Motorola's New Chip

[arn]

MacCentral confirms (http://maccentral.macworld.com/news/0201/29.motorola.php) that the new PPC in the 1-GHz PowerMac is indeed an Apollo processor (MPC 7455, MPC 7445) with use of SOI and 2MB DDR L3 Cache.

A key quote regarding the new chip is that it is a first installment of the chip and "The big news here is that this device is certainly not topped out" according to a Motorola rep.

[ftaok]

For everyone who keeps bashing MOT. This is the kind of technical partner that Apple needs.

Motorola's news release about the Apollo yesterday contained mention of Apollo design wins within the networking, mobile communications and embedded markets, but no mention of its win with Apple. Swearnigen said that it was essential that Motorola did not pre-announce Apple. "In working with them [Apple], we had to be sure that we didn't steal Apple's thunder. When a customer is as strategic and important as Apple, we don't mind helping their publicity efforts," said Swearingen.

It's a wonder why Motorola has stayed so loyal to Apple when Apple bashes them over the head at every opportunity. I guess it's the same reason why women stay with guys that treat them like dirt. But that's another discussion.

[Pants]

Originally posted by ftaok
For everyone who keeps bashing MOT. This is the kind of technical partner that Apple needs.



It's a wonder why Motorola has stayed so loyal to Apple when Apple bashes them over the head at every opportunity. I guess it's the same reason why women stay with guys that treat them like dirt. But that's another discussion.


Its because, using your analogy, Motorola is an alcoholic mother who mistreats her kids by refusing to feed them unless regularly 'woken up' by a slightly more caring husband.

Im concerned thats its not 'topped out'. Theres potential here - instead of being first to push 64 bit 'consumer' desktops , motorola is busy building potential to be last to market.

[OSeXy!]

Thanks for that, Arn. That was what I was trying to find out. Excuse my optimism, but it seems to me that Apple's (short term) future looks very bright.

[Falleron]

Are people saying that there is likely to be speed bumped G4's soon? Say march??

[Unregistered]

damn, that company is pretty too much expensive...
the pc company offer us some great price... so i guess Apple could do it too!!! that's not because it's an Apple, they just don't want to!!!

anyway, i'm pretty attracted by the 933, but again... just take a look at the price! and what about Dual 1Ghz.... just forget it!

[samy85114]

damn, that company is pretty too much expensive...
the pc company offer us some great price... so i guess Apple could do it too!!! that's not because it's an Apple, they just don't want to!!!

anyway, i'm pretty attracted by the 933, but again... just take a look at the price! and what about Dual 1Ghz.... just forget it!

[Six]

There are like 10 other forums of people complaining about the price of the G4's and stuff, go there and whine about it, i have no problem with their pricing...

Back on the subject.. ;) Now that we know the new G4's are not topped out, that seems to me that its a pretty good indication that we arent gonna see G5 too soon. Which is fine..

IMHO- Keep upgrading the G4's for now, and keep working hard to develop the G5, and totally wow us when they come out.. Although Im certainly wowed now with the new G4s :)

[Xapplimatic]

Well, I think that truthfully, Apple and Motorola have a pretty much symbiotic relationship. The public chiding over processor delays and expectations and all that is just friendly warring. Motorola R&D does what it can probably about as fast as it can given their constant slashing of positions in their semiconductor divisions. Apple as I understand it is their only major computer customer for PPC chips, and the rest only go into imbedded devices which admittedly is a big market.. but during stormy financial times, Motorola can't really complain that Apple is demanding on them because Apple is also a fairly stable customer for Motorola who guarantees a certain portion of Motorola's revenues. Apple is Motorola's rock, but I'm not so sure it works out that way on Apple's side of the equation.

As for the new G4s, they rock! Appollo G4s may be bumped up to ~2 GHz before they're maxed. They already have 2x the number of addressing boxes or something like that which allows them to address more. They are a much more finely optimized G4 apparently. It will probably be the last revision to G4s while they finish ironing out the G5s. Certainly this G4 paved the way for the G5 and provides a good test medium as the G5s will all be based on the same manufacturing process (HiPerMOS) and SOI (Silicon on Insulator) fabrication technologies. SOI incidently makes it easier for Motorola to put out low-power chips for laptops to market faster as they run cooler and at lower voltages from the get-go.:)

[Unregistered]

Does this all mean that the 933 is also an Apollo chip?

[spikey]

"Eliminating data bottlenecks
Since much of the data being processed is in dedicated L3 cache, there’s less data traveling from the processor to main memory at any given time. This in turn means there’s less data traffic in the system controller, and a lower instance of bottlenecks caused by multiple data streams contending for the system bus. This leaves more bandwidth for dealing with data from other subsystems, such as AGP graphics, PCI or the I/O controller."
From the apple site.

