Alphelion is correct, a writer wouldn't throw a name up from a "rumor site" and have it published on that companies website, this definitely confirms a 970gx with references to it.
It was IBM's fault not blocking that part out, they saw their mistake and covered it before apple was all over them for it.
Why not? He's human. He could have read the GX rumors, but was actually referring to the FX. There are more reasons than just that, but it's not worth going into everything you can figure out your self.
Why not? He's human. He could have read the GX rumors, but was actually referring to the FX. There are more reasons than just that, but it's not worth going into everything you can figure out your self.
I already figured it out, the odds of that guy accidently mistaking the fx for gx and there being rumors for gx... and him LISTING the fx along side the gx makes it being a typo impossible.
IBM is not a company to publish rumors... I have no idea where you would get that notion... they have got into major trouble in the 10 years for publishing information like that. That's why they are cleaning it up so promptly.
I already figured it out, the odds of that guy accidently mistaking the fx for gx and there being rumors for gx... and him LISTING the fx along side the gx makes it being a typo impossible.
IBM is not a company to publish rumors... I have no idea where you would get that notion... they have got into major trouble in the 10 years for publishing information like that. That's why they are cleaning it up so promptly.
This is the part that gets me though. This guy writes like MOSR.
Quote:
The G5 chips include the IBM 970, 970FX, and 970GX; these are essentially POWER4? cores with an AltiVec unit bolted on.
I note the part where he says "Altivec unit bolted on"... What kind of processor information can a fool like this really have. The Bolted on notion used in the reference is strictly for idiots. I think this whole article slipped through the cracks.
Well, the Power and PowerPC platform has been touted as a platform where different pieces can be 'bolted together' to make a custom version of a PowerPC processor (Cell anyone?). IBM has used the term 'bolted' in their own presentations so it wouldn't be surprising that a writer talking with these same people would come away with the term 'bolted'.
Basically the G5 is a single core Power4 with the AltiVec unit bolted on.
Well, the Power and PowerPC platform has been touted as a platform where different pieces can be 'bolted together' to make a custom version of a PowerPC processor (Cell anyone?). IBM has used the term 'bolted' in their own presentations so it wouldn't be surprising that a writer talking with these same people would come away with the term 'bolted'.
Basically the G5 is a single core Power4 with the AltiVec unit bolted on.
Your saying an IBM engineer used the term bolted on to describe their altivec compatible SIMD engine? I'm shocked.
I totally agree. And I will go one step further and say that the 970GX is actually the lower-power version. I would shat-a-brick if Apple announces a better (or even equivalent) speed processor for the PowerMac and the PowerBook. Whatever the name is for it, IBM does recognize it as a processor and Apple will be using them. If this is the lower-wattage demon, than all bets should be that it goes in a PowerBook.
The Think Secret reports from last year (July and November, if you want to search their archives) discussed Antares - the dual-core 970MP (July), and then AntaresSP - the 970GX (Nov). Both have improved PowerTune features - the feature IBM said made the FX suitable for laptop use. That second TS article also discussed a third variant, a low-power chip said to be testing at 1.6 to 1.8GHz.
So, we might infer the processor's roles as: 970MP - high-end PowerMac; 970GX - low-end PowerMac, iMac; low-power chip: PowerBooks, iBooks and Minis.
Aside from MacOSR's stuff on processors, which most people seem to regard as completely made up, very little more has been said about this low-power chip (which I shall dub the G5M). But as I have said elsewhere, the GHz ratings of theG5M seem rather low. Aside from removing the FSB bottleneck, it doesn't seem to offer a big advantage over the current G4s or the near-future G4s. Apple know that, of course. So is the G5M perhaps more interesting? Dual core?
Well, bolted on sounds like a second hand hack, or an afterthought when the SIMD engine IBM came up with was (I think) better than Altivec as I recall. Plus it's utilizing all altivec code. That's a nothing to be scoffing at by saying bolted on IMO.
"The original version of this article contained an erroneous reference to the 970GX; this was an oversight. We apologize for any confusion this may have caused."
