It's for unspecified 15 and 17-inch models, but our unibody 15-incher said it didn't need the update, so it's probably for previous-gen models confirmed to have bad graphics cards. [Apple via ars]

Again, nothing's confirmed just yet, but preliminary testing has shown users who artificially set the MacBook Pro's cooling fan to 3,000rpm eliminate the screen burn-in issue we told you about earlier this week.

One MacBook Pro user, posting a reply in the Apple Support Forum, said the fans are intended to run at 2,000rpm until default conditions, and then spool up to 3,500 when high-temperature conditions kick in. You know, like what happens when you're using a GPU like the Nvidia 9600M.

That said, the MacBook Pro is having a little trouble getting it up, so to speak, and remains at about 2,050rpm when the temps rise to the point where the screen begins to flake out. That's bad.

What's good is this seems like a simple software fix, at least from the sound of this latest discussion. [Slashgear]

Complaints run something like this: When the more powerful Geforce 9600M is switched on, the display is ravaged by green lines and other weirdness, which gets worse as the card appears to heat up. Switching to the integrated 9400M cures the problem, as does restarting (at least temporarily). Another user notes that the problems flare up doing gaming, but calms down when they're "doing lower-performance stuff" which also seems to point to a heat issue.

This is actually the third run of MacBook Pros to be allegedly affected by defective Nvidia graphics cards with heat problems: The previous generation was packed with Nvidia cards that Apple determined were suffering from its notorious material defect that caused cards to fail at a "higher-than-normal" rates. The current generation of 15-inch MacBook Pros also had video problems (the ominous "black screen of death") which Apple also acknowledged, though no definitive blame was ever placed on faulty cards. (Although the Inquirer will readily tell you the cards are in fact, defective.)

Are these graphics cards, in fact, bad? It's too soon to tell—and very possible we're just hearing from a tiny, tiny but vocal minority that's been afflicted—especially without definitive word from Apple. I'd say, as I did before, that given the not-so-distant past, it doesn't look very good for Nvidia, whatever the actual case may be. [Apple via Engadget]

In our case, it simply didn't fix the MiniDisplay Port to Dual-Link DVI distortion problems, but according to Apple's forums, it opened up a whole new bag of hurt for some people:

I downloaded the update today and regret that I did. The problem for me is now worse. I cannot use my external samsung 22inch monitor on extended desktop mode, only mirror mode, and that is not helpful to me. In extended mode it goes black for extended times, flicking on occasionally to remind me that its there still.

If you already installed in the update, has it borked out on you? [AppleInsider]

I'm a sucker for any Apple keynote, whether I'm in the stands with the press or F5'ing away at my browser like most of you. There's a certain hypnotic timing to them, a clean presentation and a controlled, meticulous unveiling process that makes even the less exciting products seem monumental in scope.

During today's last Apple Macworld keynote, not having been there, I actually forgot at times that Steve Jobs wasn't presenting with Phil Schiller taking his place instead. Indeed, from the text/content alone, I wouldn't have even noticed. The presentation style was intact—from the clean slides to the dramatic structure, it was an Apple keynote, even if it wasn't Steve's Apple keynote. But I want to know, do you feel the same way?

From our original Macworld 2009 rumor roundup, here's the rundown on how all the rumors panned out.

Came True:
iLife and iWork '09 (with iWork in the cloud)
DRM-Free iTunes
17-inch Macbook Pro

Didn't Come True:
New Mac Mini
Updated 30-inch Cinema Displays

Didn't Come True Stretches:
iPod Touch Jumbo/XL/Pro/Netbook/Thing
iPhone nano
Home Server
Snow Leopard Availability
New iMacs

Seeing as many of the rumors were an admitted wild shot at best, we'd put the real rumor batting average around .400, even if it's closer to .266 on paper.

I tried downloading Radiohead's Bodysnatchers in an area providing 3 bars of 3G reception. With no fanfare, I was able to enter the iTunes store without a Wi-Fi hotspot and purchase the song.

So far, so good. No firmware updates needed or anything!

Then the download stalled for a moment. Then the iPhone pulled that classic "you're on 3G, oops, you're really on EDGE" maneuver. I wondered what would happen. Would the download stall and wait for 3G? Or would it slowly trudge on?

