MOT's next move... MCF7457

[DaveGee]

Well I kinda shocked nobody posted this so here goes...

PDF: http://e-www.motorola.com/collateral/SNDF2002RECAP_Q1228.pdf

JPG: http://www.ocf.berkeley.edu/~ceugene/MPC7457.jpg

Check out this one slide from the PDF... Its shows the MCF7457 ETA 03 running at up to 1.8Ghz. Still the same slow bus speed tho... Then in 04 a 'proposed' version with a RapidIO bus.

Dave

 

[DakotaGuy]

Well it makes sense, as I am pretty sure that Apple will go with IBM's PPC 970 on the future PowerMacs, the eMac, iMac, Powerbook will probably stay G4. Over time Motorola will have to continue to improve and speed up the processor. This roadmap is probably pretty accurate. A 1.8Ghz eMac or iMac would be nice!

 

[cubist]

Note that it shows 7455 topping out at 1GHz.

Yet that is the CPU used in the dual 1.25's shipping today. Not to start the overclocking argument again...

 

[DaveGee]

Re: Note that it shows 7455 topping out at 1GHz.

Originally posted by cubist
Yet that is the CPU used in the dual 1.25's shipping today. Not to start the overclocking argument again...

Okay... maybe this will help end this overclock issue.

When a target speed is hit and chips come off the line one of 4 things will happen with a given chip.

1 - BAD CHIP goes on the junk pile (this is what drives up the price on 'good' chips.
2 - GOOD CHIP @ SPECd SPEED (okay thats fine)
3 - GOOD CHIP below SPECd SPEED (move that to the slower chip pile)
4 - GOOD CHIP above SPECd SPEED (this can happen)

Now in the past those 'better than spec' chips were just thrown into the pile with the rest of the 'just spec' chips but there is nothing from stopping MOT from putting those chips in an 'over spec' pile (only makes sence if there are a good number of over spec chips to begin with) and re-rate them as 1.25Ghz CPUs.

Is it overclocking? I dunno... I know it is when 'joe hacker' OCs a CPU it is since JH is just 'guessing' how far he/she can push a CPU but if MOT has the kind of equipment required to really test the chip to make sure it's okay running at that speed then I'm not sure I'd call it overclocking.

Its all how you look at it...

Dave

 

[rice_web]

I do believe that this roadmap could prove very accurate.

Those who have tried it have been able to get the 1.25GHz processors to 1.5GHz, and this from a G4 on 180nm manufacturing. A move to 130nm could easily yield 1.5GHz, and possibly more.

Now, if only we could see 1.5GHz G4s in January...

 

[cubist]

The roadmap shows 7457 in early 2003...

... at speeds up to 1.8 GHz. Are they actually going forward with this? Someone ought to know if the chips are available in volume. With the semi division of Moto up for sale (and noise about possible purchase by ST and Hitachi), things may be iffy... Do they make money on these chips?

 

[beatle888]

i cant remember where i saw it but i remember
reading that moto will continue with there chip
manufacturing for apple. i remember –don't
count us out yet– i also remember something
about apple using the IBM chip AND the moto
chip. i dont know...we'll see soon im sure.

 

[MacCoaster]

Re: Note that it shows 7455 topping out at 1GHz.

Originally posted by cubist
Yet that is the CPU used in the dual 1.25's shipping today. Not to start the overclocking argument again...

Didn't you see that chip image of the 1.25GHz one? It's a 7455A, not a 7457.

It could be that 7455A == 7457, but I dunno.

 

[DaveGee]

Re: Re: Note that it shows 7455 topping out at 1GHz.

Originally posted by MacCoaster

Didn't you see that chip image of the 1.25GHz one? It's a 7455A, not a 7457.

It could be that 7455A == 7457, but I dunno.

That would mean that the 1.25 was being made with a .13 process READ A LOT LESS HEAT and from what I've seen the 1.25 isn't any smaller nor running cooler...

Like I said the 1.25 is just a 1.0 running better than it should...

