[eternalfrost]

so on intels pamphlet for the GMA950 it says "192MB maximum video memory"

http://www.intel.com/products/chipse...950/gma950.pdf

after just browsing around there have been alot of people talking about being able to increase the ammount of RAM available to the GMA past the 64mb standard in the Macbook, but havent explained exactly how to do it.

some people say it cant be done in OS X but can easily be changed in the control panel in XP.

basically all i could find was wish-washy info and was wondering if anyone had anything solid. it makes sence that osx would limit it for users with 512mb ram, but seems silly not to let you increase it when upgraded if it is indeed possible.

one of my biggest concerns about buying a macbook is its low 'gpu' but if it could be configured to use up to 192MB, even if only in XP where most games would be run anyways, i would feel MUCH better about my decesion

[mduell]

What apps are you running that are so dependent on VRAM?

[eternalfrost]

well im just very curious. looking around the web, all non-macbook systems with the GMA950 are listed as 128MB shared video memory. is this the same chip or am i totally off base here?

it just seems rediculus that it would be tied down to half its potential speed. especially when so many people are 'complaining' about the low graphics in the macbook.

i dont know about you, but if im going to be paying for this chip, i want to get everything out of it that it can offer.

[sieb]

Once you add more ram, if OSX needs more VRAM past 64MB, it will increase its usage to whatever it needs up to its limit.

[ghporter]

On a Windows system, there will be a BIOS setting for how much RAM to reserve for video. I don't know how they do it with the MacBook, but it's probably something burried in System Preferences somewhere (just guessing).

[jamil5454]

64MB is how much is allocated to use, but when you need to use more, it just uses part of your system RAM, AFAIK. AGP aperture worked the same way.

And the VRAM is definitely NOT the bottleneck for the horrible performance of the GMA950. The video chipset itself is like a mini-computer. The GPU is the CPU and the VRAM is the RAM. So... just as increasing RAM on your computer doesn't improve processing speeds, increasing VRAM on the video card doesn't really improve game speeds.

64MB should be enough for just about anything you need to do with a MacBook.

[eternalfrost]

Originally Posted by sieb

Once you add more ram, if OSX needs more VRAM past 64MB, it will increase its usage to whatever it needs up to its limit.

alright cool, so this just happens automaticaly without changing any settings? i still cant find a proper explination about the whole thing which i find sort of strange and frustrating.


can anyone recomend a program or soemthing that can monitor the ram vram cpu gpu etc?

[mduell]

Originally Posted by eternalfrost

well im just very curious. looking around the web, all non-macbook systems with the GMA950 are listed as 128MB shared video memory. is this the same chip or am i totally off base here?

it just seems rediculus that it would be tied down to half its potential speed. especially when so many people are 'complaining' about the low graphics in the macbook.

i dont know about you, but if im going to be paying for this chip, i want to get everything out of it that it can offer.

Yes, it's the same chip.

Having half the VRAM won't give you half the performance. It doesn't work like that.

[sieb]

alright cool, so this just happens automaticaly without changing any settings? i still cant find a proper explination about the whole thing which i find sort of strange and frustrating.

I think the max is 192MB OSX will use if you have over a gig of ram installed. I have two gigs and never have a problem. And yes, its all automatic in OSX. In XP, the more ram increases what it will use for VRam, mine says it has 128MB to use.

[buddy1065]

Still, what if the next version Macbook specs out at least 128 MB like it should have all along? I don't think it will be a great difference; shared memory is still shared memory.

The LG TX has a Geforce 7300 and its only a 12" screen. The laptop is .9" thick. The MBP has ATI 1600. There is not that great a difference to me between MB and MBP and I really believe the MB can hold some sort of dedicated GPU like the MPB in its present state. But right now the Apple bean counters won't let that happen because it's all in the get more money plan; you know; the one that has them making sure the CPU is soldered to the motherboard to prevent upgrades, forcing users to buy another laptop. When the hordes of "Get it now 'cause there will always be a new revision" folks are thru buying the initial buggy versions then perhaps Apple will make some changes. Right now there is no need to, as long as sales are going thru the roof. Apple knows there are laptop junkies out there who are willing to sell there Macbooks and re-invest in Macbook Merom Heaven with a possible dedicated GPU thrown in, on the never ending cycle of selling their laptops ONLY AFTER A YEAR or two even though they are in perfect condition, and handing over more money to Apple for the bleeding edge product. Fess up Macbookers, a lot of you will be drooling over Merom and dedicated GPU's soon enough; IMO they are definitely comming within a year and you KNOW you will want one...