It trudged on. I guess that's both a good an a bad thing. While I was able to download the song, it took me nearly 10 minutes to grab the 3.7MB file. Wilson Rothman ran the same test from nearby and was also booted to EDGE, though the download only took him about 5 minutes.

Overall, I'm glad that truly wireless iTunes downloads are here. Hopefully more and better 3G coverage will make them live up to the generally pleasant iTunes experience. [Gizmodo Macworld 2009 Coverage]

While EMI is already offering DRM-free music on iTunes, they represent only 10% of the music market. With "the big three" on board, a majority of iTunes' music could be DRM-free. But the new pricing options may be the real cost of it all to consumers. While catalog titles will now start at 79 cents, the studios will be able to price music above 99 cents (to undisclosed amounts)—though CNET's source says that all music will eventually return to 79 cents.

So while it sounds like Apple finally caved to the music industry, at least consumers may get something decent out of it. Hopefully we'll hear about it all today at the keynote. [CNET]

So a quick recap on the original issue and the back and forth between Nvidia and the Inquirer: Nvidia admitted back in July "significant quantities" of notebooks are defective, built and packaged with "weak" materials that are leading to them to overheat and fail at a "higher-than-normal" rate, but declined to state which cards specifically.

The Inquirer said every single G84 and G86 card was affected, and a GPU apocalypse was coming, and indeed, more and more models from different manufacturers came to light with the problem. The previous-gen MacBook Pro was actually the last one revealed to be smitten by the plague, like the final Cylon or something, since Nvidia was reportedly less-than-honest about the problem to Apple. Nvidia continued to reiterate most chips are just peachy.

Nvidia says that its current chips don't use the weak materials, which produce what are called "bad bumps." Obviously, the Inquirer thought they were lying, so they took a new MacBook Pro to a lab, cracked it open, sliced apart the G96 GPU and checked it out under a scanning electron microscope with an X-ray microanalysis system.

The result? The Inq says the same old bad bumps, which were composed mostly of lead, are there in the GeForce 9600 vs. the new, good eutectic bumps that are in the GeForce 9400M (the MacBook and MacBook Pro's chipset/integrated graphics). Or more straight up, "The 9600 is unquestionably using 'bad bumps', directly contradicting the statements from Nvidia...It suggests that there are 15-inch Macbook Pros being sold with 'bad bumps', the same materials that brought down so many HP, Dell and Apple parts, both laptop and desktop."

Naturally, I asked Nvidia for their reaction to Charlie's Inq piece, and a spokesperson reiterated that the GeForce 9600 GT graphics cards in the MacBook Pros "don't have bad bumps at all." He said that, "yes, they're lead bumps" but "hundreds of millions of chips have lead bumps." And it's "a different material set [from the faulty one], one they transferred to earlier" that's used in the 9400, 9600 and 9800.

Of course, the Inquirer's whole point is that Nvidia is lying. So, who to believe? Well, here's what we know for sure. When I talked to Nvidia about the original run of faulty chips, and why we saw it some systems and not others, they told me it was largely a thermal issue, which, in combination with the weak materials, would cause the kiss-of-death cracking—so you'd see it in systems that ran hot, in other words, like some notebooks or slimline desktops with poor circulation. (Which is why the "fix" for the problem were firmware updates that cranked the card's fans sooner.)

We further know that the 9600 GT cards in the MacBook Pro are currently having problems that appear to be heat-related, causing them to lock up and launch into the "black screen of death." Also, Nvidia pointed out that there aren't a whole lot of labs properly equipped to do the kind of analysis the Inquirer commissioned—you can't just walk down to your local Discovery Channel store—though they left it at that.

And that's about as definitive as we can get, for now. Two things bother us: We would've liked a slice and dice of one of the previous-gen bad chips to directly compare to the new, supposedly bad one. And Nvidia's subtle implication that the people with labs equipped to perform this kind of analysis have a vested interest in the outcome also has the magical effect of shielding them from the results.

You can believe Nvidia. You can believe the Inquirer. I just know that given the thermal problems that already clearly exist, I really hope the Inquirer is wrong. If you know something about this you wanna share, email me.

