Hey, what about some comparisons of Windows XP and Mac OS X Tiger on the exact same hardware? Will OS X be smoked compared to Windows or not?
There would be something seriously wrong with XP if it didn't do better than the developer machine's OS X--OS X still has months of optimization (especially drivers) ahead of it. It'll probably need even more time to create optimizations that compete with Altivec. That's why my guess is that they'll first release Intel for the low end G4s (mini, iBook) because a faster Intel processor might be fast enough to allow an iDVD encode/burn to be slightly faster than the previous g4 system (the ecode process is highly Altivec optimized).
In the end, this switch is likely to spur OS X to be an even faster beast, as direct compariisons CAN be made (like OpenGL performance, disk read/write performance, UBB & Firewire performance).
Sorry, I just have a REALLY hard time seeing how a G5 with faster/wider bus, AltiVec and dual processors is not just going to run circles around a single P4 in any type of serious data crunching test.
But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
Since most computing tasks that are bandwidth limited are unidrectional, the P4's bus is much faster (almost twice as fast).
And don't forget the massive 2MB cache on the P4 running at 3.6GHz.
There would be something seriously wrong with XP if it didn't do better than the developer machine's OS X--OS X still has months of optimization (especially drivers) ahead of it. It'll probably need even more time to create optimizations that compete with Altivec. That's why my guess is that they'll first release Intel for the low end G4s (mini, iBook) because a faster Intel processor might be fast enough to allow an iDVD encode/burn to be slightly faster than the previous g4 system (the ecode process is highly Altivec optimized).
In the end, this switch is likely to spur OS X to be an even faster beast, as direct compariisons CAN be made (like OpenGL performance, disk read/write performance, UBB & Firewire performance).
Virtually every assumption made here is fundamentally wrong. You would not believe how wrong this is. Wow.
The reason no one is telling you guys the kind of specific info you are looking for is because they really, really can't.
Its not because they don't *know* mind you...its because, well, Apple said so.
What you are reading here however is "fairly accurate". I would not say its "anecdotal" or "a couple of developers".
I don't believe anyone who has or is developing on the x86 version of OS X right now would disagree with the statements made in this piece.
They wouldn't get...specific tho. Publishing benchmarks is strictly and explicitly forbidden.
I think in fact the only folks that don't/won't/can't believe it are those with too vested an interest in "Intel hate".
Whoa, take it easy. I personally ain't lookin' for benchmarks and I know they can't be given out. It just seems that with hundreds of dev boxes out there, we would be hearing more than just this 3 weeks into the program. In comparison to 100's we actually are only getting anecdotal accounts from a few developers.
I am not an Intel hater and I look forward to the Mac using the best h/w platform in the business!
I'm glad to see the developers happy with the speed of the development box, but I see very little relationship between these boxes and the ones that will be shipped. It's sort of like the PMs that were sent to developers working on X-Box games - what they have now is not what they will end up with. I therefore think that you'll be seeing some very fast Mactels. I think we'll be rather pleased, except, of course, for the traditional moaning about the GPU.
Timing? Look for both how Intel is going with their 65nm chips AND how the developers are going with their conversions - especially the major players. As soon as there are a lot of apps available I think only Intel can slow things down.
Looks like it is time for AI to have a sticky list of Mactel ready apps. All Apple apps are supposed to be ready, but I think Mathemetica deserves top spot on the list because of their 2 hour conversion for WWDC.
Don't worry...buy a Yonah Powerbook instead! For most things, a dualcore Yonah PowerBook should outrun Apple's quad-PPC PowerMacs. It will be the first time in 7 years that PowerBooks are faster than PowerMacs.
Yonah laptops running faster than quad-PPC Macs? Let's not exaggerate too much. I can see a single 2.3 GHz Sossaman outrunning a single 2.5 GHz 970mp, but a dual-core laptop outrunning a quad-PPC, that's apples vs oranges.
Thats great and all, but if they have these super Yonah powered PowerBooks wouldn't it make sense for Apple to make some really BEEFY PowerMacs to compete and/or surpass them? For that, I'm willing to bite my lip and wait.
