I think this thread has pretty much stayed on topic and is very informative....
anyway...
I just came into some information on the GPUL I am still digesting. Sometimes some juicy info can come by... But L2 cache is 512KB 8-way set associative connected by a wide bus. L1 is fully inclusive and is 32KB for data and 64KB for instruction. It is held in the L2 also.
<strong>I think this thread has pretty much stayed on topic and is very informative....
anyway...
I just came into some information on the GPUL I am still digesting. Sometimes some juicy info can come by... But L2 cache is 512KB 8-way set associative connected by a wide bus. L1 is fully inclusive and is 32KB for data and 64KB for instruction. It is held in the L2 also.
I will try and make sense of other info.</strong><hr></blockquote>
Outsider that's great information and much needed. It seems that caching is very important for current Macs but since this is a new proc though our mileage might vary.
I say we keep this thread open until Oct 15 and then close it. That way we can kind of compare information pre and post Microprocessor forum.
I find it hard to believe that the GPUL powermac will wait until the summer. Power users are starved - and there's no doubt that all these rumors about the post-g4 processor are hurting powermac sales.
I personally don't believe Apple can get away with sticking G4's in powermacs again, once IBM announces the chip in just over one week. think about it. If IBM is ready to show the chip, how could it be a year from production? They already have a fab plant - it's not being built - it's making chips. They already make a power4 (and pwoer3, etc.. chip) - this is not unfamiliar territory for IBM. I figure once the powermac users find out about the new ibm chip next week, that it has altivec, and that it is technically mac-compatible, nobody in their right mind will buy another powermac g4 unless they absolutely HAVE to (ie, their machine is dead).
[quote]Originally posted by wormboy:
<strong>I think that the consensus here is that yes, IBM will be making a chip that finds its way into the powermac line. The debate is about when this will happen. Crazy optimists are pegging late winter/spring, realists are saying late summer/fall, and pessimists are muttering something about 2004. The later two groups are guessing one more stepping from the G4 on the current motherboard with the current FSB, sine motorola will not move Maxbus to DDR. When the IBM processor arrives, it will be in conjunction with a new bus interconnect. The chip controller may be on-die or not, and if not may be a hypertransport derivative called ApplePI.</strong><hr></blockquote>
<strong>I find it hard to believe that the GPUL powermac will wait until the summer. Power users are starved - and there's no doubt that all these rumors about the post-g4 processor are hurting powermac sales.</strong><hr></blockquote>
I share your enthusiasm for GPUL, but I don't share your sense of urgency. I have a dual GHZ G4 and I'm not complaining about speed. It's very powerful for the things I do (Photoshop, Video Editing, etc. ... even games - yes, I have the Geforce 4 Ti). This setup on Jag is competitive with most high end PCs. Sure, there will be some single threaded tasks that are faster on the PC, etc. In fact, for things like Photoshop, encoding MP3, etc. it's 3x faster than my 1 GHZ P3 (Dell). A 2.8 GHZ P4 is no where near 3x faster than 1 GHZ P3, more like 2x faster.
As for the G4, Mot hasn't even moved to HiP7 (.13u) yet. Likewise, assuming MOT can pull it's head out of it's a$$ and ship a G4 on .13u (possibly with larger L2 cache), the PowerMacs should be in for a nice speed increase this January / February. If Apple then follows in late summer, 2003 with GPUL, that will be a natural progression.
[quote]<strong>I personally don't believe Apple can get away with sticking G4's in powermacs again, once IBM announces the chip in just over one week. think about it. If IBM is ready to show the chip, how could it be a year from production? </strong><hr></blockquote>
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
I find it hard to believe that the GPUL powermac will wait until the summer. Power users are starved - and there's no doubt that all these rumors about the post-g4 processor are hurting powermac sales.
</strong><hr></blockquote>
Unfortunately, the need for a faster PPC in the Mac cannot speed up the process, no matter how much you try. If there are enough people working on it, adding more just slows it down. However, I still am on the optimistic side. I believe the GPUL has been in the works for quite a while, and IBM is close to having samples from a pilot run. What might delay things now is waiting for the 90 nm process for initial production. So, will we see 130 or 90 nm chips?
