The Register reports that I"BM confirms PowerPC 750GX 'Gobi' spec."
From: http://www.theregister.co.uk/content/39/31492.html
IBM has come clean and detailed the attributes of its latest 32-bit G3-class PowerPC processor, the 750GX. The new chip is the successor to the 750FX, aka 'Sahara', that powers Apple's consumer-oriented iBook notebook.
As revealed by The Register back in May, the 750GX is the chip that will take the G3 to 1GHz and beyond. The chip will launch at 733MHz to 1.1GHz, IBM said yesterday. The 750FX already clocks at up to 900MHz.
Beyond clock speed, the 750GX offers double the on-die L2 cache of its predecessor - 1MB - designed to improve performance further by ensuring data the CPU needs is less likely to have to be pulled in from main memory via the chip's 200MHz 60x frontside bus.
Other improvements are more deeply architectural. The 750GX's L2 cache is also four-way set associative - the 750FX's cache is two-way set associative - essentially breaking it down into four 256KB units. Set associativity is a mapping technique used within the cache to improve data access times. The 750GX's four cache units can be locked to ensure that specific blocks of data are retained.
IBM has also added cache line miss buffers between the L1 data cache and the L2 cache, and the L2 and the bus interface unit (BIU), which controls the flow of data and instructions in and out of the chip. Essentially this provides further fall-back positions if the data the processor needs isn't in a particular cache. The BIU can stack up to five bus transactions - the 750FX's BIU could pipeline up to three transactions - just in case the cache doesn't contain what the processor needs. This is all about readying memory transactions just in case they're needed, speeding things up if they are.
Less relevant to Apple's designers but of interest to embedded applications developers will be the 750GX's 'instruction only' cache mode, in which the L2 is used to store only instructions, not data, which is instead pulled directly from memory each time.
Early speculation suggested the 750GX would be produced using a 100nm (0.10 micron) process. This proved unfounded: like the 750FX, the 750GX will be fabbed at 130nm, again using a copper process with silicon-on-insulator technology. So it's the architectural changes that have brought the 200MHz increase in maximum clock frequency rather than the die-shrink we were anticipating.
The extra cache and circuitry means that the 750GX is bigger than the 750FX: 51.9mm squared compared to 36.6mm squared, boosting the chip's heat output. That said, it's still an impressive less than 8W at 1GHz. The core voltage stays at 1.45V.
The 750GX begins sampling in July, IBM said, followed by full-scale production in December, so don't expect iBooks based on the new chip any time soon. In any case, it's probably about time Apple upgraded its consumer notebook line to the G4 to take allow it to advantage of AltiVec. Then again, IBM's anticipated follow-up to the 750GX, codenamed 'Mojave', is believed to contain an AltiVec unit, and Apple may be targeting that processor for future iBooks.
IBM has come clean and detailed the attributes of its latest 32-bit G3-class PowerPC processor, the 750GX. The new chip is the successor to the 750FX, aka 'Sahara', that powers Apple's consumer-oriented iBook notebook.
As revealed by The Register back in May, the 750GX is the chip that will take the G3 to 1GHz and beyond. The chip will launch at 733MHz to 1.1GHz, IBM said yesterday. The 750FX already clocks at up to 900MHz.
Beyond clock speed, the 750GX offers double the on-die L2 cache of its predecessor - 1MB - designed to improve performance further by ensuring data the CPU needs is less likely to have to be pulled in from main memory via the chip's 200MHz 60x frontside bus.
Other improvements are more deeply architectural. The 750GX's L2 cache is also four-way set associative - the 750FX's cache is two-way set associative - essentially breaking it down into four 256KB units. Set associativity is a mapping technique used within the cache to improve data access times. The 750GX's four cache units can be locked to ensure that specific blocks of data are retained.
IBM has also added cache line miss buffers between the L1 data cache and the L2 cache, and the L2 and the bus interface unit (BIU), which controls the flow of data and instructions in and out of the chip. Essentially this provides further fall-back positions if the data the processor needs isn't in a particular cache. The BIU can stack up to five bus transactions - the 750FX's BIU could pipeline up to three transactions - just in case the cache doesn't contain what the processor needs. This is all about readying memory transactions just in case they're needed, speeding things up if they are.
Less relevant to Apple's designers but of interest to embedded applications developers will be the 750GX's 'instruction only' cache mode, in which the L2 is used to store only instructions, not data, which is instead pulled directly from memory each time.
Early speculation suggested the 750GX would be produced using a 100nm (0.10 micron) process. This proved unfounded: like the 750FX, the 750GX will be fabbed at 130nm, again using a copper process with silicon-on-insulator technology. So it's the architectural changes that have brought the 200MHz increase in maximum clock frequency rather than the die-shrink we were anticipating.
