Intel GMA HD - thoughts on use in MBA?

[Anonymous Freak]

Tell me then you have not seen the keynote before replying to me on this. I saw fast responsive hd graphics on a 1.5 lb device with every possible feature buit into one chip that pretty much blew my mind. Now take that new integrated speed/IO/GPU/CPU and beef it up for a large portable device then i imagine things are getting pretty amazing. Apple has stated in the video now that they are chip makers. I am thinking we will see more chips moving into more products if they are making that claim. I for one did not see stutter in any video they played. I for one wont rule it out either.

These devices contain dedicated H.264 decoder chips. The integrated CPU in that can almost certainly *NOT* handle 720p H.264 without the help of the decoder chip. The 3D graphics is on-par with the Playstation Portable, not even with Intel GMA 4500HD graphics. Of course, it is possible that the 3D chip is more powerful than it appeared during the presentation, just that there haven't been apps to take full advantage of it yet.

Of course, if you want high-quality gaming graphics, you would need a discrete GPU that by itself would draw more power than the entire iPad.

 

[Bosman]

These devices contain dedicated H.264 decoder chips. The integrated CPU in that can almost certainly *NOT* handle 720p H.264 without the help of the decoder chip. The 3D graphics is on-par with the Playstation Portable, not even with Intel GMA 4500HD graphics. Of course, it is possible that the 3D chip is more powerful than it appeared during the presentation, just that there haven't been apps to take full advantage of it yet.

Of course, if you want high-quality gaming graphics, you would need a discrete GPU that by itself would draw more power than the entire iPad.

"It is possible it is more powerful than appeared"? You got to be kidding me! Schooling me or anyone about dedicated graphics is not necessary. We all know that currently, and i mean currently, this is the paradigm that is the way to go for powerful graphics. I am talking about a macbook air and who if not Apple would like to say screw you all we designed something just for our computers and electronic devices that doesn't require us to compare our systems with Sonys and Dells. Further more we can make them faster and spec them to do specifically what we want each device to procure because we are not limited by our imagination for what is possible. We will just hire the top GPU and CPU engineers and dream!

 

[robeddie]

Can't everyone just find some common ground?

I don't know. When I read ehurtley's (and other's) first couple of posts (before things really went south), the subtext did seem to be that they would think it was perfectly acceptable to have the new, but still inferior Intel GMA in the upcoming MacBook Air.

Like Scottsdale (obviously LOL), that sentiment makes my hair stand on end. For the past 15 months, we've enjoyed and appreciated the outstanding performance of the 9400. We were also really looking forward to what Nvidia would come up with in their next generation integrated chip - and how that would make the next gen MBA even better.

But of course, Intel shut Nvidia out of the process... and that pisses me off.

So when someone suggests that the new Intel GMA would be acceptable, (when it's not even as fast as our 'old' 9400) I find that irritating.

It's too bad things got a little nuts in this thread, but I do understand Scottsdale's irritation on this one.

 

[applesupergeek]

So when someone suggests that the new Intel GMA would be acceptable, (when it's not even as fast as our 'old' 9400) I find that irritating.

It's too bad things got a little nuts in this thread, but I do understand Scottsdale's irritation on this one.

Things went a little south buddy because some people would have us believe that it is acceptable to have a new air with inferior graphics technology, which is the thing that the air hurt more to begin with. This is the plane and simple reality of the matter. They quoted some test run by sites that have most probably strong marketing ties with intel, intel being the giant it is, and even these don't conclusively display and superior performance by the intel graphics, just being worse off than the previous gen nvidia option currently in the air. But even if they did the press is so often a mouthpiece for the pr of such companies. Years and year intel was sustained by such tec sites when AMD had a far superior product. Moreover even when their alleged tests showed intel graphics as comparable to actual options, almost always the reality has been that in real life usage they just don't perform equally well. And instead of those points made being addressed some posters chose to use innuendoes and half truths derailing the discussion.

 

[madmaxmedia]

I think it's somewhat pointless to argue whether it is acceptable or not for Apple to go back to Intel integrated graphics, it's obviously a subjective argument. I think everyone would prefer Apple continue with Nvidia anyways.

I think that Apple may unfortunately be more or less forced to go back to Intel graphics though due to 'Intel vx. Nvidia', and that's the reality we're dealing with.

I guess I will hold onto my 13" MBP, it will go through the same change as well.

 

[Scottsdale]

They do not get mad when someone doubts the validity of their sources, or even the sources themselves (as what happens all the time in forums). Unlike you.

Pot calling the kettle black (as you see I am not the first to say that to you), especially when both kettles aren't even black.

Obviously you can't respectfully disagree with anyone with a different opinion from yours, especially if it's MBA related, calling them "disgusting" and such. You know what that reminds me of? Yeah, those posts concerning your "source" before September.

Exactly. Throwing out your own opinion. Link me to a source that says that. Remember, I used to state 9400M = 2x GMA HD = 4x previous GMA, and yes that was from a source (which was fairly old and that's why I've stopped saying it).

Not necessarily incorrect. Just unsourced.

Didn't I mention independent reviewers in that one post?

Nobody is asking for exact percentages. We are asking for approximate percentages. A range of 10% is good enough.

Because they have looked up sources?