If you have no back up laptop then I can see getting a Macbook. It is a nice piece. I'm sure many out there are happy with their decision to purchase one, and rightfully so. Just hope many of you will feel the same way about your Macbooks when the new revisions are out.

I just don't believe in upgrading till I see something that will knock my socks off compared to what I already have. Sadly the MB has everything going for it except the GPU. Till then I'll be happy with my trusty G4 Powerbook. I'm sure there are still a lot of folks out there just like me though. Anyway, there's always next month....and the month after that...and the month after that...where's the fire? I don't smell anything...

[mduell]

Originally Posted by buddy1065

Still, what if the next version Macbook specs out at least 128 MB like it should have all along?

What's that supposed to mean? If you have enough memory, the GPU gets more.

Originally Posted by buddy1065

The LG TX has a Geforce 7300 and its only a 12" screen. The laptop is .9" thick. The MBP has ATI 1600. There is not that great a difference to me between MB and MBP and I really believe the MB can hold some sort of dedicated GPU like the MPB in its present state.

The LG TX has no optical drive, so it can be thinner and there is more room inside for a discrete GPU. It also starts at what, $1700?
If Apple really wanted to put a dedicated GPU in the MacBook, I'm sure they could in terms of space. Maybe it would be a little bit bigger in one or two dimensions, but not much.
Can you find an $1100 notebook that has an optical drive and discrete graphics with dedicated VRAM?

Originally Posted by buddy1065

But right now the Apple bean counters won't let that happen because it's all in the get more money plan; you know; the one that has them making sure the CPU is soldered to the motherboard to prevent upgrades, forcing users to buy another laptop.

Apple uses soldered CPUs in their notebooks, just like every other sub-1.1" notebook, in order to save on thickness. The solder packaging is thinner than the socket packaging.

[buddy1065]

"Can you find an $1100 notebook that has an optical drive and discrete graphics with dedicated VRAM?"

Not sure of your point since the Macbook has no dedicated GPU either.
Nevertheless, Sonystyle's SZ240 starts at $1299, close enough to the mark IMO. Doubtless you could probably get it cheaper than Sonystyle elsewhere.



"What's that supposed to mean? If you have enough memory, the GPU gets more."

Where is the link to this info? Thru this entire thread all I have seen is claims but no links to actual data; that's all the origional poster wants, yet he is still without any documentation. I would really appreciate some concrete link; not calling you out or anything; just like the origional poster I want to see it for myself. If this is true then why doesn't Apple advertise 128 or higher? Why do other manufacturers with the GMA950 advertise 128? Are you saying the 64 MB version is the exact same as the 128 MB version? So far I find this hard to believe. Show us.


"Apple uses soldered CPUs in their notebooks, just like every other sub-1.1" notebook, in order to save on thickness. The solder packaging is thinner than the socket packaging."

Good call.
Your third point may have validity, however I still think the Macbook will have a dedicated GPU within the next two revisions in the same dimensions, especially since component size and cost is being reduced all the time. The next revisions will probably be near the same cost as the origional within a few hundred, as usually happens.

[amazing]

Maximum RAM use for GMA 950 is 80 MB. It's in the #1 note at the bottom of the specs page:

http://www.apple.com/macbook/specs.html

see also the following tech note. Note that the RAM is also for "boot processes at startup" whatever that means. For example, does that mean 64 MB for graphics and 16 MB for "boot processes at startup" which may not be returned to the stack??

http://docs.info.apple.com/article.html?artnum=303407

[buddy1065]

Check out this link also:

http://www.acer.co.uk/acereuro/page4...crc=1504108834

"ATI Radeon® Xpress 1150 integrated 3D graphics, with up to 512 MB of HyperMemory™ (256 MB of dedicated system memory; up to 256 MB of shared system memory)"

Wow, I have no clue whether this will be better or worse than the GMA950 configuration even though both CPU's here (Macbook and 12" wide screen Ferrari 1000) are 2 Ghz but the terms are interesting. Hypermemory is a combined group of system and shared memory.

Apparently there is dedicated system memory and shared system memory. Perhaps the GMA950 can be configured with 64 or 128 dedicated system memory, which will then borrow from shared memory when needed, but I suspect that sharing is a slower process than the memory already dedicated...or perhaps the proper term is "allocated memory" in Intel terms...

