And the awnser is: it depends on what you're doing with it. This thing is significantly more tightly coupled than VT's cluster, and uses shared memory as opposed to clustering, so for alot of tightly coupled problems it will be *far* more efficient.
As for raw processing power, the Itanium2 has the same theoretical peak floating point performance as a PPC970 at the same clock. In reality the Itanium is likely to come closer to achieving it's peak than the PPC970 due to it's massive cache (9MB compared to the 970's 512KB). However the Itaniums in an Altix3000 are only running at 1.6Ghz according to SGI's page, while the 970s in VT's cluster are now at 2.3Ghz. So the BigMac would have some advantage on loosely coupled problems that it can fit in it's smaller cache and memory.
So while the BigMac might beat this system at Linpack, the benchmark used to determine the top500, in the domain this system is to be used for (3d modeling of nuclear blasts) it's tighter coupling and greater RAM will make it much faster.
Not going to work. Products like VMware and VirtualPC do not emulate a CPU, they only provide timeslices of (or some other technique of sharing) the real, underlying CPU. This is for performance reasons, and even still, they lose about 20% to overhead.
The upshot of this is that, when running on non-x86 CPU, VMW/VPC cannot host an x86 OS.
The PPC (i.e. Mac) version of VirtualPC most certainly does emulate the CPU, but I suppose you wintel people think that product never existed before MS bought it from Connectix and made a windows version with just the virtuialization system without the emulator.
So go ahead, get moderated "informative" for spouting uninformed opinions, and pretending that mentioning the speed hit I already said they'd incur from emulation is actually your contribution to the discussion.
Oh, yeah. I'm not bitter at all about the constant uninformed but opinionated blathering of so called nerds who wouldn't know anything that doesn't run on an x86 derivative from a hole in their head... *sigh*
I might try it for a bit as a novelty, but there's just no good reason to switch from MacOS or Linux/PPC to WindowsPPC.
Seriously, porting to PPC, or any other platform for that matter, would take away Window's biggest advantages: their wide software and hardware support due to their virtual monoply. Why would anyone want to use windows on an architecture where there are far less windows apps then mac or linux apps? I suppose MS might soften that blow by bundling an integrated version of VirtualPC to run x86 software, but it would still mean a significant speed hit.
I suppose PPCs do have advantages for certain kinds of calculations over current x86 stuff, but if you really want to do that kind of stuff you can allready do it on mac or linux, so there's not much of an opening for MS there. And for everything else you'd do under windows, you can allready do things just as well on x86 hardware..
I hope this doesn't detract from other more ambitious prizes. Just getting to space is nice, but I'd like to see bigger prizes for things like orbital flight, rather than reccuring smaller prizes for doing the same old thing a little bit better than the last guy.
I can imagine a cool concept for the X-Prize version 4.0 (or thereabouts).
Pay some space agency to launch a tiny satelite, just a transponder with a n-million dollar check rolled up inside. The first private team to go up and retrieve it in person keeps it.
Your numbers are right. The Opteron's theoretical peak double precision floating point numbers are pretty mediocre, it gets beaten soundly by Xenons and Itaniums too. It's just not what the processor was designed for, it only has one FPU iirc, where the G5 and most of the other competition has 2.
It is worth noting that these are theoretical numbers, not what you can actually achieve in reality on any given algorithm. The Opteron's Rmax is slightly more competitive(2.9GF/proc at 2Ghz in LosAlamos's Lightning), and the G5's Rmax (4.7GF/processor at 2Ghz in VT's machine) isn't quite as insanely high as it's Rpeak (8GF/proc).
I'd say OpenFirmware has fans because it's been doing most of what EFI promises to do 'real soon now' since the late '80s, and has been doing it as an IEEE standard for a decade.
And of course, there's the fact that OpenFirmware is still the only firmware standard out there with it's own official theme song. Ha!
But the point is who's being paid, not just that everyone is paying.
Does it really seem right to you that Apple pays real to use Real's DRM if Apple takes the offer Glaser implied.
