As we all know, it is practically traditional for Apple to upgrade their notebook lines during the summertime (mostly shortly before / after WWDC). At least with the upgrade of the white macbooks, we know that they won't be deviating from their annual schedule! I would say that their consistency so far this summer in upgrading the notebooks (and today's screen upgrade) is an indicator that they are sticking with their practice.
Unless they upgrade battery with considerably larger one and put adequate P-series CPU instead of T's, all these so-called memory and HD "upgrades" will be a sick joke and not worth even mentioning.
(I suppose you didn't mean 256mb PB22-J's as possible HD bump cause that'll be cooooooll)
Intel has had a mobile quad-core suitable for the iMac or MBP since late last year. Very expensive processors, though.
Expensive and somewhat slower cores if we are talking about the same thing. Apple apparently decided to focus on core clock rate instead of more but slightly slower cores. That is likely good for a good portion of Apples user base but does nothing for people who run apps that could benefit from more cores.
What Apple should gave at this point is alternatives. That is people who know that they would benefit from quad cores could manage an upgrade. Apple has upgrades for everything else in the MBP so why not a update for cores available.
Quote:
I think quads will wait until the "Clarksfield" processor around the end of this year (probably later rather than sooner). Which will still be expensive, but it'll be Apple's only option if they want to stay competitive.
apple needs to do more than just stay competitive, they need a platform that addresses the needs of power users. Of course even the desktop line has that huge gulf where many customers are left wanting.
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.
Please don't think I'm yelling in the title. It is just that I believe many have the wrong idea with respect to Snow Leopard. They have this idea that eveything will be dramatically faster. It's better to say that a portion of the common user programs have the potential to be dramatically faster.
Quote:
Originally Posted by Sybaritic
I'm assuming that this is because Snow Leopard will offer significant speed improvements for multi-core machines but only modest ones for dual-core?
There are two things that come to mind. One is that some software will work dramatically better with the additional cores. The second thing is that new features mean more background processes to support them.
However it isn't as simple as that as SL has OpenCL which means that some threads may not see the CPUs and instead run on the GPUs. So while SL is going a long way to improve multithreading one can't say for sure where all those threads will be running. In any event there is already software available for the Mac that can leverage additional cores beyond two. That puts the MBP at a disadvantage already, it will just get worst when software becomes SL targetted.
Software for SL can likely be divided into three groups. That would be: 1. Software that will benefit very little if any, 2. Software that will see improved performance immediately and 3. Software that will need to be rewritten.
The second group is most interesting in my mind because there is a bit of guess work here. Some will perform better due to better multi processing / threading support. Others will benefit from improved system libraries or other features the app makes use of. Hopefully we will see some surprises.
At the moment I believe the bigest set of apps are in group #3. That is the apps will need new releases to really benefit from SL. I suspect a good portion of these apps will be rewritten with OpenCL. In mind. This the potential for really dramatic performance increases. Take this to mean that initially SL might not look all that impressive. We will have to wait for software revs before the benefits are truely obvious. Maybe more than one rev will be required.
In otherwords the benefits of quad cores and SL may not be immediate for all users. Apples system software is likely to impress though. Really though I see dual core CPUs as a thing of the past.
That's a bit pessimistic isn't it? If SL is faster and has more features on the same hardware you can't complain. A Cocoa finder will benefit just about every Mac user SL can be installed on! We've been waiting a long time that's for sure!!
Quote:
Originally Posted by wizard69
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.
However it isn't as simple as that as SL has OpenCL which means that some threads may not see the CPUs and instead run on the GPUs.
The MacBook Pro has two GPUs.
So if you're running SL on a current MacBook Pro you'll have both CPU cores and both GPU cores. Now, obviously, this is only potential, but any application that realizes it will be significantly accelerated. GPUs do what they do very, very well.
If the MBP is plugged in, the 9400M is just sitting there waiting for something to do. Snow Leopard can use it as a coprocessor--with the results depending, again, on how many tasks are suited to computation on a GPU, and how well they're suited.