The extra cache and circuitry means that the 750GX is bigger than the 750FX: 51.9mm squared compared to 36.6mm squared, boosting the chip's heat output. That said, it's still an impressive less than 8W at 1GHz. The core voltage stays at 1.45V.
The 750GX begins sampling in July, IBM said, followed by full-scale production in December, so don't expect iBooks based on the new chip any time soon. In any case, it's probably about time Apple upgraded its consumer notebook line to the G4 to take allow it to advantage of AltiVec. Then again, IBM's anticipated follow-up to the 750GX, codenamed 'Mojave', is believed to contain an AltiVec unit, and Apple may be targeting that processor for future iBooks.
Comments
Not much more than a (very moderate) speed increase.
Well, maybe not 2GHz, but definitely somewhere in the mid-1GHz range and Altivec.
Man, we're so messed up. Makes me sad to think where we'd be now had things not stalled and petered out these past few years.
Thanks, M'ola...
Originally posted by Matsu
No altivec? No point. The G3 has been perpetually stunted and late, more so than even the G4, according to early statements by IBM, we all shoulda been using 2Ghz altivec G3's months ago.
100% agreed! Looks pretty lame compared to a 7447. Let's hope that the positive signs regarding the release date of the 7447/7457 are turning into real facts.
End of Line
I hope Apple doesn't use it.
remember it is a pin-compatible drop in replacement for the current 750fx that is currently used in ibooks.
the chip we are waiting for will clock much higher than this.
i believe it will come soon but in what form i dont know.
i feel that apple will use this new as yet to be released chip in a new low cost line of desktops,you heard it here first!
apple needs to get into the sub-1000 dollar space badly.
these new modified "GOBI" chip will be the ticket.
they will ramp up close to 2 ghz.
you will all be amazed.....again!
Originally posted by Matsu
No altivec? No point. The G3 has been perpetually stunted and late, more so than even the G4, according to early statements by IBM, we all shoulda been using 2Ghz altivec G3's months ago.
...but but, wouldn't it be a 'G4' if it had Altivec? Is the SIMD unit not THE main difference between the two? Also, I reckon that this is an incremental release, not the 'Gobi' but whaddoIknow.
kroehl
Originally posted by kroehl
...but but, wouldn't it be a 'G4' if it had Altivec? Is the SIMD unit not THE main difference between the two? Also, I reckon that this is an incremental release, not the 'Gobi' but whaddoIknow.
kroehl
The 750VX is the altivec G3. No, the G3 would not be a G4 with Altivec. In addition to Altivec, the G4 also is a MERSI chip, meaning that you can use it in dual CPU configurations. Thr G3 doesn't do MERSI, so it will always be a single CPU chip (it can't get a date).
The 750VX could be a great chip if it is released anytime soon. Unfortunately, it won't be.
Originally posted by geekmeet
. . . the chip we are waiting for will clock much higher than this.
i believe it will come soon but in what form i dont know.
i feel that apple will use this new as yet to be released chip in a new low cost line of desktops,you heard it here first! . . .
I'd like to believe what you say. The IBM newsletter speaks of several improvements in the 750GX for embedded applications, so the 750GX may be only for this market. If you are correct, IBM and Apple are working together on a second processor for the Mac. It might have a good FSB, AltiVec and a memory controller on the chip. It would be on the 90 nanometer process and clock at about 2 GHz. Why not?
Apple had the foresight to develop the 970 with IBM. IBM was motivated to go along because the 970 is useful to them in the low end server market. IBM now has something to compete with the Xeon. With this processor that you say is coming soon, Apple gets a chip for the low end, and IBM gets a still more powerful chip for selling to the embedded market. I hope you are right.
Originally posted by Yevgeny
The 750VX is the altivec G3. No, the G3 would not be a G4 with Altivec. In addition to Altivec, the G4 also is a MERSI chip, meaning that you can use it in dual CPU configurations. Thr G3 doesn't do MERSI, so it will always be a single CPU chip (it can't get a date).
I know you do not wish to call Mojave a G4, but what else would it be called? It will run applications that require a G4, as stated in their system requirements. Calling it a G3 would confuse customers who want to find out which applications will run, and which will not. The customer will care less whether the CPU is capable of running as a dual if they didn't buy a dual.
By the way, I've also heard that Mojave will have SMP.
Originally posted by Junkyard Dawg
Yawn. Except for the 130 nm process, this is the sort of CPU Motorola could fab.
I hope Apple doesn't use it.
Why wouldn't you hope Apple uses the chip? Sure, it's not a breakthrough like the 970, but it gets the iBook up above 1ghz. While it might not mean much to us, it could mean something to PC swithchers who see the iBook with faster speeds.