Yeah they should! maybe 4GHz??? If people think that the 3.6 GHz PMs are fast just wait till they sport Dual 4GHz processors.
Cool. I was looking for info on the dev machines last night and couldn't find much.
Is apple sending out updates often? Is there some kind of automatic Software Update or do you have to get them on your own? Is intel OSX polished enough that a new intel system could ship if the hardware were ready?
I'm not very much concerned about browsers and those types of applications. I hope my HP all-in-one printer will be supported on the Intel machines.
This is great news, I'm glad it boots faster. Windows XP has many flaws and it crashes all the time and has lots of spyware, virus, etc., but it boots blazingly fast. I'm glad we're gonna catch up in that area.
I'm not very much concerned about browsers and those types of applications. I hope my HP all-in-one printer will be supported on the Intel machines.
I think support for non-current hardware should be a concern. Unless these companies port the drivers to OS X on Intel, I imagine support just won't be there, unforunately.
I'm not very much concerned about browsers and those types of applications. I hope my HP all-in-one printer will be supported on the Intel machines.
This is great news, I'm glad it boots faster. Windows XP has many flaws and it crashes all the time and has lots of spyware, virus, etc., but it boots blazingly fast. I'm glad we're gonna catch up in that area.
My XP box boots very, very fast to the login screen, and gets to the desktop really fast, but at that point its pretty much still worthless until I wait for the virus software to load, for some other services to start up, etc. I try to show a Windows explorer window (opened on login automatically, not by me manually) when it first gets to its 'ready' state and it takes forever for it to appear, then forever for the sub-folders to appear, and then forever for my inputs to actually take affect.
But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
Since most computing tasks that are bandwidth limited are unidrectional, the P4's bus is much faster (almost twice as fast).
And don't forget the massive 2MB cache on the P4 running at 3.6GHz.
That's incorrect because each processor in a Powermac has its own bus. This isn't true even for Xenon's at this point. They have to share a single bus.
The 800MHz bus is quad pumped. It's really a 200MHz bus. That doesn't detract from it's performance, but some things are misleading.
for example, the G5 bus is unidirectional on each of its directions, but the 800MHz has to share the bandwidth.
[B]But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
I'm not at all understanding the statement "most computing tasks that are bandwidth limited are unidrectional". I can't imagine bandwidth sensitive applications where you simply suck vast amounts of data IN to a processor and have little to no output. I'd welcome education on that.
Your also ignoring that the G5 system's each have two independent front-side busses. Each bus is capable of 3.6 up AND down simultaneously. That's a total inbound of 7.2GB/s and total outbound of 7.2GB/s or a grand total of 14.4GB/s of data in-flight to and from the G5s. Compared to the grand total of 6.4GB/s on the P4.
The G5s can saturate memory, AGP and storage busses with their bandwidth capacity.
That particular comment makes little sense. It's not like Safari pegs the Mac's CPU(s), so there's absolutely no way that could ever be a valid benchmark of microprocessor performance.
I never suggested that web browsers would be a valid benchmark or that the hack as anything to do with CPU utilization.
For what it's worth, what I've heard (from the horses mouth), it is really too early to tell how performance compares.
Heh...sure it is.
If your er, horse, is um, running code on both platforms that they compiled and are running side by side then your horse probably knows better than that
I never suggested that web browsers would be a valid benchmark or that the hack as anything to do with CPU utilization.
It isn't an indication of how fast the cpu's work. It's an indication as to how more efficient the routines in the OS and program are in an x86 system.
If your er, horse, is um, running code on both platforms that they compiled and are running side by side then your horse probably knows better than that
Considering that neither the OS or the box itself is ready yet, and likely doesn't reflect what the final units will be, it can't indicate more than a vague idea of performance.
Comments
Originally posted by vas
Hey, what about some comparisons of Windows XP and Mac OS X Tiger on the exact same hardware? Will OS X be smoked compared to Windows or not?
There would be something seriously wrong with XP if it didn't do better than the developer machine's OS X--OS X still has months of optimization (especially drivers) ahead of it. It'll probably need even more time to create optimizations that compete with Altivec. That's why my guess is that they'll first release Intel for the low end G4s (mini, iBook) because a faster Intel processor might be fast enough to allow an iDVD encode/burn to be slightly faster than the previous g4 system (the ecode process is highly Altivec optimized).