Dave

 

[DakotaGuy]

Originally posted by beatle888
i cant remember where i saw it but i remember
reading that moto will continue with there chip
manufacturing for apple. i remember –don't
count us out yet– i also remember something
about apple using the IBM chip AND the moto
chip. i dont know...we'll see soon im sure.

Well yeah I would think that Apple will continue to keep the i and e consumer lines on the 745x series of processors when and if the PowerMac goes to the PPC 970, so the chip will continue to develop much like the PPC G3 750 has since it left service in the B&W G3 towers.

 

[Anon]

Re: Re: Re: Note that it shows 7455 topping out at 1GHz.

Originally posted by DaveGee


That would mean that the 1.25 was being made with a .13 process READ A LOT LESS HEAT and from what I've seen the 1.25 isn't any smaller nor running cooler...

Like I said the 1.25 is just a 1.0 running better than it should...

Dave

Usually, after they have made the chip they go through it and look for bottlenecks and look for places where they can optimise it and get some speed increases. This is usually where you get the incremental speed increases on the same process. On the other hand Motorola could have pumped the bus to 166mhz and found the chip to be stable for a large enough number or chips to ship it. Either way, as long as Motorola is will to stake its reputation and stamp on the chip itself that it is a 1.25ghz chip, I don't consider it overclocking. But then who cares what I think.

 

[scem0]

This sure doesn't go along with moto selling their semiconductor unit...

 

[Megaquad]

Cool, now we know for sure that Apple will never use IBM's high end cpu's in any future PowerMac.
Why? They cannot implement AltiVec into Power4 processors! OS X would be slow on those machines because it is optimized for AltiVec, Dual G4's etc. Apple's programmers would have to do everything from beggining and consumers would suffer. No more AltiVec enhanced scrolling, MPEG-2 encoding, real-time effects with FCP, fast iMovie rendering, iTunes...
Not to mention other pissed off 3rd party developers (especially Adobe)! So shut up about IBM chips, its not going to happen! Macs would become overally slower then with G4's.

As for revolution (aka new powermac which is 2x faster then anything before). Well, its not going to happen either. Marketing doesn't allow it, overall manufacturing costs dont allow it and many other things dont allow it. It is always going to be like this.. So, 20% speed bump is best we will ever get, which is ok with me.

 

[barkmonster]

Cool, now we know for sure that Apple will never use IBM's high end cpu's in any future PowerMac.
Why? They cannot implement AltiVec into Power4 processors! OS X would be slow on those machines because it is optimized for AltiVec, Dual G4's etc. Apple's programmers would have to do everything from beggining and consumers would suffer. No more AltiVec enhanced scrolling, MPEG-2 encoding, real-time effects with FCP, fast iMovie rendering, iTunes...
Not to mention other pissed off 3rd party developers (especially Adobe)! So shut up about IBM chips, its not going to happen! Macs would become overally slower then with G4's.

whoever said apple were going to use high end IBM cpus ?

The GPUL is a desktop cpu like the G4 is, it's just got features from the Power4 that would make it as powerful as a highend cpu, especially in a dual config with it's 6.4Gb/s of bandwidth and 2 SiMD (Altivec/VMX/Velocity Engine) units per cpu.

Silicon Strategies Article

I'm sure you've read this already but here's a little more info on the PPC* 970 :

Essentially a derivative of the company's Power4 microprocessor, IBM's PowerPC 970 adds 64-bit PowerPC compatibility, an implementation of the Altivec multimedia instruction-set extensions and a fast processor bus supporting up to 16-way symmetric multiprocessing.

Here's more info :

IBM's approach to implementing a 32/64-bit architecture appears straightforward. The 970 supports full 64-bit registers and addressing. When a flag bit is sent it triggers a 32-bit mode in which the high-order words on an arithmetic logic unit and on memory addresses are ignored. In either 64- or 32-bit mode, the processor issues up to eight instructions per clock cycle.