And other companies pay *Real* to use *Apple's* DRM, if Real follows through with it's previous threat to sell Harmony to third parties.
And yet no one pays Apple to use Apple's DRM?
Real is offering to sell the rights to use their DRM, while effectively undercutting Apple from making that same offer by selling the rights to use their hack on Apple's DRM.
Glaser:
We would be happy to cross-license our DRM and formats to Apple to enable exactly the kind of interoperability you propose.
Translation: Apple should pay us for the right to do the same thing we're done to them without paying (and are threatening to sell to others the means to do).
Hello Hipocracy!
Apple provides some pretty versatile generic drivers with the OS. If your hardware is somewhat standards based, there's a pretty decent chance it will just work weather it's officially supported or not. I know my USB card, my mouse, and 2 of my ethernet cards are not officially mac-compatible, but that didn't stop them from working beautifully as soon as I pluged them in and powered on.
I don't know about the Escape, but the Toyota Prius isn't much more expensive than a similar non-hybrid toyota, especially once you take out the $2000 tax credit for owning a low-emission vehicle (although I heard they might have changed the relevant tax law, does anyone have more info on that? If they did, it seems sort of asinine.)
I actually would have gotten a Prius last year, but they were completely sold out in the Chicagoland area. There wasn't a single new Prius to be had for 6-8 months. That's what kept me from having one.
If they started actually advertising their hybrids by more than word-of-mouth, and made sure dealers had at least a demo model to show off, they would sell *a lot* more imho.
What's really odd is that the new GeForce 6800 Ultra that they require to use the 30" Cinema Display HD isn't in that list.
I'm guessing this is a case of the right hand (Core Image team) not talking to the left hand (whoever worked the deal to use the GeForce 6800 Ultra with the display). Hopefully the new card, and thus the big new LCD, will be supported by the time Tiger ships.
It looks and feels just like an iPod! (of course, they don't mention it working, sheesh).
I can just see MS coming out with something the size of an ipod, but with a tiny flash memory instead of a hard drive. And if MS markets it hard enough and makes it ubiquitous enough, the uninformed consumers will just slurp it up and think they're getting a huge bargin (despite there having been flash players that cheap for a long time).
The numbers you quote are the listed average bitrates for all of the songs togeather, the grandparrent seemed to be talking about the bitrates for one particular song/test. He wasn't wrong, but the average numbers you give are certainly more interesting to most of us.
The problem is your average numbers don't add up. If you look at the WMA column in the chart you got those average bitrates from, you'll see only 2 numbers under 128k. That fits better with an average bitrate of 129k for WMA unless you just truncate instead of rounding down..
In any case, iTunes certainly was at a disadvantage in these tests because of it's lower average bitrate than any of the other top 4 codecs.
I think Apple's lossless codec might be MPEG-4-ALS (the lossless codec for MPEG-4).
There are a few reasons to go that way, not the least of which is Apple's heavy backing of MPEG standards. And then there's the format's use of MPEG-4 container files and streams, and the fact that they're probably already paying for licensing it anyway with all their other MPEG-4 stuff in iTunes and Quicktime.
And of course, the guys who developed ALS having the iTunes logo at the top of their website is kind of a hint...
Is it just me, or is bubble fusion a really cool concept?
This is the kind of thing real progress comes from! Not the big nasty brute force machines we've been trying to coax a usable fusion reactor out of for decades, but a clever application of the laws of physics to get tiny pockets of fusion at much more sane average temperatures and pressures. Temperatures we can work with without having to contain them in giant magnetic toruses, temperatures we don't need petawatt lasers to generate for a fraction of a second.
I can see this development panning out, but even if it doesn't I'm still in awe of it's elegance.
Huh? What in particular are you talking about? I'm talking about the great-great-etc-grandparent, post (#8899636). It was whining about how Apple releases it's changes back to the community in bulk rather than in little pre-chewed morsels.
I was under the impression you were defending that post's position, if that impression was mistaken, then nevermind.