Maybe dedicated system memory and allocated memory mean the same thing. If so the available shared memory, whether ATI or Intel, may be used additionally when needed, boosting overall memory as according to demand and what is available in your system. Could be wrong though.

[soonlar]

I always thought that the main thing the vram is used for is supporting larger video resolutions, like n the case of the MB 1280 x 800 or up to 1920 x 1200 pixels on an external display, both at millions of colors.

[sieb]

This has all been argued to death. It's cheaper to use Intels already speced out cpu/gpu chip combo to provide a higher profit margin. It also provides seperation between MBs and MBPs value. Most consumers don't care about a dedicated GPU so it makes sense to not have it as that is the market the MB is directed towards. MBP is directed towards power users who can take advantage of it and are willing to pay the premium. It's not rocket science.

[buddy1065]

What market the Macbook is directed towards is not the issue of this thread. The initial poster has heard there is a way to increase the allocated shared memory of the GMA 950. He simply wanted to know if it was true, but it appears to be hearsay. Info and opinions were presented on the nature of shared memory and what it does; some say it is not necessary to increase 64 to 128 MB. My opinion is that there is a difference between allocated system memory and the additional shared memory which is borrowed when allocated shared system memory runs out. I believe the additional is slower than what is allocated, but acknowledge I could be wrong, but my belief would explain why some manufacturers claim 64, some 128, some 256 MB, etc. It remains to be seen so far if these numbers are just manufacturers hype. But like I said; whichever way is described, shared memory is just shared memory.

The initial poster also has concerns about light gaming, which some say can be done. I know my Sony T250 1.2 Ghz Centrino and GME855 64 MB shared can handle games like Midtown Madness and even more recent games on low 640x480 like Tomb Raider Legends which has slow frame rates though it is quite playable. How the Macbook would do in comparison is a bit of a mystery. All the forums I've asked Macbook owners to download the free TRL demo (www.tombraider.com) and try it using Boot camp have not answered; the thread dies! So being one who really believes in time the Macbook will improve in this regard eventually, I will wait patiently unless there is to be an ultralight PC offering which will fare considerably better than my T250, preferably in 12" wide screen glossy or less. Time will tell.

[buddy1065]

Nevertheless, as far as marketing is concerned, didn't even the latest revision iBook have a dedicated graphics chip instead of shared memory? Wasn't it in the $1100 range? Isn't the iBook even smaller than the Macbook? How soon we forget! All the more reason to believe it is possible to have a dedicated GPU, possibly as high as 64 MB (so as to keep it out of competition with the MBPs) in the larger Macbook, perhaps sooner than some may think. The iBook was doing it before the Macbook came along. Marketing is one thing, but remaining a step backward with the idea the Macbook is some kind of ultralight that cannot be allowed a real GPU is not rocket science either. The Apple Macbook is in competition with real midsized 13" Intel laptops now, not my little 10.6" shared memory Vaio. Dell has the 12" M1210 and Sony doesn't seem to have a problem selling tons of their 13" SZ notebooks amid their larger ones, to name a few instances. But I believe Apple will step up to the plate soon enough.

[Hi I'm Ben]

Originally Posted by buddy1065

Nevertheless, as far as marketing is concerned, didn't even the latest revision iBook have a dedicated graphics chip instead of shared memory? Wasn't it in the $1100 range? Isn't the iBook even smaller than the Macbook? How soon we forget! All the more reason to believe it is possible to have a dedicated GPU, possibly as high as 64 MB (so as to keep it out of competition with the MBPs) in the larger Macbook, perhaps sooner than some may think. The iBook was doing it before the Macbook came along.

smaller screen yes, thinner, no. The GMA950 also allowed for features the Book's crappy dedicated 32mb video card didn't like CoreImage

[buddy1065]

Seriously doubt if Apple will use yesterdays ATI 32 MB video card, and yes; graphics chips do fit in the thinner MBP's, so it is possible to fit one in a Macbook IMO. But thanks for the info. Didn't realize the iBooks were so thick.

[mduell]

Originally Posted by buddy1065

"Can you find an $1100 notebook that has an optical drive and discrete graphics with dedicated VRAM?"

Not sure of your point since the Macbook has no dedicated GPU either.
Nevertheless, Sonystyle's SZ240 starts at $1299, close enough to the mark IMO. Doubtless you could probably get it cheaper than Sonystyle elsewhere.

