Re: Re: Re: G5 heat/PB FAQ & Summary
Click on his username and then search for all his posts. It's easier than way.
Click on his username and then search for all his posts. It's easier than way.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
New PowerBooks on Tuesday.
- Thread starter MacRumors
- Start date
- Sort by reaction score
Right. More directly, if your goal was to save power you would take the fastest chips and clock them as slow as possible.
Chip speed has a number of factors figuring in-- among them are process (and intra-wafer) variation and operating voltage. To get minimum power, you'd find chips with the fastest process parameters, reduce their operating frequency and probably reduce the voltage to match.
Re: Mobile G5?
Sigh. Okay, how about "a processor more suitable for mobile applications" then? The 7457 G4 as opposed to a 7455 G4. A 90nm G5 as opposed to a 130nm G5. Yes, the current 970 G5 (130nm) has speed slewing but... oh forget it. You're right, the will be G5 PowerBooks tomorrow no doubt about it!
Actually, if it were'nt for these 7457 G4 chips I wouldn't even been wondering about it. They could very easily use slower-clocked G5s in the new PowerBooks and it would even explain the loooong wait we have had. But its those damn 7457 G4 chips! I just don't want to get my hopes up.
Sigh. Okay, how about "a processor more suitable for mobile applications" then? The 7457 G4 as opposed to a 7455 G4. A 90nm G5 as opposed to a 130nm G5. Yes, the current 970 G5 (130nm) has speed slewing but... oh forget it. You're right, the will be G5 PowerBooks tomorrow no doubt about it!
Actually, if it were'nt for these 7457 G4 chips I wouldn't even been wondering about it. They could very easily use slower-clocked G5s in the new PowerBooks and it would even explain the loooong wait we have had. But its those damn 7457 G4 chips! I just don't want to get my hopes up.
Re: Re: Re: Heat and noise (human, not computer)
I don't know, I might want to buy a G5 just for "mine is bigger than yours" reasons. But from a practical standpoint, you might be right.
Then again, a G5, even at low clock speeds, allows a higher frequency bus and true use of DDR RAM, which would be an incredible speed boost on its own. But then again, I'm probably just rationalizing my desire for a G5 PowerBook just for "mine is bigger than yours" reasons, which, let's face it, are a strong motivation for most men.
I don't know, I might want to buy a G5 just for "mine is bigger than yours" reasons. But from a practical standpoint, you might be right.
Then again, a G5, even at low clock speeds, allows a higher frequency bus and true use of DDR RAM, which would be an incredible speed boost on its own. But then again, I'm probably just rationalizing my desire for a G5 PowerBook just for "mine is bigger than yours" reasons, which, let's face it, are a strong motivation for most men.
Re: Re: Re: Re: my 2cents
IIRC, at least Moto produces desktop/mobile variants. "7447" is the upcoming mobile versoin of the "7457" desktop chip. The main difference is footprint size, as well as dropping of L3 cache support.
http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=MPC7447
http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=MPC7457
Note that both processors come in 1.0V low-voltage/low-power variants, and have identical power consumption characteristics (at least to the level of detail Moto provides); the ONLY advantage of the "mobile" version is the physical size.
I haven't heard of IBM planning a mobile version of the 970. Yes, a die shrink is good for mobile processors (reducing power consumption of current frequencies and reducing footprint), but such a move would go across the 970 line, not just for "mobile" procs (unless there were severe production difficulties).
IIRC, at least Moto produces desktop/mobile variants. "7447" is the upcoming mobile versoin of the "7457" desktop chip. The main difference is footprint size, as well as dropping of L3 cache support.
http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=MPC7447
http://e-www.motorola.com/webapp/sps/site/prod_summary.jsp?code=MPC7457
Note that both processors come in 1.0V low-voltage/low-power variants, and have identical power consumption characteristics (at least to the level of detail Moto provides); the ONLY advantage of the "mobile" version is the physical size.
I haven't heard of IBM planning a mobile version of the 970. Yes, a die shrink is good for mobile processors (reducing power consumption of current frequencies and reducing footprint), but such a move would go across the 970 line, not just for "mobile" procs (unless there were severe production difficulties).