Apple and IBM surly were aware of the anticipation a chip like this would produce, once people know about it. I don't think they would have scheduled an October 15 presentation if the chip were far from ready.
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
Steve</strong><hr></blockquote>
Yeah, but if this *is* apple's next generation processor, I think it will be extremely difficult for apple to keep interest high on a g4 line that we all but *know* is on its last legs.
I hear you about speed and the new machines - in fact, I'm setting up a dual 867 tonight for a friend - it's bloody fast. no doubt. But consider that most mac users don't upgrade their machine every year and that most people (myself included) are going to look for the biggest bang for the buck, I don't think it's a stretch to say assume that if people are expecting a next-gen system in July, they'll wait for it. Which makes me wonder, how is Apple going to keep people interested in powermacs (even faster ones) until they ship with a GPUL?
Rebates? Bogus photoshop stats showing the G4 to be 10X faster than a P4? Monitor savings? All of this hasn't helped much over the past year. Sales are way down.
I mean, once the cat's out of the bag (one might argue, it already is) what is the incentive to buy a powermac G4 other than immediate need?
Is Apple ready to throw it's high-margin powermac sales out the window for 6 months? Isn't it at least possible that the machines will be ready to ship by jan?
Good old MacOS Rumors... Whenever you need a good laugh they are always willing to help.
"Although Apple and IBM continue to work on a solution that some have dubbed GPUL (Giga-Processor Ultra Lite - although this is most certainly not the name anyone involved is using; certain rumormongers just made it up, as they have numerous other "codenames")"
I agree, the GPUL is closer than many are letting on. IBM wouldn't announce Apple's next generation CPU unless Apple was nearly ready to announce the same CPU.
Also, rumors suggest that as long as a year ago, this chip was being tested in Apple boxes. IBM has had plenty of time to make this CPU, the plant is up and running--the time is here now, folks.
<strong>the time is here now, folks.</strong><hr></blockquote>
Stupid question, but if the GPUL and new board to go with it were ready to go would Apple really have spent time and effort on a new board for "GoFasterAirScoops" (GFAS)?
I can't see it really. Any new chip is a fair way off yet - I'd say at the earliest June ? perhaps even a year from now.
Stupid question, but if the GPUL and new board to go with it were ready to go would Apple really have spent time and effort on a new board for "GoFasterAirScoops" (GFAS)?
I can't see it really. Any new chip is a fair way off yet - I'd say at the earliest June ? perhaps even a year from now.</strong><hr></blockquote>
The chipset in the new PowerMacs are an evolution of the ones that preceded it. They are mostly unchanged than what was presented over 2 years ago when the original Sawtooth was showed off. UniNorth lives on still. This new chipset will be a revolution. And how hard is it to beleive t hat Apple would have 2 simultaneous chipsets in development? Not only do I find that reasonable but necessary.
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
Steve</strong><hr></blockquote>
Yes, and most rumors point to a 2nd half 2003 release.
The chipset in the new PowerMacs are an evolution of the ones that preceded it.</strong><hr></blockquote>
Chipsets have ALWAYS been evolutionary. You just can't start from scratch with something completely new. With the U2 (UniNorth IS gone, U2 is the successor) controller, we have dual ATA-100 and DDR RAM controllers. With UniNorth 7 (from memory) we had gigabit ethernet. The first g4 uninorth saw the introduction of AGP. It's always going to be gradual. The GPUL chipset will probably be like the current one but with ApplePI and maybe one or two new features.
I know the odds are against it but I would not be shocked if this new IBM shows up in a Macintosh much sooner than is speculated.
What good reason was there for Apple to do a major overhaul of the PowerMac internals recently, focusing on a massive upgrade in cooling? Could it have been a real-world shakedown of the mechanicals which will (speculating) be needed for the new chip? If the answer to that is yes, then why would they have released the double mirror design a year or so ahead of the chip? Surely they wouldn't have needed the boxes in the field for that long to see how they hold up.
well, for what it's worth, the info i got was that the ibm chip started sampling some time around april or may of this year. i don't know how long it take to go from sampling to appearing in a product, but that info is probably pretty darn accurate. it was from someone who knows.
