Cooling issues?!?
With some recent rumors claiming PowerMac G5 delays are based on cooling issues, could that even be close to correct? Offical notes form IBM have put the 90nm part running one hell of a lot cooler than the 130nm part. And a theroetical 3.0GHz 970FX would run about as hot as a 2.0GHz 970 (according to numbers run here at the AI forums). So are rumor sites grasping at straws, or could there potentially be something to all of this?
MOSR just jumpped on this bandwagon of cooling issues, claming the next PowerMac update will require liquid cooling... Bunk I say!
MOSR just jumpped on this bandwagon of cooling issues, claming the next PowerMac update will require liquid cooling... Bunk I say!
Comments
As for the rumor: The 970fx runs cooler at equivalent clockspeed, but if you crank it up as far as Apple cranked up the 970 then you're back in the same neighborhood, heat-wise (although the 970fx might be a bit cooler). In addition, if Apple is altering the case, as some have speculated, to allow for more space for internal drives, then there's that much less room in the CPU cooling zone.
Additionally, depending on exactly what kind of drives they're planning on putting in there, the cooling issues could be at the top of the case. Assuming two optical drives and three hard drives, imagine two SuperDrives burning away while something else streams a few gig of data off a three-disk RAID - there's a cooling nightmare right there.
Originally posted by kupan787
With some recent rumors claiming PowerMac G5 delays are based on cooling issues, could that even be close to correct? Offical notes form IBM have put the 90nm part running one hell of a lot cooler than the 130nm part. And a theroetical 3.0GHz 970FX would run about as hot as a 2.0GHz 970 (according to numbers run here at the AI forums). So are rumor sites grasping at straws, or could there potentially be something to all of this?
MOSR just jumpped on this bandwagon of cooling issues, claming the next PowerMac update will require liquid cooling... Bunk I say!
I would be more inclined to think that PM updates are waiting on debugging the latest rev of OS X and supply of chips from IBM. Earlier builds of 10.3.3 were causing some serious problems with 20" iMacs if I remember the reports correctly, and ironing out these bugs could be a realistic reason for a delay.
Also the delay of the Xserve suggests that either they have some serious backlog of unofficial pre-orders to government, Pixar and Virginia Tech; IBM is not able to build a large enough supply for initial Xserve release; or there is some other hardware issue that crept up that is hurting the delivery of these computers. Given that the original mother board has had some success, and the Xserve board is probably based on this I doubt that it is holding up the release. Software or chip issues sound most likely.
Originally posted by Amorph
As for the rumor: The 970fx runs cooler at equivalent clockspeed, but if you crank it up as far as Apple cranked up the 970 then you're back in the same neighborhood, heat-wise (although the 970fx might be a bit cooler). In addition, if Apple is altering the case, as some have speculated, to allow for more space for internal drives, then there's that much less room in the CPU cooling zone.
I wonder if it isn't actually a combination of the rumors that is the real story. Rumors have revolved around cooling issues and power supply issues. Possibly, the new power supplies that work with the power-draw management issues of the 970fx are having issues with cooling. Maybe, there is a confusion in rumor sources that the power supplies needed for working with the cooling function of the 970fx are having problems?
Who knows. But, it may be possible that the seemingly differing rumors may actually be one and the same and is the true reason...