Re: Re: Re: Re: Heat and noise (human, not computer)
Not every task is memory bound. True DDR would certainly help in memory bound tasks, but all the DDR speed in the world can't help a CPU bound task.
The point is that the G5 laptop would not give enough of a speed boost across the board to make it the choice now.
You'd be falling over yourself in a hurry to get the next iteration of G5 laptops, the ones that allow the true speed of the G5 to shine. Do you really want a neutered G5 laptop?
I'm waiting for the "real" G5 laptops. I got my 12" PB a couple months ago and I'm already planning on selling it come next year when we're looking at the real deal G5 PBs.
Not every task is memory bound. True DDR would certainly help in memory bound tasks, but all the DDR speed in the world can't help a CPU bound task.
The point is that the G5 laptop would not give enough of a speed boost across the board to make it the choice now.
You'd be falling over yourself in a hurry to get the next iteration of G5 laptops, the ones that allow the true speed of the G5 to shine. Do you really want a neutered G5 laptop?
I'm waiting for the "real" G5 laptops. I got my 12" PB a couple months ago and I'm already planning on selling it come next year when we're looking at the real deal G5 PBs.
Re: Re: Re: Re: Re: my 2cents
1) The 7457 line completely replaces the 7455 line according to Moto.
2) The 90nm line of G5s will likewise completely replace the 130nm line of G5s.
In order to have desktop/mopb ile variants, you must have both lines being advanced (updated/refined) at the same time. Otherwise, one is generally a replacement of the other.
1) The 7457 line completely replaces the 7455 line according to Moto.
2) The 90nm line of G5s will likewise completely replace the 130nm line of G5s.
In order to have desktop/mopb ile variants, you must have both lines being advanced (updated/refined) at the same time. Otherwise, one is generally a replacement of the other.
hiding something in paris
I suppose you've all seen this thread by now.
Even though the argument seems to have been "new G5 powerbooks tomorrow" vs. "new G4 powerbooks tomorrow", I hadn't been convinced that it won't be "no new powerbooks tomorrow."
But the locked cabinets and black covers in Paris are making that third possibility look more remote. Enough that I just ordered a stock 1GHz 15" TiBook from the Apple Store in hopes of getting an AlBook instead. Wish me luck.
I suppose you've all seen this thread by now.
Even though the argument seems to have been "new G5 powerbooks tomorrow" vs. "new G4 powerbooks tomorrow", I hadn't been convinced that it won't be "no new powerbooks tomorrow."
But the locked cabinets and black covers in Paris are making that third possibility look more remote. Enough that I just ordered a stock 1GHz 15" TiBook from the Apple Store in hopes of getting an AlBook instead. Wish me luck.
Re: Re: Heat and noise (human, not computer)
I expect that with the G5, we're going to see a larger gap between PowerMac and PowerBook performance than we've seen with the G4. IMO, the only reason that the G4 PBooks and PMacs have been as close as they have (not that they're really *that* close) is simply that the G4 wouldn't scale up to clock speeds where the power consumption became unreasonable for a notebook. The G5 has a lot of room to scale to clock speeds and power consumption levels that aren't going to work in a notebook; The PowerMacs will be able to use G5s at those speeds, and the PowerBooks won't.
So, sure, we might see a 90nm ~1.8 GHz 970 in the PowerBook early/middle of next year, but by that time, the G5 PowerMac will have taken a step forward, as well, and quite probably a bigger one. Remember Apple's promise to have 3.0GHz G5s within a year?
I expect that with the G5, we're going to see a larger gap between PowerMac and PowerBook performance than we've seen with the G4. IMO, the only reason that the G4 PBooks and PMacs have been as close as they have (not that they're really *that* close) is simply that the G4 wouldn't scale up to clock speeds where the power consumption became unreasonable for a notebook. The G5 has a lot of room to scale to clock speeds and power consumption levels that aren't going to work in a notebook; The PowerMacs will be able to use G5s at those speeds, and the PowerBooks won't.
So, sure, we might see a 90nm ~1.8 GHz 970 in the PowerBook early/middle of next year, but by that time, the G5 PowerMac will have taken a step forward, as well, and quite probably a bigger one. Remember Apple's promise to have 3.0GHz G5s within a year?