That was my point. An $1100 13" notebook with optical drive and discrete graphics with dedicated VRAM is too good to be true. Sure, for an extra $200, Sony managed to squeeze in discrete graphics with 32MB dedicated VRAM. Meanwhile they went with a slower CPU, less cache, slower FSB, and oh yea, missing an entire CPU core. Also note no camera, slower memory, no Bluetooth, smaller hard disk, no DVI, no GigE, and no powered Firewire. Add all of those back (where possible), and you're at a base price of $1500, now $400 off the original target.

Originally Posted by buddy1065

"What's that supposed to mean? If you have enough memory, the GPU gets more."

Where is the link to this info? Thru this entire thread all I have seen is claims but no links to actual data; that's all the origional poster wants, yet he is still without any documentation. I would really appreciate some concrete link; not calling you out or anything; just like the origional poster I want to see it for myself. If this is true then why doesn't Apple advertise 128 or higher? Why do other manufacturers with the GMA950 advertise 128? Are you saying the 64 MB version is the exact same as the 128 MB version? So far I find this hard to believe. Show us.

There is only one version of GMA950. Memory usage is just a configuration setting. Intel.com has tons of documentation, some of it incredibly technical.

Originally Posted by buddy1065

Apparently there is dedicated system memory and shared system memory. Perhaps the GMA950 can be configured with 64 or 128 dedicated system memory, which will then borrow from shared memory when needed, but I suspect that sharing is a slower process than the memory already dedicated...or perhaps the proper term is "allocated memory" in Intel terms...

No. There is dedicated video memory and there is shared video/system memory. GMA950 does not support any dedicated video memory. All of the memory for GMA950 comes from the system's memory pool

Originally Posted by buddy1065

Nevertheless, as far as marketing is concerned, didn't even the latest revision iBook have a dedicated graphics chip instead of shared memory? Wasn't it in the $1100 range? Isn't the iBook even smaller than the Macbook?

It sure did. And the discrete graphics chip in the iBook was so old it didn't even support Core Image; GMA950 fully supports Core Image and Core Video.
The iBooks were 25% thicker than the MacBooks are.

[buddy1065]

"That was my point. An $1100 13" notebook with optical drive and discrete graphics with dedicated VRAM is too good to be true. Sure, for an extra $200, Sony managed to squeeze in discrete graphics with 32MB dedicated VRAM. Meanwhile they went with a slower CPU, less cache, slower FSB, and oh yea, missing an entire CPU core. Also note no camera, slower memory, no Bluetooth, smaller hard disk, no DVI, no GigE, and no powered Firewire. Add all of those back (where possible), and you're at a base price of $1500, now $400 off the original target."

Mduell, I would be willing to pay more than $1100 for a Macbook with a dedicated GPU, DVD burner, etc; wouldn't most of you? Again I say, what's your point? To use a super low impractical price for what? Doesn't make sense to me. There will be many willing to pay a higher price for such a well equipped Macbook when it is available, myself included. Personally I would be happy to pay over $1500 and would consider it money well spent.

My point is (again) with falling prices and decreasing component size I believe we will see a dedicated GPU in the Macbook within the next 2 revisions.

"There is only one version of GMA950. Memory usage is just a configuration setting. Intel.com has tons of documentation, some of it incredibly technical."

Hey, not disputing there is only one version, but if it's a setting is there documentation that the setting cannot be increased as other manufacturers are apparently claiming as some say 64 while others say 128? Oh well, no problem. Maybe it's just too technical as you say and we will never really know the true answer.

"No. There is dedicated video memory and there is shared video/system memory. GMA950 does not support any dedicated video memory. All of the memory for GMA950 comes from the system's memory pool"

Actually I said "dedicated system memory" not dedicated video memory, which I believe is still shared memory. Guess I didn't make it clear enough. Of course there is no dedicated video memory. Otherwise the Macbook would be a better gaming machine.

"It sure did. And the discrete graphics chip in the iBook was so old it didn't even support Core Image; GMA950 fully supports Core Image and Core Video.
The iBooks were 25% thicker than the MacBooks are."

Like I said to a previous poster, I seriously doubt Apple will use an old ATI 32 MB GPU in the Macbook. Again, as far as thickness is concerned the MBP has ATI 1600. If the Macbook is even thicker then I suppose it is even more possible to fit one. Sorry, still don't see the impossibility that you are implying of putting a dedicated GPU in a Macbook. In fact I believe they can do it even in the next revision if it weren't for the bean counters, which you evidently forgot I mentioned well before you implied the economics of it not being feasible marketing strategy.