Thus, 32-bit PowerPC applications run unchanged on the 970. However, a 32-bit operating system would have to support new data structures in its memory management unit and new interrupt handlers.

and this pretty much explains why the IBM chip is very real possibility :

Besides the 64-bit capability, the PowerPC 970 includes IBM's first support for what Motorola calls the Altivec instruction set and Apple refers to as its Velocity engine. IBM simply cites the capability of the 160 vector instructions in its SIMD engine to speed graphics and multimedia operations. IBM had opted not to support the instruction extensions in its 32-bit PowerPC family but will bring them into the new 64-bit line. "Adobe has been the premier company taking advantage of [the Altivec extensions]," said Krewell of Microprocessor Report.

It has got Altivec (just a motorola name for the 128bit SiMD unit anyway) it's got 2 of them aswell. 1 PowerPC cpu with 2 Altivec units could be like a dual G4 on 1 chip when running certain tasks like audio/video compression, software synthesis and those worthless RC5 and SETi crunch tests people like to refer to when comparing Mac and windows PC performance.

The 970 also sports a cache-coherent, 900-MHz processor bus capable of data rates up to 6.4 Gbytes/second. It will support symmetric-multiprocessing configurations of up to 16 CPUs. That capability could be valuable for Apple's new line of Xserve entry-level servers.

6.4Gb/s memory bandwidth, The G4 can only handle a measly 1.3Gb/s not even enough to fully saturate the 1 altivec unit the PPC 74xx series has.

So I think we'll see Motorola chips at 1.5Ghz in the next mac line up and then when the GPUL is ready, apple will move to the IBM chip.

* PPC = PowerPC as in 601,603e,604,604e,750,7400,7410,7450,7455 or G1,G2,G3 and G4 if you prefer. IT ISN'T A POWER4

 

[technocoy]

the powerPC 970

I did a post about this on friday...
the powerPC 970 will be in future macs... trust me

technocoy

 

[technocoy]

one more thing

someone, in all these conversations, has failed to mention what seems to be obvious.... I have the new dual 867 @ work and it runs hot as hell, has a huge heat sink, and the fan sounds like a jumbo jet ready for take-off, if the 1.25 is an overclocked 1.0 (which it isn't since the manufacturer doesn't overclock) then going by all the "logic" around here, the dual 867 must be an overclocked 1.0 as well!!! ... c'mon people!

technocoy

 

[scem0]

Originally posted by Megaquad
Cool, now we know for sure that Apple will never use IBM's high end cpu's in any future PowerMac.
Why? They cannot implement AltiVec into Power4 processors! OS X would be slow on those machines because it is optimized for AltiVec, Dual G4's etc. Apple's programmers would have to do everything from beggining and consumers would suffer. No more AltiVec enhanced scrolling, MPEG-2 encoding, real-time effects with FCP, fast iMovie rendering, iTunes...
Not to mention other pissed off 3rd party developers (especially Adobe)! So shut up about IBM chips, its not going to happen! Macs would become overally slower then with G4's.

As for revolution (aka new powermac which is 2x faster then anything before). Well, its not going to happen either. Marketing doesn't allow it, overall manufacturing costs dont allow it and many other things dont allow it. It is always going to be like this.. So, 20% speed bump is best we will ever get, which is ok with me.

We might see a high end IBM processor that has altivec support... Just because there are none out now doesn't mean there will never be any out. And even without AltiVec on a fast system, the OS will be fast - altivec or not. Not as fast as it would be w/ altivec, but fast enough. Adobe might be pissed, but who cares. They can deal.

As for no 2x speed increases - that is not true at all. I could totally see that happening. Why would marketing not allow that? Wouldn't they welcome it? If apple figured out some semiconductor manufacturing strategy right now, and started being able to produce 2.8 GHz g4's at half the cost of the current ones, why the hell wouldn't apple be able to sell it??? Im not saying that will happen, but technology is unpredictable... Be ready for anything.

 

[ddtlm]

A 130nm G4 is the obvious thing for Moto to do... and I am very much not surprised that they will keep the FSB the same. This sounds good. Should work well in laptops, iMacs, eMacs and even in lower-end PM's. Will also mean that upgrade card makers have access to faster G4's for some time. Woo hoo, my dual 800 will live on!

 

[Catfish_Man]

.13 micron G4+...