There isn't a Cray system that can touch the brute parallel power of a big cluster like Virginia Tech's G5s. But depending on the kind of problem you're working on, there are Cray systems that would walk all over that G5 cluster.
With problems that can be split up into hundreds or thousands of more-or-less independent subtasks, a cluster is the way to go. But for problems that can't be divided up like that, a smaller system with a few very tightly coupled extremely fast vector processors, like what Cray specializes in, is what you need.
There are certainly plenty of HPC problems that aren't well suited for large clusters, but it sounds like the Cray guy might have been significantly overstating his point.
Q: How many arch's can you run OSX on?
A: One (PPC), two if you count the x86 version of Darwin, as many as you want if you feel like porting XNU (it is open source afterall, and very portable (far more so than monolithic linux in fact, since you only need to port the microkernel, and everything else is hardware independent))
Q: How many File Systems do you have to choose from with OSX?
A: HFX, HFS+, UFS, UDF, FAT32, and various other common filesystems work out of the box, and you can (and people do) write your own plugins for other FSes like ext2 and NTFS.
Q: How many desktop UI's do you have available to you?
A: Aqua, X11, and any theme or window manager you feel like running on top of either of those.
Q: How "customizable" is your interface (aqua).
A: Very, you can change the visual elements of your applications without recompiling or any special tools thanks to apple's application packaging system, and with themes and one of the common tools for managing them you can change the appearance of your whole system quite dramatically.
Q: How portable is the cocoa framework?
A: It is based on an OpenStep, a open specification that has been around for quite a while, as has GNUStep, a open source implementation of that spec. In addition, Apple has in the past prototyped it's own "yellowbox" version of Cocoa running on top of Windows.
Q: How many vendor's do you have to choose from if the one you're with takes a direction that you don't like or can't work with?
A: All of them except microsoft really. If you get pissy about buying Apple's OS updates, you can always install linux or BSD or something on your macs.
Slashdot Mirror
It makes headlines on /. every time you upload a new comic.
Hmm... their homepage says they are "a dicom solution".
A versiontracker search for "dicom" under MacOSX returns these programs.
Or you could just use the ubiquitous GraphicConverter which handles just about everything, including dicom images iirc.
And the awnser is: it depends on what you're doing with it.
This thing is significantly more tightly coupled than VT's cluster, and uses shared memory as opposed to clustering, so for alot of tightly coupled problems it will be *far* more efficient.
As for raw processing power, the Itanium2 has the same theoretical peak floating point performance as a PPC970 at the same clock. In reality the Itanium is likely to come closer to achieving it's peak than the PPC970 due to it's massive cache (9MB compared to the 970's 512KB). However the Itaniums in an Altix3000 are only running at 1.6Ghz according to SGI's page, while the 970s in VT's cluster are now at 2.3Ghz. So the BigMac would have some advantage on loosely coupled problems that it can fit in it's smaller cache and memory.
So while the BigMac might beat this system at Linpack, the benchmark used to determine the top500, in the domain this system is to be used for (3d modeling of nuclear blasts) it's tighter coupling and greater RAM will make it much faster.
VT has officially got the BigMac up and running faster than ever at 12.25TF with 1150 dual 2.3Ghz XServes.
Check out the announcment.
I wonder how many Centrises that equates to...
...is going to have to find something better than "an iPod killer" to label themselves.
Originality people!
The PPC (i.e. Mac) version of VirtualPC most certainly does emulate the CPU, but I suppose you wintel people think that product never existed before MS bought it from Connectix and made a windows version with just the virtuialization system without the emulator.
So go ahead, get moderated "informative" for spouting uninformed opinions, and pretending that mentioning the speed hit I already said they'd incur from emulation is actually your contribution to the discussion.
Oh, yeah. I'm not bitter at all about the constant uninformed but opinionated blathering of so called nerds who wouldn't know anything that doesn't run on an x86 derivative from a hole in their head... *sigh*
I might try it for a bit as a novelty, but there's just no good reason to switch from MacOS or Linux/PPC to WindowsPPC.