Is this just ************ from the apple site to detract from no DDR RAM???

[dongmin]

Yeah Moto should be commended for getting the Apollos to market, finally. But this is the first bit of good news from Moto in a while. They've really been dropping the ball the last few years.

Switching over to the Apollo now probably means Apple won't move to the G5 for at least another year. MWSF 2003 is probably a good bet. This also means we won't see any big new technologies (like DDR RAM, Hyper transport, etc.) coming to the powermac line anytime soon.

I'm skeptical of the claims in the YourMacDaily.com rumor that the PowerPC will pass the Pentiums by the end of 2003. They're already at like 2.4 Ghz, right? By the end of 2003, they should be pushing 4+ ghz. I don't see the PowerPC making that kind of jump.

[spikey]

I dont see it possible to catch up with intel but with the Athlon it is possible.
It isnt moving very fast really.

Your theory on the G4 powermac is very valid, but i think the reason you give is wrong. Apple wouldnt release a G4 powermac and max it out before they give it to the imac. In my opinion they would let the appollo max out in the imac.

[theranch]

Why is it that Apple is always playing catch up in the processor speed ? I agree with dogmin when he said:
"I'm skeptical of the claims in the YourMacDaily.com rumor that the PowerPC will pass the Pentiums by the end of 2003. They're already at like 2.4 Ghz, right? By the end of 2003, they should be pushing 4+ ghz. I don't see the PowerPC making that kind of jump."

That statement to me just reiterates the same thing that has been going on....Motorola will be playing catch up forever. Right now when I view a flash site on a mac it runs slow but on a peecee it is fast...the same with gaming. Why should I shell out a ton of money for a mac when I can spend half and get a faster machine with a monitor. I know..the OS is more stable on the mac and the looks are great as well as the ease of use but the cost is killing Apple sales. People in general aren't convinced by the speed story that Apple is feeding us.
I love my mac but the speed jumps aren't there yet. I do like the lower price released yesterday.

[eric_n_dfw]

Originally posted by theranch
Right now when I view a flash site on a mac it runs slow but on a peecee it is fast...the same with gaming.
I suspect that the Mac version of Flash is poorly written. Look at other apps that run plenty fast and do a lot more like Final Cut Pro or iMovie.

I would not be at all surprised if Flash is running in some kind of emulated PC mode liek MS Office did a few years back. Windows core code compiled on top of some kind of Win32/Carbon shim or something.

Just a guess.

[dongmin]

I don't have as much problem with flash running slow as I do with internet explorer or netscape. Even with a G4 800 (I tried out the new iMacs today), it seems sluggish to me. Apple needs to do something about that. They need to throw some serious support behind a browser developer--I don't care who--and make it fast. It's probably the application that the average user spends the most time with.

[evanmarx]

It is important to note that this is only the first installment in this processor family. "The big news here is that this device is certainly not topped out," said Swearingen. Mac users can reasonably expect incremental performance increases from this processor family in the future, taking Apollo to its promised GHz+ performance levels. Swearingen could not comment on when these increments will come or what size they will be. The Apollo G4s are available now at speeds between 600MHz and 1GHz.

looks to me as if it's not so sure that the G5's are underway very soon ... they could still go with the g4 until fall, and only then (in winter) introduce the G5 as a top top end ... "vertainly not tooped out" ...hmmmm

[G4scott]

About the price, it is probably the cheapest power mac to date. the mid range is only 2299, compared to about 2500-2600 a year ago. also, dual ghz processors for under 3000, what do you want? a quad processor G5 with a 4x bus speed, dual processor graphics cards and 16x agp for under $1000? If you want a lower priced dual processor, find one of the dual 800's and quit your bitchin...

[ThlayliTheFierce]

If you read the article, they say that the Apollo chips run at 600-1000 Mhz. I think 600-700 Mhz Tibooks are in the not-too-distant future.

[me hate windows]

i am quite happy now that prices have gone down on G4s.
100 on basic
300 on middle
500 on best
and now there is a custom built super one that is only 150 more than the old top-ends:D :p

[Too X and beyond]

I wonder if the Apollo's are in the imacs. This would make sence and help with heat in the small inclosures.

I still hope and think G5's in N.Y. with the switch to imac G4's. Yes the Apollos have head room but so do the imacs. The G5's are timed for the low end release at 1Ghz.

I hope I'm not wrong/disapointed. I've been fallowing Apple long enough to know that I need to check my enthusasm. But, hey they still suprise me time and again.


--------- All my life I've wanted to be somebody. Now I know I should have been more specific. -------------
-Lillie Tomlan (R.I.P.)