Comments
<strong>Sounds like all the other rumours</strong><hr></blockquote>
Sure enough - but did you see "G5 and GPUL" elsewhere?
I suppose there is some sense in it if G5 is going for the low-power sector.
<strong>Sure enough - but did you see "G5 and GPUL" elsewhere?</strong><hr></blockquote>
Yes. What you posted was just a rehash of the eWeek article.
anyway...
I just came into some information on the GPUL I am still digesting. Sometimes some juicy info can come by... But L2 cache is 512KB 8-way set associative connected by a wide bus. L1 is fully inclusive and is 32KB for data and 64KB for instruction. It is held in the L2 also.
I will try and make sense of other info.
<strong>I think this thread has pretty much stayed on topic and is very informative....
anyway...
I just came into some information on the GPUL I am still digesting. Sometimes some juicy info can come by... But L2 cache is 512KB 8-way set associative connected by a wide bus. L1 is fully inclusive and is 32KB for data and 64KB for instruction. It is held in the L2 also.
I will try and make sense of other info.</strong><hr></blockquote>
Outsider that's great information and much needed. It seems that caching is very important for current Macs but since this is a new proc though our mileage might vary.
I say we keep this thread open until Oct 15 and then close it. That way we can kind of compare information pre and post Microprocessor forum.
Kill it already.
I personally don't believe Apple can get away with sticking G4's in powermacs again, once IBM announces the chip in just over one week. think about it. If IBM is ready to show the chip, how could it be a year from production? They already have a fab plant - it's not being built - it's making chips. They already make a power4 (and pwoer3, etc.. chip) - this is not unfamiliar territory for IBM. I figure once the powermac users find out about the new ibm chip next week, that it has altivec, and that it is technically mac-compatible, nobody in their right mind will buy another powermac g4 unless they absolutely HAVE to (ie, their machine is dead).
[quote]Originally posted by wormboy:
<strong>I think that the consensus here is that yes, IBM will be making a chip that finds its way into the powermac line. The debate is about when this will happen. Crazy optimists are pegging late winter/spring, realists are saying late summer/fall, and pessimists are muttering something about 2004. The later two groups are guessing one more stepping from the G4 on the current motherboard with the current FSB, sine motorola will not move Maxbus to DDR. When the IBM processor arrives, it will be in conjunction with a new bus interconnect. The chip controller may be on-die or not, and if not may be a hypertransport derivative called ApplePI.</strong><hr></blockquote>
<strong>I find it hard to believe that the GPUL powermac will wait until the summer. Power users are starved - and there's no doubt that all these rumors about the post-g4 processor are hurting powermac sales.</strong><hr></blockquote>
I share your enthusiasm for GPUL, but I don't share your sense of urgency. I have a dual GHZ G4 and I'm not complaining about speed. It's very powerful for the things I do (Photoshop, Video Editing, etc. ... even games - yes, I have the Geforce 4 Ti). This setup on Jag is competitive with most high end PCs. Sure, there will be some single threaded tasks that are faster on the PC, etc. In fact, for things like Photoshop, encoding MP3, etc. it's 3x faster than my 1 GHZ P3 (Dell). A 2.8 GHZ P4 is no where near 3x faster than 1 GHZ P3, more like 2x faster.
As for the G4, Mot hasn't even moved to HiP7 (.13u) yet. Likewise, assuming MOT can pull it's head out of it's a$$ and ship a G4 on .13u (possibly with larger L2 cache), the PowerMacs should be in for a nice speed increase this January / February. If Apple then follows in late summer, 2003 with GPUL, that will be a natural progression.
[quote]<strong>I personally don't believe Apple can get away with sticking G4's in powermacs again, once IBM announces the chip in just over one week. think about it. If IBM is ready to show the chip, how could it be a year from production? </strong><hr></blockquote>
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
Steve
<strong>
I find it hard to believe that the GPUL powermac will wait until the summer. Power users are starved - and there's no doubt that all these rumors about the post-g4 processor are hurting powermac sales.
