Yeah, but isn't there an inherent short-coming to making processors that use longer pipelines (a la the MHz myth animation thing Jobs did last July)? That is, with the shorter pipelines, current G4 chips greatly outpace Intel and AMD chips at similar clock speeds (and even moderately higher clock speeds), because they use much longer pipelines. So if we go in that direction with the G5, aren't we basically shooting ourselves int he foot...creating faster clock speeds at the expense of real-world performance?
Someone help me with this one because i thought I read that the Itanium thing from Intel is going in the reverse direction...lower clock speeds, but shorter pipeline and therefore better performance. Why would we then go in the opposite direction? Seems Intel is playing into our hands by reducing the focus on clock-speed...so why not take advantage of that fact?
Sure sounds good enough for me. No matter what happens (if something is released as a G5), we'll start a thread "Is this really a new class of chip or a G4 pretending to be a G5" and get the help of our fellow gossip hounds here. I can't hold you to your word, and you can't hold me to mine. One of us could always disagree so I think we let the EXPERTS here that have knowledge of chips be the deciders.
Tomorrow is Payday, I'm setting aside that 1st $13.
Yeah, but isn't there an inherent short-coming to making processors that use longer pipelines (a la the MHz myth animation thing Jobs did last July)? That is, with the shorter pipelines, current G4 chips greatly outpace Intel and AMD chips at similar clock speeds (and even moderately higher clock speeds), because they use much longer pipelines. So if we go in that direction with the G5, aren't we basically shooting ourselves int he foot...creating faster clock speeds at the expense of real-world performance?
This is true... to a point. But the G5 compensates by adding a bigger cache, faster bus, and more execution units and also by processing more per cycle. Look at it this way: you give some you lose some. I think over all the G5 will be at almost parity with the G4 when it comes to MHz for MHz processing power. The G5 will just have the high Mhz advantage.
The long vs short pipeline has to do with "branch prediction". That's when the code hits a branch in the code and has to do one of two things based on the out come of an "if" statement for example.
The CPU tries to guss which branch it will take and load it up for the pipeline. It has to guess because it may be waiting on the result of a proceding instrcution in the pipeline. If it guesses right then no problem. If it's wrong then it had to clear the pipe and load up the other branch. Longer pipes take longer to clear and load.
Intel overcomes that with fancy ass "branch prediction units" BPU that are better than most at gussing the correct branch. Of course that uses up more space on the CPU and cost more power and is larger and not very RISC like.
Apple/Mot/IBM deal with it by having shorter pipes.
<strong>When you say do more per cycle, do you mean in the sense of processing data in 64 bit chunks, or something else?</strong><hr></blockquote>
Think of washing dishes: You could have a 1-stage "pipeline," in which the entire process of washing a dish were done by one person in one sink. Contrast that with a 3 stage pipeline, where one person washed, one person rinsed, and one person dried. Each person would be doing (roughly) one third the work, and dishes could be washed (roughly) three times faster, because a new dish could be introduced into the pipeline in the time it took to complete one of the three tasks.
That metaphor doesn't work apart from illustrating what's meant by "doing more per cycle" - for instance, dishwashers will rarely run across a conditional branch.
That metaphor doesn't work apart from illustrating what's meant by "doing more per cycle" - for instance, dishwashers will rarely run across a conditional branch. </strong><hr></blockquote>
Well? Just write on something "If this is a cup dry the blue plate next." When the next thing is a fork the toss the fork and whatever the person washing has (lets say a spoon) in the trash. The only way to do that is to pass them down to the dryer. The the washer has to grab the blue plate and the rinser and dryer sit there and do nothing.
I think I follow what you two are saying. Sorta. So what then is the biggest reason for increasing the number of pipeline stages, beyond just clock speed? I mean, are we asking this processor to do so much more than the G4, that it *requires* more stages, or are we just trying to be good marketers?
AMD plans to release the Athlon "Hammer" in late 2002 (or early 2003). They say it will reach 1400 SpecInt2000 @ 2ghz - twice as fast as the current P4. Intel will probably have a 3ghz Pentium by then.
The (rumored) G5 specs are really great now, but will they still be up to date when the chip is finally shipping???
Are you sure that the white keyboard isn't just shipping with the iMac? The pages for pro mouse and keyboard still show the usual ones. It looks to me like they just modified the two to match the imac, but it doesn't really look too good, more like a quick fix. I would expect a redesigned keyboard even if the Power Mac went white. In other words, I see no indication that the keyboard and mouse shipping with the QS are white, and in fact it appears to be the opposite.</strong><hr></blockquote>
I reach my conclusions based on the fact that Apple has never shipped 2 different sets of keyboard and mouse. If they did or have, then yes, I can see the old style black/clear plastic remaining. But, because they have never had 2 sets, I think the current black/clear plastic is done and we will see a case to match the white keyboard & mouse.