Here they are, the MacBook and MacBook Pro graphics benchmarks you've been waiting for. Our basic Mac-only benchmarks used in our dual review were great at showing how close the CPU performance was, but not great at illustrating the disparity between chunky-hunky 9600 GT and the comparatively weak integrated 9400M that we saw firsthand while playing Spore. With a little time, we've been able to:
• Install Windows XP and run 3D Mark
• Play a little Crysis—yeah, Crysis
• Get ahold of Adobe CS4 and run some rendering jobs
The conclusion? If you're a serious gamer or work with graphics of any kind and you need an Apple laptop, you're gonna want the MacBook Pro. Here are the raw numbers to prove it.

Crysis

Pretty gigantic gap in Crysis performance here. It's a bit more than the 50ish% difference tossed around a few weeks ago when describing the 9600 vs. the 9400, but that's also due to the faster CPU, more RAM and increased L2 cache found in the MacBook Pro. Still, though, you can gauge from the two charts that Crysis is kinda playable on the MacBook on low, and playable on the MacBook Pro on medium. Both are chunktastic on high, and both get slower when you start shooting people in the forest.

3DMark 06
Big ass difference once again between the MB and MBP. The kicker is that the CPUs are almost identical, a fact illustrated in the benchmarks contained in our original review. It's really the card that's going to make the difference in gaming on your MacBook.

After Effects CS4 and Photoshop CS4
Rendering a project in After Effects uses a little bit of GPU action, but not very much. The difference between 9400 and 9600 on the same MacBook Pro is only two seconds of render time on a nearly seven minute project. Negligible at best.

There's essentially no difference between the two cards on the MBP on the Photoshop test, but a slight slowness when you swap over to the MacBook.

Verdict
The takeaway from these charts is that yes, there really is a difference between the MacBook and MacBook Pro in terms of graphical capability. The two benchmarks we used before, GeekBench and XBench, do a poor job of utilizing the graphics card. Whether that's because the benchmarks are two years old or because they're just poor benchmarks is irrelevant—you can totally see the difference in gameplay and in more established metering programs such as 3D Mark. If you want to be playing one of the handful of games released for Mac every year (or dual boot to Windows), you should really pick up a MacBook Pro. [MacBook and MacBook Pro Review]

Since the hardware is capable of all of these things, it means that they can all be enabled by a software/firmware/driver update. Whether or not that happens is entirely up to Apple. While you can argue that Hybrid SLI—using both GPUs at once—has a limited, balls-to-the-wall utility, being able to switch between the integrated 9400M and discrete 9600M GT on the fly without logging out would obviously be enormously easier than the current setup, and allow for some more creative automatic energy preferences—discrete when plugged in, integrated on battery. Hell, you can do it in Windows on some machines.

But since it's Apple it's also entirely possible we'll never see any of this to come to pass—GPU-accelerated video decoding has totally been possible with the 8600M GT in the previous-gen MacBook Pros, and well, you know where that stands. [Apple & Nvidia Coverage@Giz]

They noted this apparent GPU acceleration when playing back 1080p trailers from Apple's site, which saw CPU utilization drop from 100 percent in previous gen MacBook Pros to just 28 percent in current MacBooks, even though the CPU is the same speed and the Pro has more RAM (new MacBooks have a faster FSB, but it wouldn't account for this kind of bump). Nvidia cards do have hardware support for video decoding, so this would make sense, but Apple hasn't used it before.

The other possibility that MacRumors floats is that these changes could be based on improvements slotted for Snow Leopard, namely Quicktime X, "which optimizes support for modern audio and video formats resulting in extremely efficient media playback." Apple hasn't confirmed this, so if you've got one of the new MacBooks and old one to compare, check it out and let us know how it looks. [MacRumors]

Two graphics cards? It sounds crazy, preposterous, retardiculous. It's actually not. It's not unique to the MacBook Pro at all. PC users might be more familiar with Nvidia's Hybrid SLI, which pulls similar dual-card wizardry. In a nutshell, it lets you use the less power-hungry integrated graphics processor when you're doing lighter stuff to save battery, and then when you want a lot of video-crunching Mr. T powah, you can flip on the discrete graphics card. Of course, there's balls-to-the-wall full SLI too, which uses two entirely separate graphics cards in one notebook for Hulk power and about 45 seconds of battery life, like in one of Alienware's beasts.