</strong><hr></blockquote>
Unfortunately, the need for a faster PPC in the Mac cannot speed up the process, no matter how much you try. If there are enough people working on it, adding more just slows it down. However, I still am on the optimistic side. I believe the GPUL has been in the works for quite a while, and IBM is close to having samples from a pilot run. What might delay things now is waiting for the 90 nm process for initial production. So, will we see 130 or 90 nm chips?
Apple and IBM surly were aware of the anticipation a chip like this would produce, once people know about it. I don't think they would have scheduled an October 15 presentation if the chip were far from ready.
<strong>
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
Steve</strong><hr></blockquote>
Yeah, but if this *is* apple's next generation processor, I think it will be extremely difficult for apple to keep interest high on a g4 line that we all but *know* is on its last legs.
I hear you about speed and the new machines - in fact, I'm setting up a dual 867 tonight for a friend - it's bloody fast. no doubt. But consider that most mac users don't upgrade their machine every year and that most people (myself included) are going to look for the biggest bang for the buck, I don't think it's a stretch to say assume that if people are expecting a next-gen system in July, they'll wait for it. Which makes me wonder, how is Apple going to keep people interested in powermacs (even faster ones) until they ship with a GPUL?
Rebates? Bogus photoshop stats showing the G4 to be 10X faster than a P4? Monitor savings? All of this hasn't helped much over the past year. Sales are way down.
I mean, once the cat's out of the bag (one might argue, it already is) what is the incentive to buy a powermac G4 other than immediate need?
Is Apple ready to throw it's high-margin powermac sales out the window for 6 months? Isn't it at least possible that the machines will be ready to ship by jan?
<img src="confused.gif" border="0">
"Although Apple and IBM continue to work on a solution that some have dubbed GPUL (Giga-Processor Ultra Lite - although this is most certainly not the name anyone involved is using; certain rumormongers just made it up, as they have numerous other "codenames")"
"certain rumormongers just made it up"
<img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" /> <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
It's no wonder so few have a good thing to say about them.
D
Also, rumors suggest that as long as a year ago, this chip was being tested in Apple boxes. IBM has had plenty of time to make this CPU, the plant is up and running--the time is here now, folks.
<strong>the time is here now, folks.</strong><hr></blockquote>
Stupid question, but if the GPUL and new board to go with it were ready to go would Apple really have spent time and effort on a new board for "GoFasterAirScoops" (GFAS)?
I can't see it really. Any new chip is a fair way off yet - I'd say at the earliest June ? perhaps even a year from now.
<strong>
Stupid question, but if the GPUL and new board to go with it were ready to go would Apple really have spent time and effort on a new board for "GoFasterAirScoops" (GFAS)?
I can't see it really. Any new chip is a fair way off yet - I'd say at the earliest June ? perhaps even a year from now.</strong><hr></blockquote>
The chipset in the new PowerMacs are an evolution of the ones that preceded it. They are mostly unchanged than what was presented over 2 years ago when the original Sawtooth was showed off. UniNorth lives on still. This new chipset will be a revolution. And how hard is it to beleive t hat Apple would have 2 simultaneous chipsets in development? Not only do I find that reasonable but necessary.
<strong>
There's a big difference between showing a technology and shipping a completed product. IIRC, the current G4s were described at this same processor forum about a year before they shipped, correct?
Steve</strong><hr></blockquote>
Yes, and most rumors point to a 2nd half 2003 release.
Barto
<strong>
The chipset in the new PowerMacs are an evolution of the ones that preceded it.</strong><hr></blockquote>
Chipsets have ALWAYS been evolutionary. You just can't start from scratch with something completely new. With the U2 (UniNorth IS gone, U2 is the successor) controller, we have dual ATA-100 and DDR RAM controllers. With UniNorth 7 (from memory) we had gigabit ethernet. The first g4 uninorth saw the introduction of AGP. It's always going to be gradual. The GPUL chipset will probably be like the current one but with ApplePI and maybe one or two new features.
Barto
What good reason was there for Apple to do a major overhaul of the PowerMac internals recently, focusing on a massive upgrade in cooling? Could it have been a real-world shakedown of the mechanicals which will (speculating) be needed for the new chip? If the answer to that is yes, then why would they have released the double mirror design a year or so ahead of the chip? Surely they wouldn't have needed the boxes in the field for that long to see how they hold up.