AMD to Build G4 CPUs?

the eric conspiracy writes "Bloomberg news is reporting that Motorola and AMD are in talks to include manufacturing of Motorola CPUs at AMD's Dresden facility. This could help cash-strapped AMD particularly if its x86 compatible line runs into problems. Motorola and AMD already have cross-licensing agreements - AMD gets its copper technology from Motorola, while Motorola uses AMD's specialized RAM chip technologies. "

Comments

Re:AMD making Alpha chips?

[Christopher B. Brown]

There was a technology transfer that has recently become productized in the form of the EV7 "bus protocol" that DEC created for Alpha, and which Athlon now brings to IA-32.

It is entirely possible that the "talk" could have been a corrupted understanding of this transfer.

I'm sure AMD is happy enough to see some extra business come their way that isn't solely predicated on head-to-head battle with Intel.

It would be rather neat if this resulted in there being a third-party source for PPC motherboards, as that is a Critical Resource.

It looks like the AMD involvement hasn't led to cheap Alpha motherboards, which means that it's not time to replace my Multia/UDB yet; probably the same for you, too...

Strange..

[TurkishGeek]

I personally find this very strange. Both AMD and are suffering from capacity shortage problems. AMD's stock is at its pathetic levels because analysts slammed the company when they could not provide enough AMD K6's to the market. Just recently, Gateway announced that it would not be using AMD processors since AMD can not supply them in volume. How can AMD consider sharing its production capacity with Motorola at this critical time when Athlon is just out and in great demand? If they can not meet the demand for Athlon and provide enough Athlons to OEMS, Wall Street will KILL AMD this time. And unfortunately for us techies, technical excellence does not make a company successful by itself. AMD is having a lot of financial trouble, and I don't think taking the risk of not being able to produce enough Athlons is worth the money that Motorola will provide them in return for this favor..

Here is what I believe: There must be something else behind this, if it is true. Motorola must have made a really attractive offer for AMD to have taken this risk..

Re:Strange..

[Anonymous Coward]

Yup, that's about right. One thing that people aren't really aware of is that the Dresden fab is a)huge and b)empty. AMD had the cash to build it, but not fill it. This sort of arraingement is ideal because it gives them contracts that they can use to prove an intent to buy, thus allowing their long-suffering creditors to let them borrow more to pay Applied Materials et al for more fab hardware.

This is a win-win deal for AMD. They get to equip their fab with .18 micron copper hardware that they can certainly use later, they have an extant contract that they can use to stave off the vultures, and they get a small margine from manufacturing the chips.

When the Dresden plant is up to speed, they will start migrating K6-3 (their Celeron killer at the low end) over to copper at Dresden and phase out production in Austin, while the K6-2 will continue as an embedded CPU only while Austin is rebuilt to manufacture at .18 micron and with copper. AMD seems to think that everything that they make should go to copper and SOI ASAP, except for the stuff that must be as cheap as possible for the time being, so we will be seeing soon seriously fast K6-2s (for portables as well) followed by a full move to K6-3s on copper, at .18 micron for about half the cost of the K7s while they move to putting the cache on-die and expanding the size hugely while cutting the power use and working out the inevitable bugs.

AMD isn't being very cagey with their plans, either. I think that this is one reason why Intel is nervous. Yes, the K7 is good. But remember, the K6 was good as well. For that matter, the K5 kicked some serious ass. (And I guess that I am not too jaded because I recall how delighted I was to finally get a 66MHz 486!) These chips did not help AMD too much. What kept AMD in the status of also-ran? Remember, they have had a lot of excellent embedded products for years -- they are a well respected company in the market outside of x86 CPUs (and would bet better respected inside the x86 market if Jerry Sanders could keep his salesdroid mouth shut -- the bulk of the problems that AMD has had have to do with Sanders' selling the idea that AMD will ramp up better and faster than anyone ever before to the press; in reality, AMD's fabs have run better than most and the ramp-up has rarely been slow, Sanders had just worked everyone into a froth ahead of time so that there was no way to meet demand). What beat Intel was marketing and Intel managing to turn the Pentium into a brand. Now that AMD is learning to play the same game and people are seeing that there isn't much of a difference/AMD is competing on MHz which people look for, AMD is doing better. When they are at a decent volume, I think that the simple economics will push the majors (IBM, Compaq) to go heavily AMD, because this will be happening that the same time (end of this year, beginning of next) that the direct channel will be putting pressure on them to cut costs.

So, I think that AMD will do fine.

Wow -- sorry about the rambling. Back to coffee and work.

Re:silicon on insulator and bad old ISA

[SEE]

hmm everyone seems to say copper is good but I don't see any specs

Electricity passing through copper encouters less resistance and therefore produces less heat than electricity passing through aluminum (the current standard). This is why aluminum wiring in houses caused so many house fires and has been outlawed in the U.S. This is also why copper interconnects on a chip are better than aluminum -- less heat.