I think we'll get RIO G4's before we get altivec G3's, just a feeling.
But, if Apple has gone to IBM for a low power 32bit chip, then that should tell you something about the overall heat and cost numbers of the G5 platform. I think it's kinda funny that all the people who insist Apple/IBM will move G5's into every product also insist that IBM is pursuing a G4 equivalent. Why would they do that if it were possible to move the G5 into those spaces within a short space of time?
Easy answer? It isn't, not for IBM, and not for Apple.
Originally posted by Matsu
But, if Apple has gone to IBM for a low power 32bit chip, then that should tell you something about the overall heat and cost numbers of the G5 platform. I think it's kinda funny that all the people who insist Apple/IBM will move G5's into every product also insist that IBM is pursuing a G4 equivalent. Why would they do that if it were possible to move the G5 into those spaces within a short space of time?
Easy answer? It isn't, not for IBM, and not for Apple.
I think Apple/IBM are making several chips together, not just the 970 we already have. Apple is sure in a quandry if they are not making 32-bit chips. Why? Sure, they could still rely on Motorola, but for what? No chip advancement. Not bloody likely! They've been down that road. No Motorola is really no longer an option for advancing any product line.
Besides, Apple sells lotsa 32-bit stuff, G3 and G4-based systems that will not advance through use of Motorola-made crap. That fact alone tells me something else is coming. Either that, or Apple is abandoning iMac/eMac/iBook/PowerBook (for now) lines altogether. They sure ain't gonna see a G5 any time soon. Gotta have a lower-end, lower-power, lower-cost solution for the masses. Not everyone is buying a Power Mac G5.
G5 across the line? Bah, what would be the point of spending $3k+ on a Power Mac when you could get it in an iMac for half the price. Just as happened when the intro'd the LCD iMac with a G4, they shot themselves in the arse by canabalizing Power Mac sales.
If you ask me, Apple is a little fat in product line right now. They need to slim down to the 2x2 strategy again - iMac/Power Mac and iBook/PowerBook. Screw the eMac, it's a dead horse anyway, the last CRT dinosaur still around, no expansion, you've heard it all before.
The iMac should separate from the AIO strategy and just use a NeXTstation like pizza box with basic IO one PCI-X slot and the normal accompaniment of hardware.
My 2 cents...
Originally posted by Rhumgod
I
G5 across the line? Bah, what would be the point of spending $3k+ on a Power Mac when you could get it in an iMac for half the price. Just as happened when the intro'd the LCD iMac with a G4, they shot themselves in the arse by canabalizing Power Mac sales.
If you ask me, Apple is a little fat in product line right now. They need to slim down to the 2x2 strategy again - iMac/Power Mac and iBook/PowerBook. Screw the eMac, it's a dead horse anyway, the last CRT dinosaur still around, no expansion, you've heard it all before.
The iMac should separate from the AIO strategy and just use a NeXTstation like pizza box with basic IO one PCI-X slot and the normal accompaniment of hardware.
My 2 cents...
that could quite possibly be the worst 2 cents ever.
Originally posted by applenut
that could quite possibly be the worst 2 cents ever.
Originally posted by Matsu
I think you would have to call any altivec G3 (with improved fp) a G4. DP is really not needed in any applications where Apple would use it.
I think we'll get RIO G4's before we get altivec G3's, just a feeling.
But, if Apple has gone to IBM for a low power 32bit chip, then that should tell you something about the overall heat and cost numbers of the G5 platform. I think it's kinda funny that all the people who insist Apple/IBM will move G5's into every product also insist that IBM is pursuing a G4 equivalent. Why would they do that if it were possible to move the G5 into those spaces within a short space of time?
Easy answer? It isn't, not for IBM, and not for Apple.
Ok, so I am just being picky. G4's have MERSI, G3's don't. I agree that from a consumer standpoint, a G3 with Altivec might as well be a G4.
A die shrunk RIO G4 would be quite a good thing for powerbooks. My concern is that Moto hasn't said anything about this yet (RIO). Have I missed something on their roadmap(s) or have my expectations for Moto sunk so low?
A better question for IBM would be why we need a G4 clone in eight months when the G5 is nearing a die shrink. Either IBM thinks you can't put a G5 in a laptop (maybe), or that Apple will want to differentiate its machines (most likely), or that Apple wants to ditch Moto altogether (depends on Mr. Job's mood), or they are making it for the embedded market (very likely).
I see the 750VX as the ibook chip a year from now once the G5 is in powerbooks. A 1.5GHz 750VX would make for an incredible iBook. I don't think that Apple will go with G4's long term because Moto hasn't the ability or inclination to scale the CPU speed. Hence a low end G4 replacement for the low end machines.