Charlie at it Again: 9400M Problematic?

http://www.theinquirer.net/gb/inquirer/news/2008/11/29/nvidia-heat-causing-macbooks

Well, The Inquirer is running another story on problematic nVidia GPUs and this time they implicated the 9400M. Supposedly, the 9400M is running hot and using more power than it's supposed to which may lead to problems down the line.

Their proof is related to insider questioning and the delays associated with getting the MCP79 out the door, including the fact that it launched on B2 stepping. Now there's no way to know the validity of insider talk, but it is a fact that the MCP79 has been horribly delayed since it was supposed to ship before Montevina, was rumoured to be cancelled, then didn't show up again until Jobs announced it. Similarly, it is true that most GPU and chipsets launch on A1 steppings so if they had to launch at B2, that would indicate a lot of issues required many design iterations to fix. There are also a strange lack of desktop 9300 and 9400 motherboards so OEMs do seem to be holding back.

It should be noted that the B-stepping may not necessarily be related to failure of design targets. Jobs said that the MCP79 was originally targeted at desktops and Apple asked nVidia to redesign the part for mobile use. It's quite possible the A-steppings were the original desktop implementation which was scrapped and they moved to B-stepping for the mobile design. If they started from a desktop design rather than from scratch, it isn't surprising if the MCP79 is more power and heat hungry than would be preferred for a mobile part, although you do get the performance benefit which everyone was complaining about from Intel IGPs.

This isn't quite the same as the problems with previous G8x generation parts, since even Charlie notes that nVidia has been cherry-picking parts for Apple so the parts in Apple products meet agreed-upon design guidelines. As such, there shouldn't be a grave concern of failure. No doubt, Jobs has extracted strict promises from nVidia that there won't be further problems. (Probably threatening to never buy from them again if there are). And Charlie reports that there is a B3 stepping coming in January to further reduce heat, which conveniently comes just in time for a unibody 17" MBP and iMac refreshes.

 

Well, considering nvidia has just been trying to ignore the problem all this time, sweep it under the rug, without any real indication they actually corrected the problem, none of this comes as any surprise to me. I actually take what the inquirer says with a grain of salt, however I think there could be some validity to their arguments. How much time has nvidia had to fix this, and apparently hasn't done anything? Of course it doesn't mean that every nvidia based system will fail, just that more than the usual are failing, and to me that's not a chance I'm willing to take 2 or more years down the road, no thanks.
Also I'd like to point out that I believe Charlie is wrong, as far as I know, the 9400M is actually a rebadged 9300M exclusively for apple, which is a mobile chip.

 

Well in this case, The Inquirer isn't reporting any packaging material weaknesses as with the previous G8x series. Rather, the 9400M on average isn't running as cool as the mobile part was originally envisioned to be, while the parts that Apple is getting supposedly are specially picked to meet spec. It's probably more of a customer inconvenience factor with the unit being a bit hotter than it could have been rather than a direct failure issue.

And of course, it's an issue for nVidia since after Apple takes all the good ones, there probably aren't any left for other laptop makers and nVidia is left with all the hot IGPs that nobody wants, not even desktop mobo makers given the few 9300/9400 desktop motherboards.

 

So according to the article, anyone thinking about buying a unibody MB or MBP (MBA ?) right now should wait until late January when the B3 stepping shows up in those models.

 

Its just charlie rambling on about nothing.