Welcome to the MacNN Forums.

Eug · 04:42 PM

Came across some interesting info at Ars. See this Motorola 7447 PDF document.

They specifically mention that it's a chip used for laptops and elsewhere in the same document then say it's for Apple.

Part numbers:

MC7447RX1333WB - 1.3 V (runs at 1250 or 1333 MHz)
MC7447RX1000WB - 1.2 V (runs at 1000 MHz)
MC7447RX1333YB - 1.32 V (runs at 1333 MHz)

The weird part is that it specifically states it's a 0.18 um (SOI) part, NOT a 0.13 um part.

EDIT:

The link no longer works. It seems the document has now been removed from the website. Hmm...

C-Bear · Sep 17, 2003, 04:31 PM

Originally posted by Eug:
The weird part is that it specifically states it's a 0.18 um (SOI) part, NOT a 0.13 um part.

Eug, please explain the significance of this "weird part?"

thunderous_funker · Sep 17, 2003, 04:35 PM

"The MPC7447 is manufactured on Motorola's 0.13-micron HiPerMOS silicon-on-insulator (SOI) copper interconnect process technology, enabling it to deliver superior performance over bulk CMOS technology."

Link

But are you sure?

The Register reported the new machines are using the 7457 chip. Wouldn't the 7447 be older than the 7455 that the machines were already using??

Eug · Sep 17, 2003, 04:39 PM

Originally posted by C-Bear:
Eug, please explain the significance of this "weird part?"

I'm not a CPU engineer, but here are 2 pennies...

The 7447/7457 is supposed to be Motorola's transition chip to 0.13 um, to take the G4 up to 1.8 GHz.

However, assuming it's not a typo, they've chosen to use 0.18 um, which makes me wonder if they're having problems transitioning to the 0.13 um process. This would be worrisome, because the G4 on 0.18 um isn't really meant to go higher than where it is now. ie. The current CPU is already at the end of the line in terms of speed.

Eug · 04:50 PM

Originally posted by thunderous_funker:
"The MPC7447 is manufactured on Motorola's 0.13-micron HiPerMOS silicon-on-insulator (SOI) copper interconnect process technology, enabling it to deliver superior performance over bulk CMOS technology."

Link

But are you sure?

The Register reported the new machines are using the 7457 chip. Wouldn't the 7447 be older than the 7455 that the machines were already using??

No I'm not sure. I was wondering if the 0.18 was just a typo. Who knows, it just might be... The document is now gone from the site. Clicking on the link takes you to the main page.

The 7447/7457 is newer than the 7445/7455. The 3rd digit's 5 just indicates it supports L3 cache. The 4 series does not.

Everyone was assuming the PowerBooks would use the 7457 because they were assuming they'd have 1 MB L3 cache. We now know that none of the PowerBooks have L3, so there is no need for the 7457. The 7447 is the same chip but just without L3 cache support.

NightEyes · Sep 17, 2003, 04:49 PM

Jobs specifically stated that the new PB's used the 7457 during the keynote.

Eug · 04:57 PM

Originally posted by NightEyes:
Jobs specifically stated that the new PB's used the 7457 during the keynote.

Well, that's interesting. I have not yet seen the keynote. That doesn't quite make complete sense though since we already know that the old 12" PowerBooks (which didn't have L3) didn't use the 7455. They used the 7445 AFAIK. (ie. Apple used both the 7445 and the 7455 depending on the model.)

However, even if Jobs did say 7457, one wonders why Motorola would release a document saying the 7445 was used by Apple and in laptops. Plus there's no point to use the 7457, which presumably is more expensive than the 7447, if Apple isn't going to use L3 anyway.

But then again the document has now disappeared so who knows.

The other option is the 7447 is for the iMacs, but I had thought the iMacs use the 7445, the document stated laptops, and there is no such thing as 1.33 GHz iMac either.

schk · Sep 17, 2003, 05:02 PM

Well a similar document exists here where it mentions .18 and Apple at the bottom of the document:

http://e-www.motorola.com/files/32bi...477_457RQI.pdf

At the top it says .18 um, but the process is HiP7SOI whereas the 7455 was HIP6SOI, so someone could have just copied the 7455 template and not edited the .18. Also, the main doc for the 7447/57 explicity state .13 um and so does the webpage for it. Most likely a typo. The whole point of the new revision was to move to a lower size which would benefit power usage and allow faster speeds to be reached.

Eug · Sep 17, 2003, 05:05 PM

Originally posted by schk:
Well a similar document exists here where it mentions .18 and Apple at the bottom of the document:

http://e-www.motorola.com/files/32bi...477_457RQI.pdf

At the top it says .18 um, but the process is HiP7SOI whereas the 7455 was HIP6SOI, so someone could have just copied the 7455 template and not edited the .18. Also, the main doc for the 7447/57 explicity state .13 um and so does the webpage for it. Most likely a typo. The whole point of the new revision was to move to a lower size which would benefit power usage and allow faster speeds to be reached.

Hmmm... It looks like the same document, but just with a different name.

I hope for Motorola's sake (and Apple's) that this indeed is just a typo.

BTW, the main point was to shrink the process, but the other point was also to increase performance by doubling the L2 cache size.

thunderous_funker · Sep 17, 2003, 05:20 PM

Interesting.

So it seems we have one doc saying .18 and another saying .13

Considering Apple waited as long as it did for the update, wouldn't it seem logical that they waited for a .13 chip instead of just swapping one .18 chip for another?

I suspect the .18 is a type and that the new chip is .13.

Ratm · Sep 17, 2003, 05:26 PM

The benefits of these new chips can be seen here ---> http://barefeats.com/al15.html

crispinwilliams · Sep 17, 2003, 05:30 PM

Since the operating voltage is between 1.1 and 1.3 V .. I think that it is reasonable to believe that this is 0,13 micron processing.. 0.18 micron likes a higher voltage. Thus I am fairly confident that this is a typo

Eug · Sep 17, 2003, 05:37 PM

Originally posted by crispinwilliams:
Since the operating voltage is between 1.1 and 1.3 V .. I think that it is reasonable to believe that this is 0,13 micron processing.. 0.18 micron likes a higher voltage. Thus I am fairly confident that this is a typo

I agree it's probably a typo, but 7455 chips can run at 1.3 V IIRC.

all2ofme · Sep 17, 2003, 06:25 PM

So are all three of the new models using the same model, though - whether that be 7447 or 7457?

Does that mean that there's even less of a reason that the new 12" should run slower than the 17" machine I have at present?

I understand that the original 12" could have done with the level 3 cache, but given that even the new 15" and 17" don't have it then perhaps it's not the issue it once was...

Eug · Sep 17, 2003, 06:28 PM

Originally posted by all2ofme:
So are all three of the new models using the same model, though - whether that be 7447 or 7457?

Does that mean that there's even less of a reason that the new 12" should run slower than the 17" machine I have at present?

I understand that the original 12" could have done with the level 3 cache, but given that even the new 15" and 17" don't have it then perhaps it's not the issue it once was...

Apple got rid of the L3 cache because with the doubling of the L2 cache size performance was good enough. Get rid of the L3 and save some bux.

I can't say 100% if the all three use the 7447, but I would expect so, since a 7457 without L3 is simply the same thing functionally as a 7447, for presumably extra cost.