I wouldn't say that it's a complete loss. Frankly, the people buying quad cores now generally fall into two camps: either they need as many cores as they can get their hands on or they have two or three idle cores 90% of the time and bragging rights. The transition to software that is truly designed around multicore processing will be slow and painful, with the biggest bread-and-butter applications coming to the party last.
Well if you are lucky enough own one of the recent models it does.
Quote:
So if you're running SL on a current MacBook Pro you'll have both CPU cores and both GPU cores. Now, obviously, this is only potential, but any application that realizes it will be significantly accelerated. GPUs do what they do very, very well.
True but the real important part of that statement above is "GPUs do what they do very, very well". If your multithreaded code simply doesn't run well on a GPU then your speed up might not be that great. Generally though even slow hardware cores are better than time slicing across one or two cores.
Quote:
If the MBP is plugged in, the 9400M is just sitting there waiting for something to do. Snow Leopard can use it as a coprocessor--with the results depending, again, on how many tasks are suited to computation on a GPU, and how well they're suited.
Its is the question of suitability that has me wondering just how well a GPU can be leveraged. The way I understand OpenCL (granted little formal training here) is that the Mac only hands off routines to the cores on the GPU, you aren't really running a full thread there. I could be wrong here as I've had no development experience what so ever with this feature.
Quote:
I wouldn't say that it's a complete loss. Frankly, the people buying quad cores now generally fall into two camps: either they need as many cores as they can get their hands on or they have two or three idle cores 90% of the time and bragging rights.
Yes exactly! For the people that fall into that first camp they don't have an Apple alternative to run their apps on. At least not a low cost or portable machine. So it is a complete lost for people that have a demand for that type of hardware.
As to the future I believe that SL will make far better use of those cores even for the person that falls into camp two. It is like the move to dual core processors where many of the same arguments where made, in general though dual core lead to more responsive machines for the general user thus it went from bragging rights to being mainstream simply because it offered a better experience. At times that better experience came at the cost of single threaded performance, it was the overall gain that drove the demand for dual cores though. Especially after OS'es and apps became better adapted to the capability.
So I see SL driving the demand of Quad cores in a similar manner. Yeah it depends just how well certain apps can leverage the GPU, which is already there. Interestingly benchmarking is going to be very interesting in the future as not only will the CPU have an impact but so will the GPU. With the wide variety of GPU's, finding out which platform is best for an app will take a lot of testing.
Quote:
The transition to software that is truly designed around multicore processing will be slow and painful, with the biggest bread-and-butter applications coming to the party last.
Well no not really as some software already exists and other apps will never convert to heavy threading. GPU leveraging may or may not be easy for a given app. I don't see individual apps as the big driver here anyways, it is the capability to do multiple things at once that will make such machines interesting.
I only hope that they will update the MBP line soon after the WWDC2009.
And hopefully they terminate that aweful G9400M/9600M GT and replace it with
some of these:
GeForce GTX 280M (optional 17" MBP)
GeForce GTX 260M (standard 17" MBP)
GeForce GTS 160M (optional 15" MBP)
GeForce GTS 150M (standard 15" MBP)
Do you really want 60W and 75W GPU in MBP's?
They won't do this obviously:
1) Battery life. 15" MBP with current battery won't be able to last even 1 hour under normal load. Less than 1.5 hours when idle and < 30min under full load (3D games, ect)
2) 4500-5000 rpm when idle with current heatsink and super-hot enclosure.
3) In order to provide efficient cooling and acceptable battery life (btw only current 17" unibody is good enough) they'll have to put at least 200Wh battery and much heavier heatsink. Each MPB will become 50% thicker and 1kg heavier.
And GTX260M is only 70-80% faster than GT 130M. Still not enough for modern 3D games. In other words it's not worth it.
Only 70-80% faster? That's almost twice as fast...
But your powerconsumption evaluation is sound. Hadn't thought of it to be honoust. As long as they upgrade the GPU to something better then the 9400M/9600M GT I'm satisfied.