If there was no such thing as a 970GX, they would have said something like "contained an erroneous reference to a '970GX'", not "the 970GX", which, as Aphelion points out, implies that it is an existing thing. But even stranger is calling the error an "oversight", which means something was omitted.
with altivec that I posted about 2 years ago after reading
a list of publications for an australian microprocessor conference. I had thought this would be the 750GX but unfortunately no altivec. Perhaps they have finally decided to produce it and this writer thought it was a new 970 variant.
If there was no such thing as a 970GX, they would have said something like "contained an erroneous reference to a '970GX'", not "the 970GX", which, as Aphelion points out, implies that it is an existing thing. But even stranger is calling the error an "oversight", which means something was omitted.
It now reads 'a "970GX"'. Do you think no one reads these silly threads.
I think its quite clear that between IBM failing to produce 3Ghz G5's, let alone in high yields, to failing on being able to produce a G5 laptop with low heat specs Apple will make sure they downplay or completely never mention Laptop and G5 until they have hundreds of thousands of chips in their possession.
It now reads 'a "970GX"'. Do you think no one reads these silly threads.
Hi Steve.
*Confirmed*! The 970GX has Altivec II !!!
It's all very logical: If the 970GX is *confirmed* because the reference was edited out, it *must* exist; Since every G5 has Altivec and a Power4-derived core but *the* 970GX doesn't fill that bill and was edited out to not cause confusion, it *must* have another incarnation of Altivec: Altivec 2!
Well, I mean logical in a Dougls Adams-kind of a way...
970GX was mentioned erroneously my ass! Apple caught wind, probably from this very thread, and suddenly the lacky decides to deny it.
If I were a betting man, I would say that this is damn near confirmation that the 970GX is being mass-produced as we speak, and that final system tests are being concluded (or have been already) and software tweaks are being done. Perhaps for 10.3.9 or Tiger, whichever rears its head first.
Comments
Originally posted by emig647
Based on what?
Alphelion is correct, a writer wouldn't throw a name up from a "rumor site" and have it published on that companies website, this definitely confirms a 970gx with references to it.
It was IBM's fault not blocking that part out, they saw their mistake and covered it before apple was all over them for it.
Why not? He's human. He could have read the GX rumors, but was actually referring to the FX. There are more reasons than just that, but it's not worth going into everything you can figure out your self.
Originally posted by onlooker
Why not? He's human. He could have read the GX rumors, but was actually referring to the FX. There are more reasons than just that, but it's not worth going into everything you can figure out your self.
I already figured it out, the odds of that guy accidently mistaking the fx for gx and there being rumors for gx... and him LISTING the fx along side the gx makes it being a typo impossible.
IBM is not a company to publish rumors... I have no idea where you would get that notion... they have got into major trouble in the 10 years for publishing information like that. That's why they are cleaning it up so promptly.
Originally posted by emig647
I already figured it out, the odds of that guy accidently mistaking the fx for gx and there being rumors for gx... and him LISTING the fx along side the gx makes it being a typo impossible.
IBM is not a company to publish rumors... I have no idea where you would get that notion... they have got into major trouble in the 10 years for publishing information like that. That's why they are cleaning it up so promptly.
This is the part that gets me though. This guy writes like MOSR.
The G5 chips include the IBM 970, 970FX, and 970GX; these are essentially POWER4? cores with an AltiVec unit bolted on.
I note the part where he says "Altivec unit bolted on"... What kind of processor information can a fool like this really have. The Bolted on notion used in the reference is strictly for idiots. I think this whole article slipped through the cracks.
Basically the G5 is a single core Power4 with the AltiVec unit bolted on.
Originally posted by PBG4 Dude
Well, the Power and PowerPC platform has been touted as a platform where different pieces can be 'bolted together' to make a custom version of a PowerPC processor (Cell anyone?). IBM has used the term 'bolted' in their own presentations so it wouldn't be surprising that a writer talking with these same people would come away with the term 'bolted'.
Basically the G5 is a single core Power4 with the AltiVec unit bolted on.