I dont see that a couple interchangable plastic parts and a whole lotta keys are much more complex that several interchangable parts in 5 different translucent colored plastics.
I'm betting that the new PowerMac is gonna be pretty badass looking.
Chromy dark, dark cloudy plastics to match THE KEYBOARD of all things.
<strong>Sure sounds good enough for me. No matter what happens (if something is released as a G5), we'll start a thread "Is this really a new class of chip or a G4 pretending to be a G5" and get the help of our fellow gossip hounds here. I can't hold you to your word, and you can't hold me to mine. One of us could always disagree so I think we let the EXPERTS here that have knowledge of chips be the deciders.
Tomorrow is Payday, I'm setting aside that 1st $13.</strong><hr></blockquote>
Sounds like we're on. I'm putting a couple of bucks away even as we speak.....Now where is that damn shoe box.....
I reach my conclusions based on the fact that Apple has never shipped 2 different sets of keyboard and mouse. If they did or have, then yes, I can see the old style black/clear plastic remaining. But, because they have never had 2 sets, I think the current black/clear plastic is done and we will see a case to match the white keyboard & mouse.</strong><hr></blockquote>
Well either your memory is short, or you are too young, or you weren't paying attention in '89 through '94 or so. The Apple Standard and Apple Extended keyboards were both available. I wish this was still the case because the standard keyboards are too small for me.
Well either your memory is short, or you are too young, or you weren't paying attention in '89 through '94 or so. The Apple Standard and Apple Extended keyboards were both available. I wish this was still the case because the standard keyboards are too small for me.</strong><hr></blockquote>
Well, I am niether that young nor is my memory that short, simply my mac history has a gap during that period. I learned on a mac in HS about 88-89 and then didn't touch a computer until 97 or so. So, I had no idea that had offered 2 verisons of keyboards. OF course I notice you didn't mention the mouse. So, I am still partly right
Apple Engineers working on the 8500 chip...awhile now.It should be ready.They have the Technology,they will rebuild a Powerful WorkStation.It has to look different to differentiate it from the QuickSilver.It will probably be short and wide...like an oven of yesteryear.The hype about the G-5 is all over the news and the web.The first revision of the chip should be out in less than three months.I will probably be White.Sherwin-Williams probably gave Apple a contractors discount price.Lets all make a toast to the NeXT PowerMac.Its coming....
<strong>Yeah, but isn't there an inherent short-coming to making processors that use longer pipelines (a la the MHz myth animation thing Jobs did last July)? That is, with the shorter pipelines, current G4 chips greatly outpace Intel and AMD chips at similar clock speeds (and even moderately higher clock speeds), because they use much longer pipelines. So if we go in that direction with the G5, aren't we basically shooting ourselves int he foot...creating faster clock speeds at the expense of real-world performance?
This is true... to a point. But the G5 compensates by adding a bigger cache, faster bus, and more execution units and also by processing more per cycle. Look at it this way: you give some you lose some. I think over all the G5 will be at almost parity with the G4 when it comes to MHz for MHz processing power. The G5 will just have the high Mhz advantage.</strong><hr></blockquote>
I disagree -- I think the G5 will have a larger per-cycle advantage over the G4. If you think of a pipeline as making the chip "longer", then increasing the number of execution units (i.e. the number of pipelines) is making it "wider". The speed of the machine is the overall area, which is "length" times "width".
Increasing pipeline length means each stage of the pipeline does less, and thus is simpler. It is easier to make simpler things run at higher clock rates. The downside of longer pipelines is primarily the problem encountered at branches, as described above. If you have 20 stages and you guess a branch wrong, you have to toss out about 18 stages worth of instructions... and then the later stages have to wait for up to 18 cycles before the next instruction reaches them.
Increasing pipeline length can also run into dependency problems. If one instruction needs the results of a previous instruction, but the previous instruction hasn't gotten far enough through the pipeline to produce results, then a wait has to be introduced and you see a "bubble" in the pipe where the pipe stage is idle.
The PowerPC has always had a branch prediction unit as well, by the way. Even on shorter pipes it is useful. The longer your pipelines, the more circuitry its worth throwing at the prediction. Any branch prediction is based on what has happened before, so if your code always tends to branch the same way (looping over lots of data, for example) then you can predict it pretty well. If your decisions are essentially "random", however, the branch prediction doesn't work worth beans. I've seen the P4 fall over real bad on code like this.
Going "wider" (i.e. more execution units) means you try to get more instructions executing in parallel at the same time. The G4 has a pair of simple integer math units, a complex integer math unit, a branch unit, a load/store unit, a floating point unit, and a couple of vector units. The exact breakdown changed in the 7450 to try and counter-act the problems with longer pipes. It was largely successful as long as the code is compiled with the 7450 in mind.