Re: Re: Mobile G5?
Well...all of Moto's chips are ACTUALLY designed for embedded use. So that's why they don't need a "mobile" counterpart. The IBM PPC 970 was designed:
A. For Apple
B. For IBM's High-Density Blade Servers
Please note that that is in no particular order except I thought it was funny to post "A is for Apple" :| I'm betting that IBM was partly willing to do all this chip R&D only because they knew they'd be able to use the chips for blades, that would make selling parts to apple a secondary motivation. And if IBM thought the chip was a good idea for a high-density blade, well, a powerbook is a MUCH more forgiving thermal/space environment.
Well...all of Moto's chips are ACTUALLY designed for embedded use. So that's why they don't need a "mobile" counterpart. The IBM PPC 970 was designed:
A. For Apple
B. For IBM's High-Density Blade Servers
Please note that that is in no particular order except I thought it was funny to post "A is for Apple" :| I'm betting that IBM was partly willing to do all this chip R&D only because they knew they'd be able to use the chips for blades, that would make selling parts to apple a secondary motivation. And if IBM thought the chip was a good idea for a high-density blade, well, a powerbook is a MUCH more forgiving thermal/space environment.
Re: ????
Overall case temperature directly affects the CPU's temperature, and vice versa.
Nothing in the PowerMac exists in a vacuum. If you think the other "thermal zones" have no bearing on the CPU temp, I encourage you to try yanking the extra fans out and monitoring your CPU temperature.
While you're at it, stuff something into the case to take up all that empty space. There's your G5 PowerBook. Ignore the sounds of melting.
Overall case temperature directly affects the CPU's temperature, and vice versa.
Nothing in the PowerMac exists in a vacuum. If you think the other "thermal zones" have no bearing on the CPU temp, I encourage you to try yanking the extra fans out and monitoring your CPU temperature.
While you're at it, stuff something into the case to take up all that empty space. There's your G5 PowerBook. Ignore the sounds of melting.
Put a couple of 8" squirrel cage fans in your PowerBook!
Actually, the BladeCenter is a much more forgiving environment.
The chassis has 4800 watt power supplies (4 * 1200 watt), and two large squirrel-cage blowers, plus network and fibre channel and management controllers. The chassis weighs 123kg and is 31 x 45 x 71 cm - about 37 times larger than a 17" PB.
The blade is 3cm x 24cm x 44cm - or about 20% larger by volume than a 17" PowerBook. It doesn't need to fit CD, screen, keyboard, PC card slots, 802.11 card, fans or external ports in that space.
It's designed for the climate-controlled computer room - no worries about sitting on a lap or a bed.
Go to http://www-132.ibm.com/webapp/wcs/s...Id=2575417&storeId=1&catalogId=-840&langId=-1 and click on "Visual Tour" just under the picture - you'll be able to see how much space.
Actually, the BladeCenter is a much more forgiving environment.
The chassis has 4800 watt power supplies (4 * 1200 watt), and two large squirrel-cage blowers, plus network and fibre channel and management controllers. The chassis weighs 123kg and is 31 x 45 x 71 cm - about 37 times larger than a 17" PB.
The blade is 3cm x 24cm x 44cm - or about 20% larger by volume than a 17" PowerBook. It doesn't need to fit CD, screen, keyboard, PC card slots, 802.11 card, fans or external ports in that space.
It's designed for the climate-controlled computer room - no worries about sitting on a lap or a bed.
Go to http://www-132.ibm.com/webapp/wcs/s...Id=2575417&storeId=1&catalogId=-840&langId=-1 and click on "Visual Tour" just under the picture - you'll be able to see how much space.
Re: Re: ????
Hmmm, I seem to have made no mention of the case existing as a vacuum or that temperature in other compartments didn't matter. There are divisions in the case and though overall case temperature does affect the processor, the 9 fans' primary purpose is not for cooling the processor, as I have already stated. While it's true that removing all the other fans would have an impact on the system and CPU temp that is NOT the argument here. The argument is based on the heat generated by the processor alone. Thus to truly get an accurate picture you'd have to take out the RAM and HDs which are also MAJOR sources of heat, and thus, yes, the G5 could be cooled with just that single fan. Also, if one high RPM fan can cool a 3.0 GHz P4 or an any Athlon XP one low RPM fan can certainly cool a processor like the G5 which has power requirements nearly half that of a P4 and a larger heatsink....and I never said the G5 Powerbook was coming, I've said it was possible but that I doubt that Apple has fully finished it, or if they've just decided to wait for the 90nm process (though I suppose they'd be doing MB development for that now if IBM could offer them any data on it).