In the end, this switch is likely to spur OS X to be an even faster beast, as direct compariisons CAN be made (like OpenGL performance, disk read/write performance, UBB & Firewire performance).
Originally posted by Gerardj
Sorry, I just have a REALLY hard time seeing how a G5 with faster/wider bus, AltiVec and dual processors is not just going to run circles around a single P4 in any type of serious data crunching test.
But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
Since most computing tasks that are bandwidth limited are unidrectional, the P4's bus is much faster (almost twice as fast).
And don't forget the massive 2MB cache on the P4 running at 3.6GHz.
Originally posted by mcdawson
There would be something seriously wrong with XP if it didn't do better than the developer machine's OS X--OS X still has months of optimization (especially drivers) ahead of it. It'll probably need even more time to create optimizations that compete with Altivec. That's why my guess is that they'll first release Intel for the low end G4s (mini, iBook) because a faster Intel processor might be fast enough to allow an iDVD encode/burn to be slightly faster than the previous g4 system (the ecode process is highly Altivec optimized).
In the end, this switch is likely to spur OS X to be an even faster beast, as direct compariisons CAN be made (like OpenGL performance, disk read/write performance, UBB & Firewire performance).
Virtually every assumption made here is fundamentally wrong. You would not believe how wrong this is. Wow.
Low end? Unoptimized?
Heh.
Originally posted by RnSK
The reason no one is telling you guys the kind of specific info you are looking for is because they really, really can't.
Its not because they don't *know* mind you...its because, well, Apple said so.
What you are reading here however is "fairly accurate". I would not say its "anecdotal" or "a couple of developers".
I don't believe anyone who has or is developing on the x86 version of OS X right now would disagree with the statements made in this piece.
They wouldn't get...specific tho. Publishing benchmarks is strictly and explicitly forbidden.
I think in fact the only folks that don't/won't/can't believe it are those with too vested an interest in "Intel hate".
Whoa, take it easy. I personally ain't lookin' for benchmarks and I know they can't be given out. It just seems that with hundreds of dev boxes out there, we would be hearing more than just this 3 weeks into the program. In comparison to 100's we actually are only getting anecdotal accounts from a few developers.
I am not an Intel hater and I look forward to the Mac using the best h/w platform in the business!
Thinking that a Hyperthreaded P4@3.6GHz can't possibly surpass say, a 2x2Ghz G5 Powermac would be...
Inaccurate.
Timing? Look for both how Intel is going with their 65nm chips AND how the developers are going with their conversions - especially the major players. As soon as there are a lot of apps available I think only Intel can slow things down.
Looks like it is time for AI to have a sticky list of Mactel ready apps. All Apple apps are supposed to be ready, but I think Mathemetica deserves top spot on the list because of their 2 hour conversion for WWDC.
Originally posted by Existence
Don't worry...buy a Yonah Powerbook instead! For most things, a dualcore Yonah PowerBook should outrun Apple's quad-PPC PowerMacs. It will be the first time in 7 years that PowerBooks are faster than PowerMacs.
Yonah laptops running faster than quad-PPC Macs? Let's not exaggerate too much. I can see a single 2.3 GHz Sossaman outrunning a single 2.5 GHz 970mp, but a dual-core laptop outrunning a quad-PPC, that's apples vs oranges.
Originally posted by baranovich
Thats great and all, but if they have these super Yonah powered PowerBooks wouldn't it make sense for Apple to make some really BEEFY PowerMacs to compete and/or surpass them? For that, I'm willing to bite my lip and wait.
Yeah they should! maybe 4GHz??? If people think that the 3.6 GHz PMs are fast just wait till they sport Dual 4GHz processors.
Is apple sending out updates often? Is there some kind of automatic Software Update or do you have to get them on your own? Is intel OSX polished enough that a new intel system could ship if the hardware were ready?