So give a few instead. Articles will help me; long posts won't.

Where in this thread have I advocated the use of the Intel IGP in the MBA?

To the point of fanboyism.

Good, so you won't reply to this post. Or for that matter, why did you make this post anyway?

Likewise, I'm done here. (At least that's what I hope, because I don't want to waste my time with irrational fanboys.) I'll just quote one of the most reasoned posts in the last two pages.

Of course it ticks me off when someone calls me a liar. I take pride in the fact that I am an extremely honest person. While my opinions may differ from yours, I don't stand up in the middle of a different debate and call you a liar regarding some other topic; and you did just this by bringing up my believing and sharing certain information that didn't pan out with the latest MBA update. It wasn't even information that I would have wanted the MBA to have in any way, shape or form. There was no reason in the world for me to fabricate the information as you have suggested by bringing it up.

You don't know anything about me or my race, creed or color, so drop that BS now!

My opinion is ALWAYS MY OPINION. It doesn't mean I am stating anything as fact when I believe a potential Intel IGP in the MBA would be a 50% to 60% loss. The only FACTUAL way to prove that would be an Intel IGP in an MBA to compare it with. NOBODY can do that but Apple. You are failing in your own opinion to understand DETAILS. The detail you are most lacking of is the difference between a test with Windows and OS X. Just because Intel's IGP was in one article/advertisement (with ties to Intel) and showed an improvement of Intel's IGP doesn't mean it will even have the same result if examined with OS X. The bottom line is Windows and OS X are very different, and taking one article that may or may not be embellishing Intel's IGP and applying it over to OS X is a ridiculous assumption.

Not only that, do the math. Intel's 3100 was 20% of the capabilities of Nvidia 9400m, ACCORDING TO APPLE. However, I remember reading other articles that basically said their test results were similar (may or may not be true, it was just printed on the Internet). Now, you're implying that Intel's IGP in two updates has improved 500%. Is that likely? Where are the numbers to support that with regards to the MBA? Have you read any scientific study, using the current MBA's components, the potential MBA's components, and the same exact OS X operating system and application software? If you haven't read that exactly, you are making too many assumptions in supposing something as factual. Forget what you read on the Internet, it's not all fact!

Who are you to determine whether a review is "independent" or not? Anyone can go publish ANYTHING on the Internet. Some will choose to believe certain information is "fact," when in actuality it's not factual. Just because you cite a source doesn't make the information factual! Do you understand that? You tell me to cite my source, but anyone can go prove anything on the Internet, and that doesn't make them correct. Can you understand that the Internet is full of incorrect information?

My opinion remains that if we had an MBA with a new Arrandale CPU in it and only the IGP, that it would be a 50% to 60% loss in total graphics performance (I do believe that's accurate within 10%). It will always be an opinion until we have both MBAs in hand and we can research it, scientifically prove it is the standard repeatable performance, and make ourselves believe that our past opinion was indeed correct or incorrect. No matter what Internet article you link me to it doesn't mean that I am incorrect. It just means that you perceive the article's information to be factual and differ from my opinion. To be factual that that article must have accurately performed a scientific study that could be repeated using the hardware that was in the MBA AND MORE IMPORTANTLY THE SAME SOFTWARE, OS X. Just because something is true with Windows does not make it true with OS X.

I do really love the MBA. Am I a fanboy? It depends on your definition of fanboy. I can definitely point out all of our compromises in using OS X over Windows and an MBA over an MBP. I can definitely admit to all of the MBA's problems. I definitely understand that Apple's Macs are far behind the competing PCs on the market. I definitely have called out Apple's problems in the past. I typically think of a fanboy as someone who cannot see the truth. I don't think that's me, but that's your own opinion, and you're entitled to it whether you're correct or not. Do you understand what I am saying? YOU ARE ENTITLED TO YOUR OWN OPINION.

We all make our arguments for and against certain statements, concepts, or beliefs. It is definitely always just our opinions. It's an opinion to believe in something posted on the Internet. A lot of people have said they enjoy my posts, and they're entitled to their own opinion. You, and a lot of other people, don't like my posts, and all of them are entitled to their own opinion. I don't believe you have cited "factual" information in how the MBA would perform with OS X and only Intel's IGP, and I am entitled to my own opinion. Whether each person is for or against someone's statement is their own opinion.

I believe your replies to me have been naive and insulting. While you may choose to believe something different is your own opinion. I didn't intend to try to insult anyone here. My intention is to share the fact that I believe Intel's IGP would be a terrible mistake as a sole solution for the graphics used in the MBA (the FACT in this statement is I believe that Intel's IGP would be a terrible mistake). We each post our own OPINIONS here. It is the duty of each reader to decide for himself/herself whether they choose to believe opinions or not. Some people enjoy my posts. Some people hate my posts (like yourself). You're entitled to your own opinion, but to go on with personal attacks to try to discredit a source is, quite frankly, a sad way to make people believe your own arguments over my arguments.

 

[Anonymous Freak]

They quoted some test run by sites that have most probably strong marketing ties with intel...

No, it went South because of baseless unsubstantiated claims such as this.

So I take it that the *ONLY* source you trust for independent reviews of *ANYTHING* is Consumer Reports? Your arguments smack of "This review disagrees with me, therefore it must have been bribed by the company I dislike. And because that company bribes reviewers, I dislike them!"