"But right now the Apple bean counters won't let that happen because it's all in the get more money plan;" - my post.

So in conclusion some are telling me it is not good marketing strategy when I was the first to point that out. They evidently have not read my post entirely and are counterpointing instead my prediction of things to come. Apparently you are doing the same thing?

But forget my last quote about the bean counters, now I say I believe Apple even now has the technology to put a GPU in the Macbook although the marketers were are against it for the first revision.

So are we saying it is impossible fit a GPU in a Macbook, citing the iBook thickness, when a dedicated GPU is already in the thinner MBP? Tell you what; you all can say "I told you so" if it doesn't happen in the next two revisions if it makes you feel better. I am sticking with my prediction.

As to the origional poster I hope in all of this rabble (for which I apologize) you have found an answer. Maybe it's a bit too technical as Mduell stated. I am still a bit unclear, but then I am not buying a Macbook with shared memory; I am waiting, so it matters little to me.

[jamil5454]

By the time a MacBook with a decent dedicated graphics solution comes out, I (and most of you, I'm sure) could have more than made up for the difference in cost between a MacBook and MacBook Pro.

[JulesLt]

Buddy - as per comments above, the main reason Apple will have switched from the separate GPU combo on the iBook to the dual Intel set will have been to get a sweet deal on the CPU and other chipset components - it is well known that Intel give substantial discounts to vendors using multiple Intel components. Also, the G4 is near the end of it's lifecycle so will now be pretty low-cost.

This will also account for some of the price differential with the MBP. It is not just that the separate GPU costs more, but that it loses the CPU subsidy.

I also think there is the heat issue. The original roadmap seemed to be for Apple to release machines using what is now being branded as Core Duo 2, which offers the CPU power/watt advantages Jobs mentioned. Instead they went over on the existing (P4) generation technology, which is more power hungry than the G4 chips it replaced. Given how hot the current MB / MBP are running, I don't know whether they would be up to a high-performance GPU.

I think you are correct, however, that we will see a non-shared memory GPU in a MacBook within the next couple of generations. The current setup can't even support OmniDazzle, and it's going to be pretty useless if their baseline machines can't even run the fancy effects in Leopard.

And it's evident that there IS demand for a small decent spec laptop - whether it's a 13" MBP or a maxed-out MB - just as there is demand for a desktop between the Mini and MacPro (in both cases - something with about the spec of the iMac would be nice).

[ghporter]

I wouldn't call the Core Duo "existing P4 technology." It's based on the P4, but it's a couple generations past it. And the Core Duo is significantly less power hungry than the P4m mobile processors Intel boasted about only a few months ago.

Further, I disagree with the assertion that the MacBook's video solution is "not a separate GPU." It's integrated into the logicboard and uses main memory, but it's still a separate processor. There ARE truly non-GPU video solutions out there, using CPU power to generate video; they're pretty scary. While the video solution in the MB is not a separate card, it's most certainly a "separate GPU" in the sense that it consists of discrete components that handle video. This may be a minor issue, but it's still my feeling on the matter.

[amazing]

I still can't figure out how Apple ever thought that the target market for the MB wasn't interested in playing games? That's just a big puzzler for me...

[chefpastry]

Nowadays, gaming on a computer means investing in some serious hardware. Specifically, really fast (read expensive) video cards. Gamers have to look high end for serious gaming. The MacBook is the low end of the line.

The other option would be to get a PlayStation 2 or Xbox 360.

[buddy1065]

No one is expecting to jam in an ATI 7800 GPU into a Macbook and demand serious smokin' gaming, unless you want the Macbook to seriously start smoking.... All I am proposing is a dedicated GPU a bit higher than the GMA950 that will handle more of Leopards requirements, as well as do a bit better at occasional gaming, that's all. I feel a lot of folks just wanted a little more and I feel it can be done.

[mduell]

Originally Posted by buddy1065

Mduell, I would be willing to pay more than $1100 for a Macbook with a dedicated GPU, DVD burner, etc; wouldn't most of you? Again I say, what's your point? To use a super low impractical price for what? Doesn't make sense to me. There will be many willing to pay a higher price for such a well equipped Macbook when it is available, myself included. Personally I would be happy to pay over $1500 and would consider it money well spent.