-tink

[Xapplimatic]

Originally posted by spikey
I dont see it possible to catch up with intel but with the Athlon it is possible.

Ever used Athlons side by side to P4s? Actually, AMD chips are--at lower MHz --faster than P4s.. so back to the MHz Myth...

[lordsinforge]

Originally posted by spikey
"Eliminating data bottlenecks
Since much of the data being processed is in dedicated L3 cache, there?s less data traveling from the processor to main memory at any given time. This in turn means there?s less data traffic in the system controller, and a lower instance of bottlenecks caused by multiple data streams contending for the system bus. This leaves more bandwidth for dealing with data from other subsystems, such as AGP graphics, PCI or the I/O controller."
From the apple site.

Is this just ************ from the apple site to detract from no DDR RAM???


Negitive. the thing that slows down computation is going to the main memory for data, due to the fact that it is so "far" from the chip. It's like this (an example only I have no idea what the real times are):

time is in milliseconds

on chip cache ==0 time
Level 2 cache ==10 time
level 3 cache == 20 time
main memory == 100 time

[AlphaTech]

Originally posted by Xapplimatic


Ever used Athlons side by side to P4s? Actually, AMD chips are--at lower MHz --faster than P4s.. so back to the MHz Myth...

I'll testify to that fact...

[TennotX]

Originally posted by spikey
I dont see it possible to catch up with intel but with the Athlon it is possible.
It isnt moving very fast really.

Your theory on the G4 powermac is very valid, but i think the reason you give is wrong. Apple wouldnt release a G4 powermac and max it out before they give it to the imac. In my opinion they would let the appollo max out in the imac.

I think the above is not Motorola's biggest tech problem, but Apple's biggest marketing problem. Let me explain. From what I've read, which could be wrong, Mhz is a relative term that varies from processor to processor. Say, a G4 Mhz means something different from an Intel Mhz. It may even be true that a G3 Mhz is different from a G4 Mhz.

This makes sense for a couple of reasons. Number one, Mhz means millions of cycles per second (or cycles, I'm not sure which). So what is one cycle? Could it not vary. Again, this reasoning could be off, and correct it if it is, but one cycle of a bike wheel is a lot different from one cycle of a ferris wheel. The second reason, as I see it, is that the processor is not purely Mhz driven, but other features, like the fact that the Velocity Engine can process 128-bit data chunks or 4 32-bit data chunks at a time, should factor in. Right?

That's why Motorola chips do so much better than comparative Intel and AMD chips. Motorola is foolish and set its Mhz standard relatively high. To change now would be utterly confusing, and perhaps lead to law suits for misleading consumers. This means that Apple has to market the 1 Ghz chip when Intel is marketing a 2.4 Ghz chip, although the motorola's 1 Ghz might be comparable to perhaps a 2 Ghz Intel. Mind you, this still leaves Motorola significantly behind, but not so far behind that it can't catch up. If this new manufacturing process allows Motorola to make faster chips more consistently, the true power of the highest-end Motorola PPC could exceed that of the Intel and AMD processors soon.

---------------
IBM: I Buy Macs.
Microsoft: Software for users with micro-budgets, micro-needs, and micro-brains.
Dell: Sorry, I always get the "D" and "H" keys confused with each other.
Gateway: It is the gateway, to doom.
HP: <b>How</b> about a <b>Printer</b> instead?

[gbojim]

That's a really cool analogy - comparing the CPU to a bicycle.

Moving the pedals through one cycle is pretty much the same for all bicycles since they are mostly the same size. That could be directly related to the clock speed since a MHz is a MHz no matter who the chip maker is.

On the other hand, depending on the gearing, you can cover more or less distance, meaning you can complete more or less work. So you could say that the PPC, Intel and AMD chips start off with the same MHz base, but they are geared differently.

Thanks for thinking of that. I'll have to use it in my next hardware class.

BTW - anyone else wondering how Intel is going to explain how the Itanium gets so much more work done than Pentium while it is running at about half the clock speed?

[Catfish_Man]

...is an excellent way of explaining. The actual explanation is slightly more complicated (more along the lines of different gears for different tasks, so one chip will win at one thing and another at another thing), but that's basically it. Right now though, G4s only win at Altivec heavy tasks. Compared to a P4/AthlonXP their integer/floating point performance is not that great. Still, Altivec *flattens* SSE, SSE2, MMX, enhanced 3dNOW!, and 3dNOW! (the x86 equivalents to it).

[Onyxx]

they could say that the itanium processes faster because the die is phisically bigger!!! hey a new consumer scale of *judging* which processor is the fastest. Who needs Mhz (Ghz) when we have die size!!