...should have frickin awesome power dissipation (G3ish). 2 cents and a used jawbreaker says the iBook goes G4 when the 7447(low power 7457) comes out! (Then, later, the TiBook can use the low voltage version of the 970, the PowerMac can have the high voltage 970, and the iMac and eMac can have 7457s).

 

[ddtlm]

Catfish_Man:

With regards to the PBook, the way I see it Apple needs to make a choice: PC-laptop-like performance with the PPC-970 and PC-laptop-like wieght and heat output, or slim design and lower performance, as we have now.

Summary: I really don' think a fast 130nm PPC-970 will work in a 1" thick laptop frame. Maybe a low-volt, low-clock version... much much slower than the PMacs. Also note that laptops basically make feeding the PPC-970 FSB impossible... a dual-channel DDR chipset is just not very practical for a laptop. Especially not a 1" thick laptop.

 

[Beigean]

barkmonster:

according to motorola the G4 has four altivec units

7455 info

now, motorola does have different ways of counting things than other folks. my guess is that the 970's implementation of Altivec is on par with that of the 7455, hence the linear scaling of RC5 performance.

don't expect the 970 to have double the altivec performance of a G4

 

[ddtlm]

Beigean:

There is currently a lot of uncertainty about the subject, because people often refer to all four AltiVec funtional units as one AltiVec unit. I think the orginal G4 has only two AltiVec functional units, BTW.

 

[Switcher2001]

Processor speed vs. Application speed

Hi. This is my first post here, although I've been reading the forums every day for the past six months or so... ;-) For some reason, I just felt roiled up enough to register as a member and say something about this issue.

I'm really getting frustrated Apple's deceleration of processor speed increases and their tendency, lately, to come out with dreadfully slow applications. I switched to an iMac (Sage) at the beginning of 2001 because I was sick of Microsoft as a company and I was fed up with Windows crashing all the time. At the time, I was happy with the processor speed, and Apple seemed to be keeping up with the Intel, AMD, etc. processors just fine. I also got my iMac because I was looking forward to OS X.

But lately, it's amazing to me how far behind Apple is lagging in processor speed. And what is it with the latest iApps? iCalendar is the slowest application I have ever used on any platform in my life. It nearly choked when I imported three year's worth of events from Palm Desktop. PD zipped along as happy as a honeybee with three years of calendar events (it's ugly compared to iCal, and it doesn't use the iCalendar standard, and it's not a Cocoa app, but at least it's fast). I finally ended up spending about five hours going through all my calendar events from 2000 and 2001 and saving them to their own calendars. Once I archived 2000 and 2001, I saw a dramatic speed increase in iCal, but it's still a slow application even with just a year's worth of events to carry.

I just don't understand what in the world could take an application such a long time to perform the simplest tasks. How can Apple afford to make the slowest applications ever at a time when their processors are the slowest in the business? Why did it take Apple over a year to accelerate OS X's Finder so that it didn't take a full three seconds to move from file to file in icon view using the arrow keys? It just boggles my mind.

Believe it or not from the way I'm ranting, I'm an avid Mac fan. I have no desire to return to "my clunky Windows world," as one of the Switchers says. It burns me, though, when I'm using Word X on my 600 MHz iBook with 384 MB RAM and it's three or four times SLOWER than Word 97 ever was on my old Pentium 125 MHz machine with 128 MB RAM! I mean, why bother having all this processor speed and RAM when applications are slower than ever? I just don't get it. I love it that my Mac hardly ever crashes, though... and it's gorgeous... and intuitive... etcetera... but, IMHO, Apple—and all other software developers—should focus more on application speed than processor speed.

 

[voyagerd]

what's the PM under the L3 cache stand for?

 

[MacTech]

Re: Processor speed vs. Application speed

Originally posted by Switcher2001

Believe it or not from the way I'm ranting, It burns me, though, when I'm using Word X on my 600 MHz iBook with 384 MB RAM and it's three or four times SLOWER than Word 97 ever was on my old Pentium 125 MHz machine with 128 MB RAM!

That sure is one hell of a RANT. I have an iBook and several Windows boxen. No way is what you say above true on my hardware or any of the ones I admin. Three to four times slower, I highly doubt that it is even slower than a P3 750.