Only 70-80% faster? That's almost twice as fast...
But your powerconsumption evaluation is sound. Hadn't thought of it to be honoust. As long as they upgrade the GPU to something better then the 9400M/9600M GT I'm satisfied.
Yes, almost twice as fast than mid-range 130M GT (130M GT is basically overclocked 9600M GT, approx. 20% faster).
But both desktop GTX285 and GTX295 are respectively 2.5x and 3.6x faster than mobile 280M (based on 3DMark Vantage tests)! That's why current mobile GPUs are obsolete for hardcore gaming (and thas's what 280M is intended for)
On the other hand, 9600M GT (and 130M GT) are most power efficient solutions for mid-sized notebooks.
So, what's the consensus-- update timeframe another 3 months?
I doubt there is a firm consensus. Last summer a bunch of us predicted that Apple would release the mobile Penryn unibody in mid-July. They didn't get around to the 17" machine until February of 2009. Needless to say, I was wrong, as were quite a few others.
Who wouldn't love to see a refresh this month? Still, it looks like Apple have a lot on the proverbial plate right now. Intel just released updated mobile cpus, but we have seen no news of limited supply when it comes to the current Macbook Pros. That's usually the most telling sign of a near-term upgrade.
On the other hand, the recent change in the 13" Macbook is salient evidence for a sooner-rather-than-later refresh — and unlike last year, the unibody is already a reality, so sometime in the next three months sounds about right. I'd love for the change to come this week, though.
? Intel just released updated mobile cpus, but we have seen no news of limited supply when it comes to the current Macbook Pros. That's usually the most telling sign of a near-term upgrade.
?
Cross fingers.
Ah, yes, stock lead time has been VERY telling. I'm wondering if I should just wait, or buy now. Yeah, I know; I said it.
Comments
--Memory upgrade
--HD upgrade
--Memory upgrade (not happening in MBP)
--CPU upgrade
(I suppose you didn't mean 256mb PB22-J's as possible HD bump cause that'll be cooooooll)
Intel has had a mobile quad-core suitable for the iMac or MBP since late last year. Very expensive processors, though.
Expensive and somewhat slower cores if we are talking about the same thing. Apple apparently decided to focus on core clock rate instead of more but slightly slower cores. That is likely good for a good portion of Apples user base but does nothing for people who run apps that could benefit from more cores.
What Apple should gave at this point is alternatives. That is people who know that they would benefit from quad cores could manage an upgrade. Apple has upgrades for everything else in the MBP so why not a update for cores available.
I think quads will wait until the "Clarksfield" processor around the end of this year (probably later rather than sooner). Which will still be expensive, but it'll be Apple's only option if they want to stay competitive.
apple needs to do more than just stay competitive, they need a platform that addresses the needs of power users. Of course even the desktop line has that huge gulf where many customers are left wanting.
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.
Dave
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.Dave
I'm assuming that this is because Snow Leopard will offer significant speed improvements for multi-core machines but only modest ones for dual-core?
I'm assuming that this is because Snow Leopard will offer significant speed improvements for multi-core machines but only modest ones for dual-core?
There are two things that come to mind. One is that some software will work dramatically better with the additional cores. The second thing is that new features mean more background processes to support them.
However it isn't as simple as that as SL has OpenCL which means that some threads may not see the CPUs and instead run on the GPUs. So while SL is going a long way to improve multithreading one can't say for sure where all those threads will be running. In any event there is already software available for the Mac that can leverage additional cores beyond two. That puts the MBP at a disadvantage already, it will just get worst when software becomes SL targetted.
Software for SL can likely be divided into three groups. That would be: 1. Software that will benefit very little if any, 2. Software that will see improved performance immediately and 3. Software that will need to be rewritten.
The second group is most interesting in my mind because there is a bit of guess work here. Some will perform better due to better multi processing / threading support. Others will benefit from improved system libraries or other features the app makes use of. Hopefully we will see some surprises.