Seriously, porting to PPC, or any other platform for that matter, would take away Window's biggest advantages: their wide software and hardware support due to their virtual monoply.
Why would anyone want to use windows on an architecture where there are far less windows apps then mac or linux apps? I suppose MS might soften that blow by bundling an integrated version of VirtualPC to run x86 software, but it would still mean a significant speed hit.
I suppose PPCs do have advantages for certain kinds of calculations over current x86 stuff, but if you really want to do that kind of stuff you can allready do it on mac or linux, so there's not much of an opening for MS there.
And for everything else you'd do under windows, you can allready do things just as well on x86 hardware..
I hope this doesn't detract from other more ambitious prizes.
Just getting to space is nice, but I'd like to see bigger prizes for things like orbital flight, rather than reccuring smaller prizes for doing the same old thing a little bit better than the last guy.
I can imagine a cool concept for the X-Prize version 4.0 (or thereabouts).
Pay some space agency to launch a tiny satelite, just a transponder with a n-million dollar check rolled up inside. The first private team to go up and retrieve it in person keeps it.
More like a 3U unit (making it 36U for 12 servers).
Since when are racks 72U high? It's more like 72 inches, or 42U, for the big ones.
Your numbers are right.
The Opteron's theoretical peak double precision floating point numbers are pretty mediocre, it gets beaten soundly by Xenons and Itaniums too. It's just not what the processor was designed for, it only has one FPU iirc, where the G5 and most of the other competition has 2.
It is worth noting that these are theoretical numbers, not what you can actually achieve in reality on any given algorithm. The Opteron's Rmax is slightly more competitive(2.9GF/proc at 2Ghz in LosAlamos's Lightning), and the G5's Rmax (4.7GF/processor at 2Ghz in VT's machine) isn't quite as insanely high as it's Rpeak (8GF/proc).
So now what are we supposed to do the next time the Goa'uld System Lords come to enslave Earth?!
I'd say OpenFirmware has fans because it's been doing most of what EFI promises to do 'real soon now' since the late '80s, and has been doing it as an IEEE standard for a decade.
And of course, there's the fact that OpenFirmware is still the only firmware standard out there with it's own official theme song. Ha!
Because MS and Intel don't control OpenFirmware.
But the point is who's being paid, not just that everyone is paying.
Does it really seem right to you that Apple pays real to use Real's DRM if Apple takes the offer Glaser implied.
And other companies pay *Real* to use *Apple's* DRM, if Real follows through with it's previous threat to sell Harmony to third parties.
And yet no one pays Apple to use Apple's DRM?
Real is offering to sell the rights to use their DRM, while effectively undercutting Apple from making that same offer by selling the rights to use their hack on Apple's DRM.
Apple provides some pretty versatile generic drivers with the OS. If your hardware is somewhat standards based, there's a pretty decent chance it will just work weather it's officially supported or not.
I know my USB card, my mouse, and 2 of my ethernet cards are not officially mac-compatible, but that didn't stop them from working beautifully as soon as I pluged them in and powered on.
I don't know about the Escape, but the Toyota Prius isn't much more expensive than a similar non-hybrid toyota, especially once you take out the $2000 tax credit for owning a low-emission vehicle (although I heard they might have changed the relevant tax law, does anyone have more info on that? If they did, it seems sort of asinine.)
I actually would have gotten a Prius last year, but they were completely sold out in the Chicagoland area. There wasn't a single new Prius to be had for 6-8 months. That's what kept me from having one.
If they started actually advertising their hybrids by more than word-of-mouth, and made sure dealers had at least a demo model to show off, they would sell *a lot* more imho.
What's really odd is that the new GeForce 6800 Ultra that they require to use the 30" Cinema Display HD isn't in that list.
I'm guessing this is a case of the right hand (Core Image team) not talking to the left hand (whoever worked the deal to use the GeForce 6800 Ultra with the display). Hopefully the new card, and thus the big new LCD, will be supported by the time Tiger ships.