Yes, because Intel would bribe a website to issue a review that shows that their latest integrated graphics are only slightly inferior to three year old low-end graphics....... Now if you find a website that has three extremely obscure games that Intel graphics win by 50% over current nVidia and AMD low-end discrete graphics, you might have a case. But when it shows Intel graphics winning one of four tests, by a whopping 3%, and losing the rest by about 3-5%, that doesn't strike me as them being in bed with Intel.

Really, if you want to convince people, you should use actual facts, not baseless conjecture.

We're both talking about what we estimate Apple may do. It's not worth descending into insanity over. If you wouldn't buy a MacBook Air with Intel graphics, fine! Don't! However to call even the mere idea of it stupid, and call anyone/everyone who suggests that it might even be a possibility idiots is just massively ignorant and insulting.

Now, you're implying that Intel's IGP in two updates has improved 500%. Is that likely? Where are the numbers to support that with regards to the MBA?

Obviously, we won't have any MBA numbers until/if Apple releases an Arrandale MBA. But if you're unwilling to accept similar comparisons in the Windows world, then you can't reasonably make *ANY* estimates about *ANY* unreleased product of any kind. By your measure, we can't even assume that the upcoming 6-core processors would be faster in the Mac Pro than the current 4-core processors.

But, yes. I am saying that it has improved 500%. According to aggregated benchmarks at Notebookcheck.net, Intel HD Graphics is, in general, 5x as fast as X3100. And is about a wash with the 9400M. It loses to the 9400M, sometimes by large numbers, but it also wins against the 9400M.

My opinion remains that if we had an MBA with a new Arrandale CPU in it and only the IGP, that it would be a 50% to 60% loss in total graphics performance (I do believe that's accurate within 10%).

I wouldn't estimate it that far, based largely on the Notebookcheck.net numbers above; but yes, it would be a performance loss. I have never attempted to argue otherwise. Yes, it would be great if Intel were to license the DMI/FDI bus to nVidia so nVidia could make a low-power integrated-GPU Southbridge. Yes, it would be great if Apple went with a discrete GPU instead of using the onboard one. Looking back, you, Scottsdale, have been 'defensive' but not 'insulting', so thank you for remaining calm (even if you are 'on the other side' from my viewpoint.) It is others who have been doing the outright attacks against anyone who even makes an attempt to argue that Intel graphics wouldn't be a horrendous choice.

Mods, please shut this thread down. If people wish to discuss the actual technical merits of Intel graphics vs. nVidia, a new, less heated thread needs to be created.

 

[Scottsdale]

Mods, please shut this thread down. If people wish to discuss the actual technical merits of Intel graphics vs. nVidia, a new, less heated thread needs to be created.

Now if we "believe" Intel's IGP is not comparable to Nvidia's 9400m, you ask the Mods to shut down the thread??? Why? Because opinions different than yours somehow is hurting you or Intel? How about Free Speech? Can we not say what we believe or share our opinions? I THOUGHT THE WHOLE POINT OF MACRUMORS.COM WAS TO SHARE OUR OPINIONS!

Whether it's correct or incorrect, I will NOT believe that Intel's Arrandale IGP will even COMPARE with Nvidia's 15-month-old 9400m GPU, UNTIL I RECEIVE SCIENTIFIC REPEATABLE PROOF TO THE CONTRARY or I can use an MBA with an Intel IGP that provides the same graphics experience as Nvidia has with its 9400m. That's the biggest point, no matter what, it doesn't seem to me, that the Intel IGP can even compare with Nvidia's 15-month-old GPU. Then when Intel cannot compete, it decides to squash the competition; I believe (again an opinion) that Intel's anti-competitive practices will cost it in the long run. Unfortunately, customers are the biggest losers in the battle. Because until Nvidia wins a court case, it cannot continue to compete fairly. We are not talking about a licensing dispute here; this is Intel flexing its muscles because it's a LOSER when it comes to graphics performance versus the competition.

Furthermore, I believe in the end Apple will be using its own processors to shutout Intel. It seems obvious to me that Apple doesn't really benefit by working with Intel, and it's obvious Apple knows it too. Why did Apple buy its own chip company? We are seeing it in the iPad, Apple is shutting out Intel and using its own chip, which creates competition and better products for consumers. Depending on other companies who act anti-competitively, like Intel, is a terrible business strategy. Businesses often work together to provide a better experience for the customer; Nvidia and Apple have tried to create an alliance to provide a better graphics experience for Mac users. It really worked out well for us, for a while.

Intel still provides great CPUs, but the real constraints are not the CPUs in current computers. The CPUs on the market are incredibly great, and we don't really need Arrandale over Penryn C2D. What computers need for improvement is better software that takes advantage of the hardware available, better drive technology that provides better throughput (speed as it's usually the bottleneck), and better graphics to provide the user with a better visual experience that compares with the processing performance of the CPU.

 

[applesupergeek]

No, it went South because of baseless unsubstantiated claims such as this.

So I take it that the *ONLY* source you trust for independent reviews of *ANYTHING* is Consumer Reports? Your arguments smack of "This review disagrees with me, therefore it must have been bribed by the company I dislike. And because that company bribes reviewers, I dislike them!"