Nvidia's standard hybrid SLI for PC actually uses both the integrated and discrete GPU at the same time when it goes into turbo mode, and it'll let you switch on the fly or have it automatically flip between the two depending on the power source. But the MacBook Pro uses Apple's spin on Nvidia's tech that simply lets you pick one or the other (not both, booooooo) and you have to manually flip the switch in system prefs, log out and back in, pretty annoying. Battery life is apparently an issue with the new MacBook Pro, considering that the integrated 9400M card now nets you five hours of go-time, the same as the separate, more power-hungry 8600M GT in the previous model, whereas the new discrete 9600M GT now gets you only four.

The other major reason for the huge upgrade to more proficient graphic cards in both the MacBook and Pro is Snow Leopard, which will be big on parallel processing and offloading work to the graphics card—graphics cards are particularly adept at parallel processing because of the way they're designed and the fact that they have a buttload of cores. (Here's a more in-depth explanation of that.) And if graphics cards are driving more and more of the general computing experience, the truly shitty ones in the last generation of MacBooks just won't cut it.

Nvidia's been heavily investing in "General-purpose computing on Graphics Processing Units" (GPGPU)—again, using the graphics card for more general applications—on its own for a while, actually. When they demoed their latest, most badass cards for me a few months ago, it was heavily tilted on those types of applications, including in-game physics and Folding@Home. They actually have their own development kit called CUDA that lets programmers leverage graphics cards using a standard programming language—PhysX, a physics gaming engine, is probably the most well-known application of it so far. (Nvidia isn't sure when PhysX come to Mac, but they're looking at it.) Not so coincidentally, CUDA for Mac came out in August. These cards also support Apple's own graphics programming language, called OpenCL.

So even if you're the type of person that browses the net, edits Office docs and fiddles around in Photoshop rather than the type that plays WoW: Wrath of the Lich King or cuts video, graphics cards will matter to you almost as much as it does to those people: They're going to be critical not just in a lot of the awesome stuff you'll see coming out in the next couple of years but increasingly so in the way operating systems run, whether it's from Apple or Microsoft or anyone else. So get ready to hear a lot more about them.

Something you still wanna know? Send any questions about games, snow kitties or pancakes to tips@gizmodo.com, with "Giz Explains" in the subject line.

The specific models affected are MacBook Pros with the GeForce 8600M GT manufactured between May 2007 and uh, last month. You know you've got a crappy video card if you see screwy or scrambled video, or worse, no video, even if the computer is turned on. Apple will fix any Nvidia graphics card problem for free within two years of the original purchase date. If you've had to pay to get the problem fixed before, Apple will give you a refund.

The fact that Nvidia apparently misled Apple is deeply disturbing, especially as they counter the Inquirer's continuing reports about defective chips. If you've had video problems with MBP, raise your hand please.

Here's what Nvidia says for their side:

We've worked diligently with Apple, as we have done with all of our customers and partners, to analyze notebooks and determine the cause of such problems.

Our analysis shows that a failure in an Apple MacBook Pro notebook is remote. However, Apple, like other OEMs, decides on their own how to handle their warranty and repair programs, based upon their own quality standards.

Bottom line, we stand by our products, thus the reason why we set aside such a large reserve, and we have and will continue to work closely with Apple and their customers.

[Apple via AppleInsider]

Namely, Intel's upcoming Nehalem chips would require them to switch right back to Intel, since no one else has the license for their QuickPath interconnect. And the pros for moving to Nvidia (PC Perspective makes a good case why it would be them) are mostly about better graphics performance, ultimately. So this seems to fall in the unlikely column, for now. [Ars]

While there's no immediate way to tell if Nvidia is in fact lying (which would be a massively boneheaded move), as the Inquirer alleged, we'll know in the coming months. Reports of massive chip failures and an even larger hit to Nvidia's financials than the $150-$200 million it already took would make it super obvious. Hopefully the Inquirer was just being bombastic and British. [Ars Technica]

The root of the problem is the substrate/bumping material, which Nvidia themselves had characterized as "weak." The G84 and G86 chipsets use the same application-specific integrated circuit, which the Inquirer has been told had no changes made to it during its life cycle—besides, Nvidia wouldn't change the assembly process or materials for a single batch, according to engineers. So Nvidia's official claim, that the weak materials were only used for a single batch, doesn't hold so much water. And as the Inquirer points out, their fix, to run the fan more to offseat the heat issue simply covers up the problem while it drains your battery.

If this is all true, Nvidia better have more than $200 million set aside for fixes. [The Inquirer