This is great news, I'm glad it boots faster. Windows XP has many flaws and it crashes all the time and has lots of spyware, virus, etc., but it boots blazingly fast. I'm glad we're gonna catch up in that area.
Originally posted by monkeyastronaut
I'm not very much concerned about browsers and those types of applications. I hope my HP all-in-one printer will be supported on the Intel machines.
I think support for non-current hardware should be a concern. Unless these companies port the drivers to OS X on Intel, I imagine support just won't be there, unforunately.
Originally posted by monkeyastronaut
I'm not very much concerned about browsers and those types of applications. I hope my HP all-in-one printer will be supported on the Intel machines.
This is great news, I'm glad it boots faster. Windows XP has many flaws and it crashes all the time and has lots of spyware, virus, etc., but it boots blazingly fast. I'm glad we're gonna catch up in that area.
My XP box boots very, very fast to the login screen, and gets to the desktop really fast, but at that point its pretty much still worthless until I wait for the virus software to load, for some other services to start up, etc. I try to show a Windows explorer window (opened on login automatically, not by me manually) when it first gets to its 'ready' state and it takes forever for it to appear, then forever for the sub-folders to appear, and then forever for my inputs to actually take affect.
And this is a 3.2GHz box with very little on it.
Originally posted by Existence
But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
Since most computing tasks that are bandwidth limited are unidrectional, the P4's bus is much faster (almost twice as fast).
And don't forget the massive 2MB cache on the P4 running at 3.6GHz.
That's incorrect because each processor in a Powermac has its own bus. This isn't true even for Xenon's at this point. They have to share a single bus.
The 800MHz bus is quad pumped. It's really a 200MHz bus. That doesn't detract from it's performance, but some things are misleading.
for example, the G5 bus is unidirectional on each of its directions, but the 800MHz has to share the bandwidth.
[B]But the G5 doesn't have a "faster/wider bus". The P4's bus is a unidirection 64bit-wide bus running at 800MHz (6.4 GBps up or down). The G5's bus in a 2.0GHz model is a 1.00GHz bi-directional 32-bit bus (3.6 GBps up and/or 3.6 GBps down, 0.8GBps overhead).
I'm not at all understanding the statement "most computing tasks that are bandwidth limited are unidrectional". I can't imagine bandwidth sensitive applications where you simply suck vast amounts of data IN to a processor and have little to no output. I'd welcome education on that.
Your also ignoring that the G5 system's each have two independent front-side busses. Each bus is capable of 3.6 up AND down simultaneously. That's a total inbound of 7.2GB/s and total outbound of 7.2GB/s or a grand total of 14.4GB/s of data in-flight to and from the G5s. Compared to the grand total of 6.4GB/s on the P4.
The G5s can saturate memory, AGP and storage busses with their bandwidth capacity.
Originally posted by Big Mac
That particular comment makes little sense. It's not like Safari pegs the Mac's CPU(s), so there's absolutely no way that could ever be a valid benchmark of microprocessor performance.
I never suggested that web browsers would be a valid benchmark or that the hack as anything to do with CPU utilization.
Originally posted by skatman
I'm curious if OSX for Intel supports hyperthreading?
This is a hardware feature, not a software. OSX sees the processor as one physical processor but two virtual. OSX uses both.
On to something not entirely different..
Mach-O is tailored fot CISC-processors. gcc is better for x86 than it is for PPC.
What will this do for extra performance on x86?
Originally posted by tink
For what it's worth, what I've heard (from the horses mouth), it is really too early to tell how performance compares.
Heh...sure it is.
If your er, horse, is um, running code on both platforms that they compiled and are running side by side then your horse probably knows better than that
Originally posted by Gerardj
I never suggested that web browsers would be a valid benchmark or that the hack as anything to do with CPU utilization.
It isn't an indication of how fast the cpu's work. It's an indication as to how more efficient the routines in the OS and program are in an x86 system.
Originally posted by RnSK
Heh...sure it is.
If your er, horse, is um, running code on both platforms that they compiled and are running side by side then your horse probably knows better than that
Considering that neither the OS or the box itself is ready yet, and likely doesn't reflect what the final units will be, it can't indicate more than a vague idea of performance.