Yes, because Intel would bribe a website to issue a review that shows that their latest integrated graphics are only slightly inferior to three year old low-end graphics....... Now if you find a website that has three extremely obscure games that Intel graphics win by 50% over current nVidia and AMD low-end discrete graphics, you might have a case. But when it shows Intel graphics winning one of four tests, by a whopping 3%, and losing the rest by about 3-5%, that doesn't strike me as them being in bed with Intel.

Really, if you want to convince people, you should use actual facts, not baseless conjecture.

The fact that intel has a strong arm in tec sites and would have them preview (mark my word again PREVIEW) something under their own terms isn't conjecture, it is a fact. All the more so when they are in the middle of one of their most anticompetitive moves recently, that of saddling their quite good chips with a worst in class gpu.

Don't worry though a Federal Trade Commission antitrust lawsuit has been already filled against them.

Of course based on some tests someone run on a piece of hardware and posted on the internet, you want us to bow down and accept them as gospel. And when those very tests prove intel's igfx still inferior to old tec by nvidia, despite selectively running their own tests, and being privy to the new igfx from intel thanks to their sponsors, intel.

Well tough luck some people here see this for what it is and they don't like it, and we wish apple would find some way to bypass that whole backward step for the air that intel signalled with their dishonest, anticompetitive practises and their crap igfx that they couldn't even sell at a swap meet.

Not only that, you want this thread closed...my, my a lot of people have intel stock nowadays it seems...

 

[findingforever]

What processors would be ideal in the next MBA? Everything I'm hearing points to a slower Rev D MBA. What solutions does the MBA have to actually improve performance? Does Arrandale have any solutions?

 

[iMacmatician]

I take pride in the fact that I am an extremely honest person.

If you really were an honest person you would not even reply to my last post because you said that you made a mistake replying to me. So are you going to keep making your "mistake" or not?

I don't stand up in the middle of a different debate and call you a liar regarding some other topic

I brought up what happened then because it relates to the point of jimboutilier's post.

My opinion is ALWAYS MY OPINION.

Exactly. Don't make it as fact or force it on anyone.

It doesn't mean I am stating anything as fact when I believe a potential Intel IGP in the MBA would be a 50% to 60% loss.

That's exactly what you've been stating/implying the whole time.

The only FACTUAL way to prove that would be an Intel IGP in an MBA to compare it with. NOBODY can do that but Apple.

So why are you throwing numbers around?

You are failing in your own opinion to understand DETAILS.

While you are mentioning them yet ignoring them in your "analysis."

The detail you are most lacking of is the difference between a test with Windows and OS X. Just because Intel's IGP was in one article/advertisement (with ties to Intel) and showed an improvement of Intel's IGP doesn't mean it will even have the same result if examined with OS X. The bottom line is Windows and OS X are very different, and taking one article that may or may not be embellishing Intel's IGP and applying it over to OS X is a ridiculous assumption.

Again, so why are you throwing around numbers? And what if OS X shows a greater improvement?

Now, you're implying that Intel's IGP in two updates has improved 500%. Is that likely? Where are the numbers to support that with regards to the MBA?

Already addressed by ehurtley.

Have you read any scientific study, using the current MBA's components, the potential MBA's components, and the same exact OS X operating system and application software which gives the HD graphics performance "50%-60% lower" than the 9400M's? If you haven't read that exactly, you are making too many assumptions in supposing something as factual.

Fixed that for you. Again, mentioning yet ignoring details.

Forget what you read on the Internet, it's not all fact!

And your OPINONS are?

Who are you to determine whether a review is "independent" or not?

Who are you to determine whether a review is "biased" or not?

Just because you cite a source doesn't make the information factual! Do you understand that?

A step above throwing numbers out of nowhere, right?

You tell me to cite my source, but anyone can go prove anything on the Internet, and that doesn't make them correct. Can you understand that the Internet is full of incorrect information?

So, you have no citation whatsoever, and when others disagreeing with you have citations, you expect us to believe your rants over them?

I typically think of a fanboy as someone who cannot see the truth. I don't think that's me, but that's your own opinion, and you're entitled to it whether you're correct or not.

You and applesupergeek are obviously fanboys, because:

1. You are overpassionate about one or a few things (GPU)
2. You ignore benchmarks with conclusions different to your beliefs, even if those types of benchmarks are widely accepted elsewhere
3. You get mad at and accuse people when they challenge your ideas
4. You make statements without any basis whatsoever to try to "challenge" 1. and 2.
5. Anyone who challenges you is either (or both) an employee (in this case of Intel) or a stockholder, and therefore cannot see clearly

I don't believe you have cited "factual" information in how the MBA would perform with OS X and only Intel's IGP, and I am entitled to my own opinion.

Neither have you. So don't go around bashing people who challenge your "50%-60%" analysis, which as far as I can tell has no source whatsoever attached to it.

I believe your replies to me have been naive and insulting.

Funny thing to say for someone who is single-minded at a set of made-up performance values and dismisses anything contrary to that.

You're entitled to your own opinion, but to go on with personal attacks to try to discredit a source is, quite frankly, a sad way to make people believe your own arguments over my arguments.