It looks and feels just like an iPod! (of course, they don't mention it working, sheesh).
I can just see MS coming out with something the size of an ipod, but with a tiny flash memory instead of a hard drive. And if MS markets it hard enough and makes it ubiquitous enough, the uninformed consumers will just slurp it up and think they're getting a huge bargin (despite there having been flash players that cheap for a long time).
The numbers you quote are the listed average bitrates for all of the songs togeather, the grandparrent seemed to be talking about the bitrates for one particular song/test. He wasn't wrong, but the average numbers you give are certainly more interesting to most of us.
The problem is your average numbers don't add up. If you look at the WMA column in the chart you got those average bitrates from, you'll see only 2 numbers under 128k. That fits better with an average bitrate of 129k for WMA unless you just truncate instead of rounding down..
In any case, iTunes certainly was at a disadvantage in these tests because of it's lower average bitrate than any of the other top 4 codecs.
I think Apple's lossless codec might be MPEG-4-ALS (the lossless codec for MPEG-4).
There are a few reasons to go that way, not the least of which is Apple's heavy backing of MPEG standards. And then there's the format's use of MPEG-4 container files and streams, and the fact that they're probably already paying for licensing it anyway with all their other MPEG-4 stuff in iTunes and Quicktime.
And of course, the guys who developed ALS having the iTunes logo at the top of their website is kind of a hint...
Is it just me, or is bubble fusion a really cool concept?
This is the kind of thing real progress comes from! Not the big nasty brute force machines we've been trying to coax a usable fusion reactor out of for decades, but a clever application of the laws of physics to get tiny pockets of fusion at much more sane average temperatures and pressures. Temperatures we can work with without having to contain them in giant magnetic toruses, temperatures we don't need petawatt lasers to generate for a fraction of a second.
I can see this development panning out, but even if it doesn't I'm still in awe of it's elegance.
Huh? What in particular are you talking about?
I'm talking about the great-great-etc-grandparent, post (#8899636). It was whining about how Apple releases it's changes back to the community in bulk rather than in little pre-chewed morsels.
I was under the impression you were defending that post's position, if that impression was mistaken, then nevermind.
There isn't a Cray system that can touch the brute parallel power of a big cluster like Virginia Tech's G5s. But depending on the kind of problem you're working on, there are Cray systems that would walk all over that G5 cluster.
With problems that can be split up into hundreds or thousands of more-or-less independent subtasks, a cluster is the way to go. But for problems that can't be divided up like that, a smaller system with a few very tightly coupled extremely fast vector processors, like what Cray specializes in, is what you need.
There are certainly plenty of HPC problems that aren't well suited for large clusters, but it sounds like the Cray guy might have been significantly overstating his point.
Q: How many arch's can you run OSX on?
A: One (PPC), two if you count the x86 version of Darwin, as many as you want if you feel like porting XNU (it is open source afterall, and very portable (far more so than monolithic linux in fact, since you only need to port the microkernel, and everything else is hardware independent))
Q: How many File Systems do you have to choose from with OSX?
A: HFX, HFS+, UFS, UDF, FAT32, and various other common filesystems work out of the box, and you can (and people do) write your own plugins for other FSes like ext2 and NTFS.
Q: How many desktop UI's do you have available to you?
A: Aqua, X11, and any theme or window manager you feel like running on top of either of those.
Q: How "customizable" is your interface (aqua).
A: Very, you can change the visual elements of your applications without recompiling or any special tools thanks to apple's application packaging system, and with themes and one of the common tools for managing them you can change the appearance of your whole system quite dramatically.
Q: How portable is the cocoa framework?
A: It is based on an OpenStep, a open specification that has been around for quite a while, as has GNUStep, a open source implementation of that spec. In addition, Apple has in the past prototyped it's own "yellowbox" version of Cocoa running on top of Windows.
Q: How many vendor's do you have to choose from if the one you're with takes a direction that you don't like or can't work with?
A: All of them except microsoft really. If you get pissy about buying Apple's OS updates, you can always install linux or BSD or something on your macs.