At the moment I believe the bigest set of apps are in group #3. That is the apps will need new releases to really benefit from SL. I suspect a good portion of these apps will be rewritten with OpenCL. In mind. This the potential for really dramatic performance increases. Take this to mean that initially SL might not look all that impressive. We will have to wait for software revs before the benefits are truely obvious. Maybe more than one rev will be required.
In otherwords the benefits of quad cores and SL may not be immediate for all users. Apples system software is likely to impress though. Really though I see dual core CPUs as a thing of the past.
Dave
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.
Dave
However it isn't as simple as that as SL has OpenCL which means that some threads may not see the CPUs and instead run on the GPUs.
The MacBook Pro has two GPUs.
So if you're running SL on a current MacBook Pro you'll have both CPU cores and both GPU cores. Now, obviously, this is only potential, but any application that realizes it will be significantly accelerated. GPUs do what they do very, very well.
If the MBP is plugged in, the 9400M is just sitting there waiting for something to do. Snow Leopard can use it as a coprocessor--with the results depending, again, on how many tasks are suited to computation on a GPU, and how well they're suited.
I wouldn't say that it's a complete loss. Frankly, the people buying quad cores now generally fall into two camps: either they need as many cores as they can get their hands on or they have two or three idle cores 90% of the time and bragging rights. The transition to software that is truly designed around multicore processing will be slow and painful, with the biggest bread-and-butter applications coming to the party last.
The MacBook Pro has two GPUs.
Well if you are lucky enough own one of the recent models it does.
So if you're running SL on a current MacBook Pro you'll have both CPU cores and both GPU cores. Now, obviously, this is only potential, but any application that realizes it will be significantly accelerated. GPUs do what they do very, very well.
True but the real important part of that statement above is "GPUs do what they do very, very well". If your multithreaded code simply doesn't run well on a GPU then your speed up might not be that great. Generally though even slow hardware cores are better than time slicing across one or two cores.
If the MBP is plugged in, the 9400M is just sitting there waiting for something to do. Snow Leopard can use it as a coprocessor--with the results depending, again, on how many tasks are suited to computation on a GPU, and how well they're suited.
Its is the question of suitability that has me wondering just how well a GPU can be leveraged. The way I understand OpenCL (granted little formal training here) is that the Mac only hands off routines to the cores on the GPU, you aren't really running a full thread there. I could be wrong here as I've had no development experience what so ever with this feature.
I wouldn't say that it's a complete loss. Frankly, the people buying quad cores now generally fall into two camps: either they need as many cores as they can get their hands on or they have two or three idle cores 90% of the time and bragging rights.
Yes exactly! For the people that fall into that first camp they don't have an Apple alternative to run their apps on. At least not a low cost or portable machine. So it is a complete lost for people that have a demand for that type of hardware.
As to the future I believe that SL will make far better use of those cores even for the person that falls into camp two. It is like the move to dual core processors where many of the same arguments where made, in general though dual core lead to more responsive machines for the general user thus it went from bragging rights to being mainstream simply because it offered a better experience. At times that better experience came at the cost of single threaded performance, it was the overall gain that drove the demand for dual cores though. Especially after OS'es and apps became better adapted to the capability.
So I see SL driving the demand of Quad cores in a similar manner. Yeah it depends just how well certain apps can leverage the GPU, which is already there. Interestingly benchmarking is going to be very interesting in the future as not only will the CPU have an impact but so will the GPU. With the wide variety of GPU's, finding out which platform is best for an app will take a lot of testing.
The transition to software that is truly designed around multicore processing will be slow and painful, with the biggest bread-and-butter applications coming to the party last.
Well no not really as some software already exists and other apps will never convert to heavy threading. GPU leveraging may or may not be easy for a given app. I don't see individual apps as the big driver here anyways, it is the capability to do multiple things at once that will make such machines interesting.
Dave
Well if you are lucky enough own one of the recent models it does.