Who here has been trying to discredit all the sources that show the HD graphics is a contender against the 9400M?

Now if we "believe" Intel's IGP is not comparable to Nvidia's 9400m, you ask the Mods to shut down the thread???

No, read his post.

It is others who have been doing the outright attacks against anyone who even makes an attempt to argue that Intel graphics wouldn't be a horrendous choice.

Mods, please shut this thread down. If people wish to discuss the actual technical merits of Intel graphics vs. nVidia, a new, less heated thread needs to be created.

How about Free Speech?

Free speech does not extend to attacks or overly heated up arguments.

Can we not say what we believe or share our opinions?

See above.

I THOUGHT THE WHOLE POINT OF MACRUMORS.COM WAS TO SHARE OUR OPINIONS!Of course, you blow up at ehurtley without even properly reading his post. Thanks for proving our point.

Whether it's correct or incorrect, I will NOT believe that Intel's Arrandale IGP will even COMPARE with Nvidia's 15-month-old 9400m GPU, UNTIL I RECEIVE SCIENTIFIC REPEATABLE PROOF TO THE CONTRARY or I can use an MBA with an Intel IGP that provides the same graphics experience as Nvidia has with its 9400m.

Given your past posts I expect you to find some bias or error in whatever scientific studies come up. If not, then congratulations.

Why did Apple buy its own chip company? We are seeing it in the iPad, Apple is shutting out Intel and using its own chip, which creates competition and better products for consumers.

They did it because Intel does not have a CPU that works in that power envelope.

What computers need for improvement is better software that takes advantage of the hardware available, better drive technology that provides better throughput (speed as it's usually the bottleneck), and better graphics to provide the user with a better visual experience that compares with the processing performance of the CPU.

More RAM too. I'm not liking the relatively low RAM ceiling in recent notebooks.

Of course based on some tests someone run on a piece of hardware and posted on the internet, you want us to bow down and accept them as gospel.

Not as much "gospel" as "more reliable and higher likelihood of being accurate than guesses."

And while you dismiss those benchmarks…

And when those very tests prove intel's igfx still inferior to old tec by nvidia, despite selectively running their own tests, and being privy to the new igfx from intel thanks to their sponsors, intel.

…you want us to believe these benchmarks?

Not only that, you want this thread closed...my, my a lot of people have intel stock nowadays it seems...

If that's true then maybe that's saying something about the company.

 

[iMacmatician]

What processors would be ideal in the next MBA? Everything I'm hearing points to a slower Rev D MBA. What solutions does the MBA have to actually improve performance? Does Arrandale have any solutions?

Arrandale LV should provide increased performance over Penryn LV. The speculated slow CPUs are the Arrandale ULVs, which given past trends probably won't go in the MBA.

EDIT: Actually read ehurtley's post, which explains it very well.

 

[Anonymous Freak]

What processors would be ideal in the next MBA? Everything I'm hearing points to a slower Rev D MBA. What solutions does the MBA have to actually improve performance? Does Arrandale have any solutions?

What would actually be *BEST* for the Air would be an Arrandale plus low-end discrete GPU. You'd trade off battery life for better graphics performance. There has been pages of debate on the possibility of using the integrated GPU.

The options Apple has:

1. Apple keeps Core 2 Duo + nVidia 9400M. This is a two chip solution (processor is one chip, GPU/northbridge/southbridge is a second chip.) The problem is that it already has the fastest Core 2 Duo at that power range. (17W Core 2 Duo SL9600 at 2.13 GHz.) They could just get rid of the the lower-end 1.86 GHz model and make the faster processor cheaper, but that wouldn't be much of an upgrade. And Intel is unlikely to release any new Core 2 Duo processors. Last year nVidia said they weren't going to do any more chipsets, so unless they've been going against that in secret, I doubt we'll see a new chipset.

2. Apple moves to Arrandale with integrated graphics. This is a two chip solution (processor/GPU/northbridge is one chip, southbridge is a second chip.) Arrandale is available with 18W total consumption chips (that's 18W for both CPU and GPU,) the fastest the Core i7-640UM, which is nominally 1.2 GHz, but can "Turbo Boost" up to 2.26 GHz. Considering further, I doubt Apple would use this, simply because while it would use considerably less power than my option 1, it would be quite a step back, in both CPU and GPU. (Its GPU is only 500 MHz.) More likely would be one of the 25W chips, which would likely put it on par with the Core 2 Duo+nVidia solution, power-wise. The likely options here are the Core i7-640LM and i7-640LM. They are 2.0/2.8 and 2.13/2.93 GHz, respectively (standard/turbo,) so even their non-Turbo speeds are on par with the current Core 2 Duo; plus they have a slightly faster 566 MHz GPU. This would be quite a bit faster at CPU work than the Core 2 + nVidia; and has debatable performance on GPU work.