My point is (again) with falling prices and decreasing component size I believe we will see a dedicated GPU in the Macbook within the next 2 revisions.

Sure; I think we'll see a 13.3" MBP within a year, with discrete graphics and an aluminum case and a $1599ish price tag.

Originally Posted by buddy1065

Hey, not disputing there is only one version, but if it's a setting is there documentation that the setting cannot be increased as other manufacturers are apparently claiming as some say 64 while others say 128? Oh well, no problem. Maybe it's just too technical as you say and we will never really know the true answer.

I think Apple advertises and defaults to 64MB because they don't want to use a quarter of their low-end machine's stock memory just for graphics. 10.4 barely runs on 512MB, much less 384MB.

Originally Posted by buddy1065

"No. There is dedicated video memory and there is shared video/system memory. GMA950 does not support any dedicated video memory. All of the memory for GMA950 comes from the system's memory pool"

Actually I said "dedicated system memory" not dedicated video memory, which I believe is still shared memory. Guess I didn't make it clear enough. Of course there is no dedicated video memory. Otherwise the Macbook would be a better gaming machine.

I still don't understand quite what you mean. There is one pool of memory; the GPU gets some and the rest of the system gets the remainder to work with.
Dedicated VRAM wouldn't help much. The system memory is already very fast (10GBps+), and the GPU is really what is holding up performance.

Originally Posted by buddy1065

"It sure did. And the discrete graphics chip in the iBook was so old it didn't even support Core Image; GMA950 fully supports Core Image and Core Video.
The iBooks were 25% thicker than the MacBooks are."

Like I said to a previous poster, I seriously doubt Apple will use an old ATI 32 MB GPU in the Macbook. Again, as far as thickness is concerned the MBP has ATI 1600. If the Macbook is even thicker then I suppose it is even more possible to fit one. Sorry, still don't see the impossibility that you are implying of putting a dedicated GPU in a Macbook. In fact I believe they can do it even in the next revision if it weren't for the bean counters, which you evidently forgot I mentioned well before you implied the economics of it not being feasible marketing strategy.

<snip>

So are we saying it is impossible fit a GPU in a Macbook, citing the iBook thickness, when a dedicated GPU is already in the thinner MBP? Tell you what; you all can say "I told you so" if it doesn't happen in the next two revisions if it makes you feel better. I am sticking with my prediction.

The MBP has 20-40% more surface area to work with; since the optical and hard drive are a fixed size, that makes for about 35-70% more area for the logic board.

Originally Posted by JulesLt

Buddy - as per comments above, the main reason Apple will have switched from the separate GPU combo on the iBook to the dual Intel set will have been to get a sweet deal on the CPU and other chipset components - it is well known that Intel give substantial discounts to vendors using multiple Intel components. Also, the G4 is near the end of it's lifecycle so will now be pretty low-cost.

This will also account for some of the price differential with the MBP. It is not just that the separate GPU costs more, but that it loses the CPU subsidy.

Apple isn't using Intel's wifi chipsets, so they're not getting the "whole package" discount. But I just want to add that GMA950 costs $4 at list prices; since Apple is buying more than 1000 units, they're probably paying slightly less.

Originally Posted by JulesLt

I also think there is the heat issue. The original roadmap seemed to be for Apple to release machines using what is now being branded as Core Duo 2, which offers the CPU power/watt advantages Jobs mentioned. Instead they went over on the existing (P4) generation technology, which is more power hungry than the G4 chips it replaced. Given how hot the current MB / MBP are running, I don't know whether they would be up to a high-performance GPU.

Originally Posted by ghporter

I wouldn't call the Core Duo "existing P4 technology." It's based on the P4, but it's a couple generations past it.

Sorry guys, but you're both slightly incorrect.
Core Duo is not Pentium 4 (NetBurst) based, but rather it is based on the Pentium M which is based on the Pentium 3. Core 2 Duo is a significant upgrade to this, but still bears some resemblance to the Pentium 3. Core 2 Duo improves performance per watt by about 20% over Core Duo, but Core 2 Duo improves performance per watt by something like 150-200% over the Pentium 4.
It's hard to find good numbers for G4 power consumption, but the CD/C2D chips aren't that far off.

[buddy1065]

Speaking of Core duo 2, looks like the Macbook will be getting it right along with the MBP as early as next month.

http://www.reghardware.co.uk/2006/08..._to_get_merom/

$1599 sounds great, but I am still expecting to see the black Macbook casing.