Less heat allows faster processors (ask any overclocker), but how much faster depends on specific factors of the implementation. Anything that says "copper processors are x% faster than the same design using aluminum" is BS unless accompanied by three pages of conditions and explanations.

Me, I want gold interconnects :-)

Re:G4 using IDE? Why?

[47Ronin]

I think it's a little narrow-minded to always stereotype Apple's customers as non-technical people with no knowledge of hardware. There is quite a percentage of Apple users that know their hardware/software and also understand the difference between SCSI and IDE. At my workplace we have the original rev.A beige G3/233 PowerMacs. They are externally connected to scanners, Syquest drives, external HDs, a CD-R, Jaz, and Zip drives. Internally there is an IDE connector. We put in a ProMax IDE card to access a 16 GB Ultra-ATA drive, and frankly it outperforms any of the SCSI devices by a factor of two. Now when people start prodding at Apple with some argument about multitasking and all that crap, I have to say that it's all trivial. The hardware never gets in the way and I have never encountered a situation where my productivity was compromised by some so-called lack of multitasking. For example, I have Photoshop rendering AppleScripted images while I write this. The nuances between the SCSI and IDE standard become less evident because Apple's implementation of these standards makes it very simple to interface. You can just as easily use an IDE drive as a SCSI drive. Plug it in and use it. If it's not formatted then the damn thing just pops a dialog asking if you want to format it. It's not rocket science. If you really need some kind of flexibility (partitioning, differing HFS or file formats,etc.) then just launch Drive Setup and do what you wish. I'd rather see the end of SCSI and move into an era of FireWire (no SCSI ID numbers, no cable length restrictions, hot-plugging enabled...)

-----
Linux user: if (nt == unstable) { switchTo.linux() }

Re:G4 Processors and Computing

[bjohnson]

The Power Computing case had nada, zip, zilch, zero to do with anything remotely related to copyright.

(sigh)

Power Computing (and the other clone manufacturers) lost their licenses because the CEO of Apple determined that if the company he was legally responsible for was to survive, the cloners' licenses had to go. The clones were in fact NOT increasing Apple's market share, which meant that they were poaching dollars from Apple, selling hardware cheap.

Apple makes its money from selling its hardware. The systems were theirs to license, or not. The systems are yours to buy, or not. Vote with your pocketbook.

Y'all rant about 'Free software' then piss and moan when the beer isn't free.

The fact that no one else is making PPC mobo's for your use is NO FAULT of Apple's; it's not exactly like Motorola would turn down other customers for their chips, they have Intel breathing down their necks.

Copper, aluminum, and gold

[overshoot]

Copper: there are three big benefits:

• Copper atoms are heavier than aluminum atoms and thus get kicked around less by high current densities (electromigration). Seriously! On-chip current densities are so high that a serious failure mode is from the wires flowing downstream.
• WRT resistance, in the last few years signals have been suffering quite a bit of slowing due to the resistive delay introduced by wiring resistance and capacitance. In GHz processors this actually becomes a limiting effect.
• Transient supply drops are so serious that they either degrade logic timing or in extreme cases flip flops. Lower supply resistance helps a lot.

As for wanting gold interconnects, no you don't. For one thing, copper is a better conductor than gold. Besides that, gold is a disaster in silicon processing (it diffuses like lightning and scavenges carriers. Low transconductance and high leakage everywhere = slow and hot.)

Re:GCC on PowerPC

[Malichus]

Apple is using EGCS/GCC 2.95 as its compiler for Mac OS X [client] (not OS X Server; that uses gcc "2.7.2.1" for now).

They've submitted a large quantity of code (mostly from the work done at NeXT) to the GCC maintainers, and work proceeds to integrate the two source bases.

Furthermore, Apple would be nothing short of braindead to release OS X [client] for G4 systems without using an AltiVec-aware vectorizing compiler to generate their code. Since GCC Is the compiler that they're using, it seems more than likely that they will expend considerable resources to making GCC's PowerPC codegen as good as possible. (And LinuxPPC will see the benefits, too. Very cool.) Perhaps they can even integrate some of their work from MrC/MrCpp, Apple's fantastically good optimizing PowerPC C and C++ compilers for MPW.

PCI64 not closed

[Geekholder]

Actually the 64 bit extension to PCI is part of the regular PCI spec. The fastest PCI bus is 64 bits wide running at 66 MHz. Some Wintel server machines implement this, as do many Sun SPARCstations. Other RISC workstations probably do too, I just don't keep up with them all.

You may be thinkng of PCI-X, which is an extension to PCI to up the clock rate to 133 MHz. PCI-X may require royalty payments, I'm not sure.

Intel intended to make NGIO an open spec. NGIO and FutureIO recently merged into a single proposal, and I'm not sure where they are going wrt licensing.