3. Apple moves to Arrandale with discrete graphics. This is a three chip solution (processor/northbridge is one chip, southbridge is a second chip, GPU a third.) To fit in the current power envelope with a discrete GPU, Apple would possibly have to resort to the 18W Arrandales. I haven't seen any tests that see how much power an Arrandale draws with a discrete GPU, it may be possible to get away with the 25W CPUs and still not have too much battery difference. But then we open up GPU performance to much better options, depending on how much power Apple is willing to budget. This also opens up the possibility of GPU switching, which is supported on Arrandale. This means that the system could run on the internal Arrandale GPU for better battery life, and on a discrete GPU for better performance. The current MacBook Pro offers this with 9400M/9600M GT switching, but that requires a log out to switch. Some PC notebooks offer this ability without a log out, although many require a log out, some even a reboot. It's all in the combination of hardware and driver design. I would hope that if Apple goes this route, they make it transparent, no log out needed.

The third option would be, by far, the best option from a performance standpoint. The question being debated in this thread is the probability of which option Apple picks.

(Of course, further options are that Apple abandons the Air in favor of the iPad; or repositions the Air as an ultra-mobile, but with how Steve slammed netbooks in the iPad intro, I doubt they'd do that.)

 

[findingforever]

Thanks ehurtley, very informative. I'm holding out for "transparent" option 3. This would almost certainly warrant/require a redesign right?

 

[jimboutilier]

Now if we "believe" Intel's IGP is not comparable to Nvidia's 9400m, you ask the Mods to shut down the thread??? Why? Because opinions different than yours somehow is hurting you or Intel? How about Free Speech? Can we not say what we believe or share our opinions? I THOUGHT THE WHOLE POINT OF MACRUMORS.COM WAS TO SHARE OUR OPINIONS!

Whether it's correct or incorrect, I will NOT believe that Intel's Arrandale IGP will even COMPARE with Nvidia's 15-month-old 9400m GPU, UNTIL I RECEIVE SCIENTIFIC REPEATABLE PROOF TO THE CONTRARY or I can use an MBA with an Intel IGP that provides the same graphics experience as Nvidia has with its 9400m. That's the biggest point, no matter what, it doesn't seem to me, that the Intel IGP can even compare with Nvidia's 15-month-old GPU. Then when Intel cannot compete, it decides to squash the competition; I believe (again an opinion) that Intel's anti-competitive practices will cost it in the long run. Unfortunately, customers are the biggest losers in the battle. Because until Nvidia wins a court case, it cannot continue to compete fairly. We are not talking about a licensing dispute here; this is Intel flexing its muscles because it's a LOSER when it comes to graphics performance versus the competition.

Furthermore, I believe in the end Apple will be using its own processors to shutout Intel. It seems obvious to me that Apple doesn't really benefit by working with Intel, and it's obvious Apple knows it too. Why did Apple buy its own chip company? We are seeing it in the iPad, Apple is shutting out Intel and using its own chip, which creates competition and better products for consumers. Depending on other companies who act anti-competitively, like Intel, is a terrible business strategy. Businesses often work together to provide a better experience for the customer; Nvidia and Apple have tried to create an alliance to provide a better graphics experience for Mac users. It really worked out well for us, for a while.

Intel still provides great CPUs, but the real constraints are not the CPUs in current computers. The CPUs on the market are incredibly great, and we don't really need Arrandale over Penryn C2D. What computers need for improvement is better software that takes advantage of the hardware available, better drive technology that provides better throughput (speed as it's usually the bottleneck), and better graphics to provide the user with a better visual experience that compares with the processing performance of the CPU.

Sharing an opinion is one thing. Yelling and trying to browbeat everyone else til they conform to your opinion (particularly when they have facts that appear to dispute your opinion) is just annoying and rude. I mean your BOLD button must just be worn to a frazzle. Relax. We're hear to help each other out.

I expect Intel will eventually loose, pay a price for its anti competitive behavior and Nvidia will be able to offer their superior products with Intel CPU's once again. In reality though that could be quite a while off and until then Apple only has four choices:
- No updates
- Stick with the current CPU/GPU but speed bump them
- New CPU and Intel IGP
- New CPU and ATI GPU

Just because something may appear to be a better theoretical solution doesn't mean its a better practical solution. The PowerPC was a far better solution than Intel was offering when Apple started offering PowerPC.s But Intel's market share allowed more iterations and better R&D and the technically inferior solution became the better practical solution. I think Apple learned its lesson there and won't be leaving the mainstream for their PC CPU's anytime soon. But thats just my opinion ;-)

 

[Anonymous Freak]

Thanks ehurtley, very informative. I'm holding out for "transparent" option 3. This would almost certainly warrant/require a redesign right?

It would require a major software update, for sure. But hardware-wise, it shouldn't be a problem. The original MacBook Air was a three-chip design. (Before they moved to nVidia chipsets, they were using Intel's CPU+two-chip-chipset design. Intel has only in the last four months moved to a CPU+one-chip design.)

 

[robeddie]

k.

3. Apple moves to Arrandale with discrete graphics. )

This, sir... would be a dream if they did this!

My question, is whether there's the space or heat headroom to accommodate this.

Damn I sure hope so!

 

[iMacmatician]

If a redesign happens, it is likely due to accommodate the higher total TDP of the three-chip solution (that is if the 3-chip solution has a higher TDP than what is currently being used).

 

[robeddie]

If a redesign happens, it is likely due to accommodate the higher total TDP of the three-chip solution (that is if the 3-chip solution has a higher TDP than what is currently being used).

Right, but that - presumably - would mean a bigger, thicker MBA, no?