Your saying an IBM engineer used the term bolted on to describe their altivec compatible SIMD engine? I'm shocked.
Originally posted by onlooker
Your saying an IBM engineer used the term bolted on to describe their altivec compatible SIMD engine? I'm shocked.
Originally posted by Rhumgod
I totally agree. And I will go one step further and say that the 970GX is actually the lower-power version. I would shat-a-brick if Apple announces a better (or even equivalent) speed processor for the PowerMac and the PowerBook. Whatever the name is for it, IBM does recognize it as a processor and Apple will be using them. If this is the lower-wattage demon, than all bets should be that it goes in a PowerBook.
The Think Secret reports from last year (July and November, if you want to search their archives) discussed Antares - the dual-core 970MP (July), and then AntaresSP - the 970GX (Nov). Both have improved PowerTune features - the feature IBM said made the FX suitable for laptop use. That second TS article also discussed a third variant, a low-power chip said to be testing at 1.6 to 1.8GHz.
So, we might infer the processor's roles as: 970MP - high-end PowerMac; 970GX - low-end PowerMac, iMac; low-power chip: PowerBooks, iBooks and Minis.
Aside from MacOSR's stuff on processors, which most people seem to regard as completely made up, very little more has been said about this low-power chip (which I shall dub the G5M). But as I have said elsewhere, the GHz ratings of theG5M seem rather low. Aside from removing the FSB bottleneck, it doesn't seem to offer a big advantage over the current G4s or the near-future G4s. Apple know that, of course. So is the G5M perhaps more interesting? Dual core?
I think we should be told.
Pip pip!
Originally posted by onlooker
Your saying an IBM engineer used the term bolted on to describe their altivec compatible SIMD engine? I'm shocked.
Sometimes you have to simplify to help tech writer to understand.
Originally posted by PBG4 Dude
Well, bolted on sounds like a second hand hack, or an afterthought when the SIMD engine IBM came up with was (I think) better than Altivec as I recall. Plus it's utilizing all altivec code. That's a nothing to be scoffing at by saying bolted on IMO.
"The original version of this article contained an erroneous reference to the 970GX; this was an oversight. We apologize for any confusion this may have caused."
If there was no such thing as a 970GX, they would have said something like "contained an erroneous reference to a '970GX'", not "the 970GX", which, as Aphelion points out, implies that it is an existing thing. But even stranger is calling the error an "oversight", which means something was omitted.
Originally posted by cubist
Strangely worded apology: ... which means something was omitted.
What was omitted was a proper proofread with regard to Apple's sensitivities.
with altivec that I posted about 2 years ago after reading
a list of publications for an australian microprocessor conference. I had thought this would be the 750GX but unfortunately no altivec. Perhaps they have finally decided to produce it and this writer thought it was a new 970 variant.
Originally posted by cubist
Strangely worded apology:
If there was no such thing as a 970GX, they would have said something like "contained an erroneous reference to a '970GX'", not "the 970GX", which, as Aphelion points out, implies that it is an existing thing. But even stranger is calling the error an "oversight", which means something was omitted.
It now reads 'a "970GX"'. Do you think no one reads these silly threads.
Hi Steve.
And for damn sure they read these forums.
Originally posted by Kurt
It now reads 'a "970GX"'. Do you think no one reads these silly threads.
Hi Steve.
*Confirmed*! The 970GX has Altivec II !!!
It's all very logical: If the 970GX is *confirmed* because the reference was edited out, it *must* exist; Since every G5 has Altivec and a Power4-derived core but *the* 970GX doesn't fill that bill and was edited out to not cause confusion, it *must* have another incarnation of Altivec: Altivec 2!
Well, I mean logical in a Dougls Adams-kind of a way...
If I were a betting man, I would say that this is damn near confirmation that the 970GX is being mass-produced as we speak, and that final system tests are being concluded (or have been already) and software tweaks are being done. Perhaps for 10.3.9 or Tiger, whichever rears its head first.
Originally posted by Kurt
Hi Steve.
Originally posted by Anders
That does it! Apple is going back to Motorola.