a great boon to Apple for PR, and to Mac users for a solid source for online gaming...
And I would bet that an Apple-hosted and maintained rendering cluster could be rented out via the Internet for use by the new wave of digital artists and content creators... (https://apple.xRender.com/login.html)
</strong><hr></blockquote>
Unfortunately video games are very sensitive to latency, and that is one thing you tend to get lots of in a cluster environment.
And while I agree this could be a real boon to scientific computing on Macs, I'm not convinced that it will work on existing code as-is. I personally think there is more to be gained by defining a new or modified development model which is friendlier to clustering and distributed computing than the standard C/C++/Fortran/etc development model. It could also be hardware independent. The current model is just too tied to the hardware and a low-level, and doesn't account well enough for new features of the hardware (SIMD, cache controls etc). The current threading model relies on finely grained shared memory and, VMS notwithstanding, that simply cannot be lightning fast unless you are very careful not to touch the same data pages with multiple threads.
food for thought, allot of market analysts think blades are the wave of the future and all that. i could see this technology becoming very popular indeed.
<strong>That cat Brisby rambled:
(snip)Xserve is the way to go these days) and look at the innovation in front of your eyes?
sym</strong><hr></blockquote>
Am not a feline. And of course I ramble.
In previous posts you may see that I AGREE about Xserves being the way to go. This technology could make them much more valuable.
As for being idiots or not working in the field? Neither are true, (but I am no expert at clustering). I do work around the clusters every day though. And I also manage the 3D G4 Lab in the Art Building. I still don't find anyone able to support that this Kernel Level code changing would actually improve things for every application. This technology is to sophisticated to make part of all of Apple's product line. Would you share your processor over a 56k dialup line? Would a network of eMacs really whoop down on a nice Sun Cluster? I really don't think that any consumer level implementation of this would be practicle or efficient. The arguments are about bandwidth, networking know how, job distribution, and much much more. Not to mention the notion of 'sharing' is detestible in a University environment.
Labs and such may exist where collective computing might help, but Apple isn't going to change the way we work (again). This is a Xserver/Selective Application software.
This is no personal argument and I hope you can keep this to the facts. Try addressing the other issue here. Mainly Presley....
Okay, an Optical Mouse, (like we have now) but which has a center rocker switch which senses a 'tilt left' as a left click, and a 'tilt right' as a right click. I pose a major problem with this. I'd wager a huge percentage of users who have been screaming for a 2 button mouse have been doing so for gaming purposes. How handy will it be to be in a game rocking a mouse left in right? To jump and then rapidly fire you'd have to quickly go from one extreme to another rather then simply a 'choice application' of your finger to the appropriate mouse button. I think at 'tilting mouse' would be a hinderance and I personally will stick with a LogicTech optical if such a tilting mouse appears.
So lets finish this thread up. I believe (pending the validity of these fake rumors) that the Wolf upgrade to software will be for OS X Server as oppposed to standard OS X. If it is applied to OS X which we all use it will only work for a select 'few' programs (ie, iShake). These programs will be sold to professionals and basically keep the confusion out of the consumer market's hair. I also believe that presley will have a huge ammount of criticism after it comes out and that a large ammount of people who wished for a 2 button mac mouse will be dissappointed and return to a 1 button mouse which won't accidentally click wrong if you rock your hand incorrectly.
That's what I've collected. Other thoughts folks?
<strong>
<snip>
As for being idiots or not working in the field? Neither are true, (but I am no expert at clustering). I do work around the clusters every day though. And I also manage the 3D G4 Lab in the Art Building. I still don't find anyone able to support that this Kernel Level code changing would actually improve things for every application. This technology is to sophisticated to make part of all of Apple's product line. Would you share your processor over a 56k dialup line? Would a network of eMacs really whoop down on a nice Sun Cluster? I really don't think that any consumer level implementation of this would be practicle or efficient. The arguments are about bandwidth, networking know how, job distribution, and much much more. Not to mention the notion of 'sharing' is detestible in a University environment.
Labs and such may exist where collective computing might help, but Apple isn't going to change the way we work (again). This is a Xserver/Selective Application software.
</strong><hr></blockquote>
Hmmm, you seem to suggest that any such WOLF would only really be useful in a few, select environments. That might at first be true.
In my lab we have two 32 x 1.7GHz Linux clusters running our CFD code. Currently, it is not possible to justify an Apple solution due to the poor performance/cost ratio. Trust me, it was one of my projects to evaluate.
However. In the office there are another 50 or so assorted very high performance computers (including two Macs...) mostly belonging to individuals and functioning to provide pre- and post-processing, visualisation, and all the other mundane things such as text editing etc. Now if we were able to harness the mostly wasted cycles of those other 50 computers with an all Mac/OS X solution, the entire economics of the situation would be flipped.
Scientific computing is big. It is high-profile and it is high-margins. This alone is reasonable incentive for Apple.
Now as to how this would spread outside of scientific computing, I'll leave that up to you to speculate. In the meantime, I can't help but realise that the chaps in the biology department have over 1000 G4s mostly sitting idle...
CGI
I will give you some history on what I know to all. The information that has been provided to me comes from the ¨ones¨ in the know. I do believe some of the info has been beefed up, but I´ve tried to post only what I know. I´ve filtered together information to give you a picture of the technology. Whether these rumors are true or not, I believe they are 100% REAL, based on the ´ones´ that provided it to me.
The most interesting codename that has been tossed around is ´David´s Stone´ This one is big, some think its the big one. I think that´s why it´s been coined with other projects. Its very frustrating for me not to have all the information to put together but I´m doing the best I can.
There a many many many many projects going on at Apple. Trust me on this, there is a road map for the future and its being executed when the technology is right to implement. Timing is the key factor for products. To Fran, there is a PDA like device coming. I have some details but not enough where I feel its good to post yet.
Could this information be disinformation? I thought at times it could be. Honestly, after the Nexus show when Glove didn´t come out, I thought maybe I was Apple´s vehicle to throw wrinkles out on the boards. But I don´t think this is the case.
Presley has NO Buttons. I just been informed its an optional upgrade mouse (pro market)
Off to a meeting. And by the way, my profession is the business of information.
-Allen
alanmcjones, if you're really working for "the business of information" it shouldn't be that hard to get ya... <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" />
[ 06-27-2002: Message edited by: Strangelove ]</p>