Which, of course, would kind of be defeating the purpose.

 

[Anonymous Freak]

P.S., as this has gotten heated, I feel I should re-iterate in larger print, and with more words, what is in my signature:

I am a contract employee working for a company other that Intel; however at present I am working temporarily inside an Intel facility. I receive no money or benefits directly from Intel; and even if I piss Intel off, my paycheck will continue to come through the third-party company.

I have nothing to do with engineering or marketing. I have no access to anything Apple, not even information about Apple.

I have no access to any unreleased notebook or desktop hardware of any kind, and don't even have any of the just-released (Arrandale mobile, Clarkdale desktop,) equipment. I have no access to internal information about it, either. (A single Clarksfield laptop with nVidia graphics recently was brought through this department, there was much ogling. It is gone now. There are apparently a couple Lynnfield dekstops hiding somewhere, I have not touched them.)

All of my statements are my own personal opinion, with my facts largely drawn from sources that are widely considered reputable (Ars Technica, Anandtech, [H]ard|OCP, and the like,) or from outright factual-statement sources (Intel's publicly-accessible support pages for specifications of chips, for example.)

I have no innate 'love' of Intel. I happen to own Intel CPU systems at home right now because it's what Apple offers; or it had the best bang for the buck when I bought it. I have happily purchased PowerPC Macs in the past, and AMD processor-based PCs, when they provided the best bang for the buck. I am GPU-vendor neutral, as well; purchasing both nVidia or AMD, whichever provided the best experience at the time. (And Matrox, and 3dfx, etc, etc...)

 

[Anonymous Freak]

Right, but that - presumably - would mean a bigger, thicker MBA, no?

It would depend on which chips they went with. If they go for the 18W Arrandale, and a very-low-wattage GPU, there would be no problem at all. But that would likely end up worse than the current Air, performance-wise.

I have not seen any numbers on how much power the 25W Arrandale draws by itself when used with a discrete GPU; it is possible that 25W Arrandale+southbridge+low-end GPU could have the same power envelope as Core 2 Duo+9400M. I can't find any power draw numbers for the 9400M by itself, but the HM55 southbridge is claimed to draw 3.5 Watts; so the total of 25W Arrandale+southbridge is 30W.

The Air's launch chipset, 945GM, drew 7 Watts for the northbridge, 3.3 Watts for the southbridge, totalling 27.3W for all three chips (using the 17W Core 2 Duo.) I would imagine that Core 2 + 9400M comes to about the same.

But, again, we don't know how much the power draw of Arrandale would be with the internal GPU going unused, so we won't know what discrete GPU would match nicely, power-wise. (One of the things about Arrandale is that it can dynamically reassign power to the CPU or GPU depending on load; so it's possible that the CPU by itself would claim the full 25W to run faster more often when the GPU is not being used.)

Edit: A quick search shows that the Radeon Mobility 5830 draws 25 Watts, so it's out of the question in the Air. The 5430 only draws 7 Watts, though; and it is still theoretically much faster than the 9400M. So it might be a good match to Arrandale. I haven't seen any hard benchmarks on that chip yet, though.

 

[Scottsdale]

It would depend on which chips they went with. If they go for the 18W Arrandale, and a very-low-wattage GPU, there would be no problem at all. But that would likely end up worse than the current Air, performance-wise.

I have not seen any numbers on how much power the 25W Arrandale draws by itself when used with a discrete GPU; it is possible that 25W Arrandale+southbridge+low-end GPU could have the same power envelope as Core 2 Duo+9400M. I can't find any power draw numbers for the 9400M by itself, but the HM55 southbridge is claimed to draw 3.5 Watts; so the total of 25W Arrandale+southbridge is 30W.

The Air's launch chipset, 945GM, drew 7 Watts for the northbridge, 3.3 Watts for the southbridge, totalling 27.3W for all three chips (using the 17W Core 2 Duo.) I would imagine that Core 2 + 9400M comes to about the same.

But, again, we don't know how much the power draw of Arrandale would be with the internal GPU going unused, so we won't know what discrete GPU would match nicely, power-wise. (One of the things about Arrandale is that it can dynamically reassign power to the CPU or GPU depending on load; so it's possible that the CPU by itself would claim the full 25W to run faster more often when the GPU is not being used.)

Edit: A quick search shows that the Radeon Mobility 5830 draws 25 Watts, so it's out of the question in the Air. The 5430 only draws 7 Watts, though; and it is still theoretically much faster than the 9400M. So it might be a good match to Arrandale. I haven't seen any hard benchmarks on that chip yet, though.

The original MBA actually had a 65nm process Merom 20W TDP CPU, not a 17W CPU as you mentioned. In addition, it included a 965GMS chipset not a 945GM. Both CPU and chipset were miniaturized versions making them much smaller than the standard versions. If you add 20W + NorthBridge + SouthBridge (if your listed TDPs are correct = 10.3) = 30.3 total not 27.3 total.

The v 2,1 MBA (we refer to as B and C here), has a 17W TDP Penryn C2D Low Voltage CPU. In addition, the Nvidia 9400m is 12W TDP. Here we are at 29W total which shows the all-around advancement in using Nvidia over Intel.