My guess about the G5 is that they are going to throw lots of circuitry at making the chip both wider and longer, as well as optimizing how data moves between areas of the chip. If you have lots of transistors to throw around (and they should have at least double the 7450's), then you can start getting extravagent about where you use them. I'd be surprised to see longer than 10 pipe stages -- many of the Athlon's stages are just decoding the x86 instructions which doesn't take so much effort on the PPC. I hope to see another floating point execution unit or two, more integer units, and better load/store throughput.
Still not betting on when we'll see any of this though.
Comments
Someone help me with this one because i thought I read that the Itanium thing from Intel is going in the reverse direction...lower clock speeds, but shorter pipeline and therefore better performance. Why would we then go in the opposite direction? Seems Intel is playing into our hands by reducing the focus on clock-speed...so why not take advantage of that fact?
[ 01-10-2002: Message edited by: Moogs ? ]</p>
Tomorrow is Payday, I'm setting aside that 1st $13.
This is true... to a point. But the G5 compensates by adding a bigger cache, faster bus, and more execution units and also by processing more per cycle. Look at it this way: you give some you lose some. I think over all the G5 will be at almost parity with the G4 when it comes to MHz for MHz processing power. The G5 will just have the high Mhz advantage.
The CPU tries to guss which branch it will take and load it up for the pipeline. It has to guess because it may be waiting on the result of a proceding instrcution in the pipeline. If it guesses right then no problem. If it's wrong then it had to clear the pipe and load up the other branch. Longer pipes take longer to clear and load.
Intel overcomes that with fancy ass "branch prediction units" BPU that are better than most at gussing the correct branch. Of course that uses up more space on the CPU and cost more power and is larger and not very RISC like.
Apple/Mot/IBM deal with it by having shorter pipes.
[ 01-10-2002: Message edited by: Scott H. ]</p>
<strong>When you say do more per cycle, do you mean in the sense of processing data in 64 bit chunks, or something else?</strong><hr></blockquote>
Think of washing dishes: You could have a 1-stage "pipeline," in which the entire process of washing a dish were done by one person in one sink. Contrast that with a 3 stage pipeline, where one person washed, one person rinsed, and one person dried. Each person would be doing (roughly) one third the work, and dishes could be washed (roughly) three times faster, because a new dish could be introduced into the pipeline in the time it took to complete one of the three tasks.
That metaphor doesn't work apart from illustrating what's meant by "doing more per cycle" - for instance, dishwashers will rarely run across a conditional branch.
<strong>
That metaphor doesn't work apart from illustrating what's meant by "doing more per cycle" - for instance, dishwashers will rarely run across a conditional branch.
Well? Just write on something "If this is a cup dry the blue plate next." When the next thing is a fork the toss the fork and whatever the person washing has (lets say a spoon) in the trash. The only way to do that is to pass them down to the dryer. The the washer has to grab the blue plate and the rinser and dryer sit there and do nothing.
[ 01-10-2002: Message edited by: Scott H. ]</p>
The (rumored) G5 specs are really great now, but will they still be up to date when the chip is finally shipping???
<a href="http://www.architosh.com/news/2002-01/2002c-mwsf-postkeynote.phtml"; target="_blank">http://www.architosh.com/news/2002-01/2002c-mwsf-postkeynote.phtml</a>;
<strong>
Are you sure that the white keyboard isn't just shipping with the iMac? The pages for pro mouse and keyboard still show the usual ones. It looks to me like they just modified the two to match the imac, but it doesn't really look too good, more like a quick fix. I would expect a redesigned keyboard even if the Power Mac went white. In other words, I see no indication that the keyboard and mouse shipping with the QS are white, and in fact it appears to be the opposite.</strong><hr></blockquote>
I reach my conclusions based on the fact that Apple has never shipped 2 different sets of keyboard and mouse. If they did or have, then yes, I can see the old style black/clear plastic remaining. But, because they have never had 2 sets, I think the current black/clear plastic is done and we will see a case to match the white keyboard & mouse.
how about 5 different colors?
I dont see that a couple interchangable plastic parts and a whole lotta keys are much more complex that several interchangable parts in 5 different translucent colored plastics.
I'm betting that the new PowerMac is gonna be pretty badass looking.
Chromy dark, dark cloudy plastics to match THE KEYBOARD of all things.
I hope.
HAL9000, here we come.
<strong>Sure sounds good enough for me. No matter what happens (if something is released as a G5), we'll start a thread "Is this really a new class of chip or a G4 pretending to be a G5" and get the help of our fellow gossip hounds here. I can't hold you to your word, and you can't hold me to mine. One of us could always disagree so I think we let the EXPERTS here that have knowledge of chips be the deciders.
