Now this just seems wierd. According to ALI's website, the ALi M1621 is the Aladdin Pro 2, which is packaged in a 476-ball BGA. The ALi M1631 is the Aladdin TNT 2, which is packaged in a 556-ball BGA. The M1621 Should not have built-in graphics. It would seem that something else is going on here, and it isn't obviously completely Asus's fault.
Intrestingly, both the 476-ball and 556-ball BGAs are available in a 35mm-square package, see e.g. http://www.spil.com.tw/pdf/pbga.pdf Both also appear to be available in a 31mm-square package. From what I can tell, the 556 is more common in the smaller package, but ALi at least admits that the M1631 is 35mm square. I can find no reference for this spec for M1621, but the 35mm 476 seems reasonably common. ALi's website wouldn't of course show us anything so useful as the ball pattern for the two chipsets, but I wouldn't be surprised if the M1621 and M1631 had compatible patterns, so that e.g. you could make a board that would use either. Perhaps ALi accidentally mislabled some M1631s as M1621s? Perhaps the M1631 will work just like a M1621 if you put it into a 476-ball package, so that they could stop making the M1621 but still sell it? Perhaps the M1631 will work like an M1621 even if it is in a 556-ball package but is jumpered externally in some way? It wouldn't be the first time a chip vendor played these kinds of games, cf. the 80486SX. Also, I agree with Eric that in most cases substitutions are reasonable; in many cases the board maker may run into an unexpected shortage and be stuck with a pile of PWBs and no source for the parts with which to populate them. If they can find other parts that will make them work according to spec (keeping in mind that, say, LinuxBIOS compatibility may not be in their specs :-) it should be OK. Still, you'd hope that at a minimum they would paint an annotaion on the final PCB to indicate a different rev, or at least change the number on a sticker. Most motherboards have barcodes and other stickers giving more specific production-run information; commonly it is on the outside of the outermost I/O slot, or sometimes on the bottom surface. Do your CUA boards have anything like this? In any event, none of this is meant to contradict your assertion that PC hardware sucks. :-) --Bob On Thu, Oct 04, 2001 at 04:57:53PM -0600, Eric W. Biederman wrote: > > Ronald G Minnich <[EMAIL PROTECTED]> writes: > > > On 4 Oct 2001, Eric W. Biederman wrote: > > > > > Is the northbridge compatible enough that you can setup the memory and > > > handle the differences in C code? > > > > It looks like it is mostly the same. We just now verified memtest. > > I'm not especially happy with part substitutions but if the parts > are compatible, I sounds like a reasonable policy. I'd love to see a > revision number increase. > > > > Ron have you looked at the printing on the motherboard and confirmed > > > the two boards claim to be the same model? > > > > the artwork is identical down to the rev #, traces, etc. > > > > I think there is a lot of old ASUS CUA stock with 1621 northbridges being > > resold. I have concluded one reason vendors don't like us to know much > > about mainboards is the truly amazing number of games they play with the > > hardware. > > That could be. > > Eric