Also, I read a Fudzilla article the other day that states the IGP with the Arrandale CPUs can be "turned off." I still didn't read anything that stated the effects of demand of power once IGP is "turned off." It would be nice to know...

Another interesting bit of information is that Asus UL80JTS notebook has an i7 CPU and Nvidia GeForce 310 graphics that automatically switch back and forth as needed. The result is claimed to be 10 hour battery usage. I know it's a different system, but the results and capabilities show what's possible and how to avoid Intel as a sole graphics solution. I really wish Apple could make something similar with less demanding graphics (like a low-end ATI dedicated solution). According to the article I read over at Engadget, the Sony Vaio Z will have something like this too. It really makes so much sense. The 80% of the time I am on the computer to write, I don't need dedicated graphics. The rest of the time, the computer can automatically select the dedicated graphics for me.

I think there are plenty of options out there for Apple to consider. It really doesn't have to just keep using C2D CPUs to avoid Intel's IGP.

 

[applesupergeek]

however at present I am working temporarily inside an Intel facility. I receive no money or benefits directly from Intel; and even if I piss Intel off, my paycheck will continue to come through the third-party company.

Really, you have some affiliation with intel? You are joking, I 'd have never guessed.

 

[iDisk]

That article was the reason I initially posted.. what do you think about it?

I was hoping that this new chip would be a bit better for gaming. I realize that 3DMark isn't the end all, but it gets a better 3dmark05 score than the 9400m. Wonder if better drivers would allow this chip to perform better & really open up its potential?

MAC OS X & INTEL GPU's


Mac OS X 10.4 had support for the GMA 950, since it was used in previous revisions of the MacBook and 17-inch iMacs. It has been used in all Intel-based Mac minis (until Mac Mini released on March 3, 2009). Mac OS X 10.5 Leopard contains drivers for the GMA X3100, which were used in a recent revision of the MacBook range.

Late-release versions of Mac OS X 10.4 also support the GMA 900 due to its use in the Apple Developer Transition Kit, which was used in the PowerPC-to-Intel transition. However, special modifications to the kext file must be made to enable Core Image and Quartz Extreme.

Although the new MacBook line no longer uses the X3100, Mac OS X 10.5 (Leopard) ships with drivers supporting it that require no modifications to the kext file. Mac OS X 10.6 (Snow Leopard), which includes a new 64-bit kernel in addition to the 32-bit one, does not include 64-bit X3100 drivers.This means that although the MacBooks with the X3100 have 64-bit capable processors and EFI, Mac OS X must load the 32-bit kernel to support the 32-bit X3100 drivers. November 9's 10.6.2 update ships with 64-bit X3100 drivers.

The newer MacBook/Airs and MacBook Pro notebooks instead ship with a far more powerful NVIDIA GeForce 9400M, and the 15" and 17" MacBook Pro notebooks ship with an additional GeForce 9600GT supporting hybrid power to switch between GPUs. The NVIDIA GeForce 9400M chipset as implemented in current Apple Macbooks does not support composite or S-video output.

INTEL & GAMING


The Intel GMA products are designed to allow Intel to offer a full system platform that includes graphics hardware. However, due to the GMA's nature as a highly cost-sensitive product, performance and functionality are limited relative to more expensive discrete graphics components. Some games and 3D applications will not recognize support for some hardware functionality because of the simplification of parts of these graphics accelerators. The GMA X3x00's unified shader design allows for more complete hardware functionality, but the line still has issues with some games and has significantly limited performance.

Also Intel has put up a page with 'Known Issues & Solutions' for each version. For Intel Graphics Media Accelerator Software Development concerns, there is the Integrated Graphics Software Development Forum.... in other words their GPU Quality is ABHORRENT!

NOT TO MENTION LARRABEE BUT I GUESS EVERYONE FORGOT ABOUT THAT DISASTER

FORGET ABOUT INTEL GPU GUYS THEY JUST DON'T HAVE IT RIGHT

 

[applesupergeek]

Also Intel has put up a page with 'Known Issues & Solutions' for each version. For Intel Graphics Media Accelerator Software Development concerns, there is the Integrated Graphics Software Development Forum.... in other words their GPU Quality is ABHORRENT!

NOT TO MENTION LARRABEE BUT I GUESS EVERYONE FORGOT ABOUT THAT DISASTER

FORGET ABOUT INTEL GPU GUYS THEY JUST DON'T HAVE IT RIGHT

Thanks buddy for chiming in with some voice of reason here, because this thread has sadly become overpopulated with very vocal proponents of whatever junk intel dishes at us, with their critical faculties in judging intel highly diminished.

I personally haven't forgotten about the Larrabee fiasco, the igpu turned software framework, aka vapourware.

And just to add to what's been said above, I too would love it if apple could add a discrete option and turn off the intel igfx, or at least have it switch on on the most basic tasks. Unfortunately the air is incredibly thin, and while I see them being able to offer such an option in macbooks and pros I really can't see how they ll fit one in the air. But even if they do we will have to wait for an incubation period to see if this actually works thermal's wise, because gen A, despite working in the labs and on paper, was a far cry from managing the thermals well. Again, due to the intel igfx.

The air is probably the most beautiful laptop on the market at the moment, let's see what apple can figure out to make up for intel's fast one.