Tomorrow is Payday, I'm setting aside that 1st $13.</strong><hr></blockquote>
Sounds like we're on. I'm putting a couple of bucks away even as we speak.....Now where is that damn shoe box.....
SdC
shoe box is my vegas money, maybe I can stuff the money in my iomega zip 250, I never put disks in the damn thing anyway.
<strong>
I reach my conclusions based on the fact that Apple has never shipped 2 different sets of keyboard and mouse. If they did or have, then yes, I can see the old style black/clear plastic remaining. But, because they have never had 2 sets, I think the current black/clear plastic is done and we will see a case to match the white keyboard & mouse.</strong><hr></blockquote>
Well either your memory is short, or you are too young, or you weren't paying attention in '89 through '94 or so. The Apple Standard and Apple Extended keyboards were both available. I wish this was still the case because the standard keyboards are too small for me.
<strong>
Well either your memory is short, or you are too young, or you weren't paying attention in '89 through '94 or so. The Apple Standard and Apple Extended keyboards were both available. I wish this was still the case because the standard keyboards are too small for me.</strong><hr></blockquote>
Well, I am niether that young nor is my memory that short, simply my mac history has a gap during that period. I learned on a mac in HS about 88-89 and then didn't touch a computer until 97 or so. So, I had no idea that had offered 2 verisons of keyboards. OF course I notice you didn't mention the mouse. So, I am still partly right
Apple Engineers working on the 8500 chip...awhile now.It should be ready.They have the Technology,they will rebuild a Powerful WorkStation.It has to look different to differentiate it from the QuickSilver.It will probably be short and wide...like an oven of yesteryear.The hype about the G-5 is all over the news and the web.The first revision of the chip should be out in less than three months.I will probably be White.Sherwin-Williams probably gave Apple a contractors discount price.Lets all make a toast to the NeXT PowerMac.Its coming....
<strong>Yeah, but isn't there an inherent short-coming to making processors that use longer pipelines (a la the MHz myth animation thing Jobs did last July)? That is, with the shorter pipelines, current G4 chips greatly outpace Intel and AMD chips at similar clock speeds (and even moderately higher clock speeds), because they use much longer pipelines. So if we go in that direction with the G5, aren't we basically shooting ourselves int he foot...creating faster clock speeds at the expense of real-world performance?
This is true... to a point. But the G5 compensates by adding a bigger cache, faster bus, and more execution units and also by processing more per cycle. Look at it this way: you give some you lose some. I think over all the G5 will be at almost parity with the G4 when it comes to MHz for MHz processing power. The G5 will just have the high Mhz advantage.</strong><hr></blockquote>
I disagree -- I think the G5 will have a larger per-cycle advantage over the G4. If you think of a pipeline as making the chip "longer", then increasing the number of execution units (i.e. the number of pipelines) is making it "wider". The speed of the machine is the overall area, which is "length" times "width".
Increasing pipeline length means each stage of the pipeline does less, and thus is simpler. It is easier to make simpler things run at higher clock rates. The downside of longer pipelines is primarily the problem encountered at branches, as described above. If you have 20 stages and you guess a branch wrong, you have to toss out about 18 stages worth of instructions... and then the later stages have to wait for up to 18 cycles before the next instruction reaches them.
Increasing pipeline length can also run into dependency problems. If one instruction needs the results of a previous instruction, but the previous instruction hasn't gotten far enough through the pipeline to produce results, then a wait has to be introduced and you see a "bubble" in the pipe where the pipe stage is idle.
The PowerPC has always had a branch prediction unit as well, by the way. Even on shorter pipes it is useful. The longer your pipelines, the more circuitry its worth throwing at the prediction. Any branch prediction is based on what has happened before, so if your code always tends to branch the same way (looping over lots of data, for example) then you can predict it pretty well. If your decisions are essentially "random", however, the branch prediction doesn't work worth beans. I've seen the P4 fall over real bad on code like this.
Going "wider" (i.e. more execution units) means you try to get more instructions executing in parallel at the same time. The G4 has a pair of simple integer math units, a complex integer math unit, a branch unit, a load/store unit, a floating point unit, and a couple of vector units. The exact breakdown changed in the 7450 to try and counter-act the problems with longer pipes. It was largely successful as long as the code is compiled with the 7450 in mind.
My guess about the G5 is that they are going to throw lots of circuitry at making the chip both wider and longer, as well as optimizing how data moves between areas of the chip. If you have lots of transistors to throw around (and they should have at least double the 7450's), then you can start getting extravagent about where you use them. I'd be surprised to see longer than 10 pipe stages -- many of the Athlon's stages are just decoding the x86 instructions which doesn't take so much effort on the PPC. I hope to see another floating point execution unit or two, more integer units, and better load/store throughput.
Still not betting on when we'll see any of this though.