[HouseSold]

Originally Posted by buddy1065

Speaking of Core duo 2, looks like the Macbook will be getting it right along with the MBP as early as next month.

http://www.reghardware.co.uk/2006/08..._to_get_merom/

$1599 sounds great, but I am still expecting to see the black Macbook casing.

Posted that article here: http://forums.macnn.com/69/mac-notebooks/305267/good-news-finally/

http://www.digitimes.com/systems/a20060811PB201.html

Still lacking details we're all dying to know, but does specify MacBook and MacBook Pro.

When is the Paris show in September or will it be a stealth Tuesday unroll here states side?; now that we can't take our Glacier water on the planes to Paris.

Luckily, Steve has his own jet.

BTW Nothing wrong with a BlackBook casing; my request is for 14" Black, backlit keys, Merom and 965 chipset or surprise us, addressable 4gb RAM (not limited to 2gb RAM), cooler temps with new heat sink, quieter/better fan control and Leopard lasting 5+ hours on battery. Almost forgot, same non-problematic 85W powerbrick from MBP.

I know it's too soon for Leopard, but ONK (one never knows)

New screen sizes SB 12" Ultralite, 14" and 17" to please most users globally (if we could only pick 3 sizes).

My 2 cents.........

[ghporter]

mduell, I'm surprised that the Core Duo is based on the P3 after all the work they did on P4 technology to get it to the point they have it with the P4m. Any clue what's special about the P3 that made them choose this? I'd still say that the Core-line is "a couple of generations past" the P4 in terms of performance, including "per watt" performance.

[mduell]

Originally Posted by ghporter

mduell, I'm surprised that the Core Duo is based on the P3 after all the work they did on P4 technology to get it to the point they have it with the P4m. Any clue what's special about the P3 that made them choose this? I'd still say that the Core-line is "a couple of generations past" the P4 in terms of performance, including "per watt" performance.

The Pentium 4 microarchitecture (NetBurst) was all about long pipelines, high-bandwidth memory, and last but not least, clocking the piss out of the sucker. When the latency of RDRAM killed the bandwith (it was providing 6+GBps in like 2000), performance took a bit of a hit. When the power consumption drops from process shrinks didn't work out quite as (everyone, not just Intel) planned, they were stuck with this volcano of a CPU. People have demonstrated P4s running stably at 5+Ghz; you just have to feed it a constant supply of liquid nitrogen to wash away the 200+W coming off of it.

The P4m wasn't to great. My parents owned an 8lb laptop based on it, and that sucker got hot. It's saving grace was dual batteries to get decent (5h) battery life. With the P4m being suitable only for DTR laptops, Intel needed something new. The P3m was doing alright for the rest of the laptop market at the time, but TMTA was coming out with their low-power chips (and in the end, low-performance, but no one knew that at the time) and there was a lot of uncertainty. The Intel group in Israel that had been working on Timna (think Celeron-with-onboard-RDRAM-controller) was now free thanks to the demise of RDRAM, so they started work on Banias. Intel calls Banias a new chip from the ground up, but it's more like a Pentium 3 with a few Pentium 4 techonologies backported (like QDR FSB and SSE2). Banias was a smash hit, and the LV/ULV Banias chips (along with the miserable performance of Crusoe), put Intel firmly at the head of the laptop market. Dothan was a 2004 minor upgrade to Banias, with twice as much cache thanks largely to a process shink (along with Intel's ability to pull of great yields), as well as a few other tweaks. With multiple cores becoming the new rage, and the "slap two dies in a package and send it off" mantra of the Pentium D not working out so well, Intel went ahead with Yonah. Basically put two Dothan cores side by side, share the cache between them, backport a few Pentium 4 pieces (like SSE3), shink the process size (without rearrangement), bump up the FSB/RAM, and you've got Core Duo.

The whole Core microarchitecture family (Merom, Conroe, and Woodcrest) is a further evolution of Yonah. You can read more than you wanted to know about the changes involved here because I'm tired of typing.

[ghporter]

Wow. That's something else-and thanks for typing all that you did. I though I was somewhat informed about processor development when the Core processors came out and cured me of that illusion. It seems that I was even more clueless than I thought. Thanks again!