The MacBook Pro has two GPUs.
was in response to:
Today is a poor time to be buying a Mac Book Pro if you expect to leverage SL on it.
Dave
And hopefully they terminate that aweful G9400M/9600M GT and replace it with
some of these:
GeForce GTX 280M (optional 17" MBP)
GeForce GTX 260M (standard 17" MBP)
GeForce GTS 160M (optional 15" MBP)
GeForce GTS 150M (standard 15" MBP)
I only hope that they will update the MBP line soon after the WWDC2009.
And hopefully they terminate that aweful G9400M/9600M GT and replace it with
some of these:
GeForce GTX 280M (optional 17" MBP)
GeForce GTX 260M (standard 17" MBP)
GeForce GTS 160M (optional 15" MBP)
GeForce GTS 150M (standard 15" MBP)
Do you really want 60W and 75W GPU in MBP's?
They won't do this obviously:
1) Battery life. 15" MBP with current battery won't be able to last even 1 hour under normal load. Less than 1.5 hours when idle and < 30min under full load (3D games, ect)
2) 4500-5000 rpm when idle with current heatsink and super-hot enclosure.
3) In order to provide efficient cooling and acceptable battery life (btw only current 17" unibody is good enough) they'll have to put at least 200Wh battery and much heavier heatsink. Each MPB will become 50% thicker and 1kg heavier.
And GTX260M is only 70-80% faster than GT 130M. Still not enough for modern 3D games. In other words it's not worth it.
But your powerconsumption evaluation is sound. Hadn't thought of it to be honoust. As long as they upgrade the GPU to something better then the 9400M/9600M GT I'm satisfied.
Only 70-80% faster? That's almost twice as fast...
But your powerconsumption evaluation is sound. Hadn't thought of it to be honoust. As long as they upgrade the GPU to something better then the 9400M/9600M GT I'm satisfied.
Yes, almost twice as fast than mid-range 130M GT (130M GT is basically overclocked 9600M GT, approx. 20% faster).
But both desktop GTX285 and GTX295 are respectively 2.5x and 3.6x faster than mobile 280M (based on 3DMark Vantage tests)! That's why current mobile GPUs are obsolete for hardcore gaming (and thas's what 280M is intended for)
On the other hand, 9600M GT (and 130M GT) are most power efficient solutions for mid-sized notebooks.
? Increase hard drive capacities across the board to 500GB & bump processors
? Make the aluminum 13" unibody MacBook the MacBook Pro
? Add 3G WWAN option
? Add a Firewire 800 port to the newly rebranded 13" MacBook Pro
? Add NVIDIA 9600m discreet graphics alongside the 13" MacBook Pro's NVIDIA 9400m
? Add the matte display option
? Increase the resolution on the 15" MacBook Pro to 1680 x 1050
? Transition the line-up to 16:9 displays
So, what's the consensus-- update timeframe another 3 months?
I doubt there is a firm consensus. Last summer a bunch of us predicted that Apple would release the mobile Penryn unibody in mid-July. They didn't get around to the 17" machine until February of 2009. Needless to say, I was wrong, as were quite a few others.
Who wouldn't love to see a refresh this month? Still, it looks like Apple have a lot on the proverbial plate right now. Intel just released updated mobile cpus, but we have seen no news of limited supply when it comes to the current Macbook Pros. That's usually the most telling sign of a near-term upgrade.
On the other hand, the recent change in the 13" Macbook is salient evidence for a sooner-rather-than-later refresh — and unlike last year, the unibody is already a reality, so sometime in the next three months sounds about right. I'd love for the change to come this week, though.
Cross fingers.
?
? Intel just released updated mobile cpus, but we have seen no news of limited supply when it comes to the current Macbook Pros. That's usually the most telling sign of a near-term upgrade.
?
Cross fingers.
Ah, yes, stock lead time has been VERY telling. I'm wondering if I should just wait, or buy now. Yeah, I know; I said it.
I'm wondering if I should just wait, or buy now. Yeah, I know; I said it.
I feel your pain. I'm in the same boat.