[HouseSold]

Originally Posted by mduell

The Pentium 4 microarchitecture (NetBurst) was all about long pipelines, high-bandwidth memory, and last but not least, clocking the piss out of the sucker. When the latency of RDRAM killed the bandwith (it was providing 6+GBps in like 2000), performance took a bit of a hit. When the power consumption drops from process shrinks didn't work out quite as (everyone, not just Intel) planned, they were stuck with this volcano of a CPU. People have demonstrated P4s running stably at 5+Ghz; you just have to feed it a constant supply of liquid nitrogen to wash away the 200+W coming off of it.

The P4m wasn't to great. My parents owned an 8lb laptop based on it, and that sucker got hot. It's saving grace was dual batteries to get decent (5h) battery life. With the P4m being suitable only for DTR laptops, Intel needed something new. The P3m was doing alright for the rest of the laptop market at the time, but TMTA was coming out with their low-power chips (and in the end, low-performance, but no one knew that at the time) and there was a lot of uncertainty. The Intel group in Israel that had been working on Timna (think Celeron-with-onboard-RDRAM-controller) was now free thanks to the demise of RDRAM, so they started work on Banias. Intel calls Banias a new chip from the ground up, but it's more like a Pentium 3 with a few Pentium 4 techonologies backported (like QDR FSB and SSE2). Banias was a smash hit, and the LV/ULV Banias chips (along with the miserable performance of Crusoe), put Intel firmly at the head of the laptop market. Dothan was a 2004 minor upgrade to Banias, with twice as much cache thanks largely to a process shink (along with Intel's ability to pull of great yields), as well as a few other tweaks. With multiple cores becoming the new rage, and the "slap two dies in a package and send it off" mantra of the Pentium D not working out so well, Intel went ahead with Yonah. Basically put two Dothan cores side by side, share the cache between them, backport a few Pentium 4 pieces (like SSE3), shink the process size (without rearrangement), bump up the FSB/RAM, and you've got Core Duo.

The whole Core microarchitecture family (Merom, Conroe, and Woodcrest) is a further evolution of Yonah. You can read more than you wanted to know about the changes involved here because I'm tired of typing.

Great Job !!!

[StickClicks]

I don't see why the allocated couldnt be raised but I don't know why you would want to. I had thought the same thing when I went in to look at the new Macbooks and when I asked the genius about it he said that 64MB where allocated at start up, and when the system requiered more (and had RAM to spare) it would take up to 128MB as need be, and then drop back down to holding 64MB. He also noted that 64MB are reserved even if the GPU doesnt need the whole 64MB. I guess maxing out the RAM on these babies isn't really a luxury


Im not that disapointed with apple. I have read somewhere that the Intel graphics out preform the ATI in the macbook pros in video playback. I believe it considering that each of those chips are really specialized when it comes to their design. ATI for 3D rendering and the GMA950 for video playback (2D processes). Muth like the mac mini the macbook is just a base level (with many perks i will admit) computer. If that GMA950 keeps the cost down, I am fine with that. Can't wait to pick mine up when the budget permits. Pens and papers are growing annoying, webmail has limitations and I cant cart me desktop around.

[mduell]

Originally Posted by StickClicks

Im not that disapointed with apple. I have read somewhere that the Intel graphics out preform the ATI in the macbook pros in video playback. I believe it considering that each of those chips are really specialized when it comes to their design. ATI for 3D rendering and the GMA950 for video playback (2D processes). Muth like the mac mini the macbook is just a base level (with many perks i will admit) computer. If that GMA950 keeps the cost down, I am fine with that. Can't wait to pick mine up when the budget permits. Pens and papers are growing annoying, webmail has limitations and I cant cart me desktop around.

GMA950 can accelerate video playback for some types of video if the driver supports it. I have yet to see anything definitive on Apple's drivers for Intel's GPU.

[blakey]

Originally Posted by amazing

Maximum RAM use for GMA 950 is 80 MB. It's in the #1 note at the bottom of the specs page:

http://www.apple.com/macbook/specs.html

see also the following tech note. Note that the RAM is also for "boot processes at startup" whatever that means. For example, does that mean 64 MB for graphics and 16 MB for "boot processes at startup" which may not be returned to the stack??

http://docs.info.apple.com/article.html?artnum=303407

I don't especially want to get tangled in the argument here, but followed both these links which actually say that this is the MINIMUM RAM not the maximum, in fact it specifically states that MORE RAM will be allocated if the GPU asks for it.

Is that so difficult to understand?