Hmmm, I seem to have made no mention of the case existing as a vacuum or that temperature in other compartments didn't matter. There are divisions in the case and though overall case temperature does affect the processor, the 9 fans' primary purpose is not for cooling the processor, as I have already stated. While it's true that removing all the other fans would have an impact on the system and CPU temp that is NOT the argument here. The argument is based on the heat generated by the processor alone. Thus to truly get an accurate picture you'd have to take out the RAM and HDs which are also MAJOR sources of heat, and thus, yes, the G5 could be cooled with just that single fan. Also, if one high RPM fan can cool a 3.0 GHz P4 or an any Athlon XP one low RPM fan can certainly cool a processor like the G5 which has power requirements nearly half that of a P4 and a larger heatsink....and I never said the G5 Powerbook was coming, I've said it was possible but that I doubt that Apple has fully finished it, or if they've just decided to wait for the 90nm process (though I suppose they'd be doing MB development for that now if IBM could offer them any data on it).
Re: ????
True. So the 9 fans thing has little to do with the processor itself, but the architecture. Then again, I anticipate the G5 architecture to be put into the PowerBook.
And it's not just the DDR RAM (although that's a big part of it), it's the architecture in general that I want. And it's the architecture in general that we might see tomorrow, or maybe not.
True. So the 9 fans thing has little to do with the processor itself, but the architecture. Then again, I anticipate the G5 architecture to be put into the PowerBook.
And it's not just the DDR RAM (although that's a big part of it), it's the architecture in general that I want. And it's the architecture in general that we might see tomorrow, or maybe not.
7457 Specs
Is it just me, or does Motorola's site say the 7457 dissipates only 10 Watts at 1 GHz??? That's awesome!!
And, with a 1.3 GHz 7457 processor, with 512k L2 Cache and 2 MB of L3 cache.. I think we'd have one FAST-AS-HELL laptop.. especially with Panther (compared to Jaguar and previous-gen G4's at least..).
I'm getting excited about tomorrow. (assuming the rumors are true...)
Is it just me, or does Motorola's site say the 7457 dissipates only 10 Watts at 1 GHz??? That's awesome!!
And, with a 1.3 GHz 7457 processor, with 512k L2 Cache and 2 MB of L3 cache.. I think we'd have one FAST-AS-HELL laptop.. especially with Panther (compared to Jaguar and previous-gen G4's at least..).
I'm getting excited about tomorrow. (assuming the rumors are true...)
Re: 7457 Specs
That actually makes a dual somewhat feasible. If a 19W G5 is feasible, so are 2 10W G4's adding up to 20 W. And you don't need the G5 architecture, which is where the real problems start to come in terms of heat and miniaturization.
That actually makes a dual somewhat feasible. If a 19W G5 is feasible, so are 2 10W G4's adding up to 20 W. And you don't need the G5 architecture, which is where the real problems start to come in terms of heat and miniaturization.
At the end of the day, it sometimes nice just to simply forget what's under the hood and realize that the new PB's are going to be fantastic machines!!Originally posted by cb911
now i'm kind of hoping that the new PowerBooks will use the 7457!
For the record: My only regret about the G4 chips has been the inability to scale the bus. The 7457 is going to be a superb chip in all ways other than the lack of DDR memory support and the fact that the bus can't be dual- (or quad- like the current Intel and AMD chips) pumped. If Moto can really produce chips as fast as they told powerlogix (1.8 I think was the highest #) then hell yes that would be great for a PB. But they'll need DDR support. The name of the game is becoming bandwidth as clock speed start getting ridiculous, we'll need to feed those chips!
Pricing lists
On my Apple retailer's pricing list, they haven't listed the new updated iMac or the iPod on it I was wondering if this was because they knew that powerbooks would be coming at the Apple Expo, laziness, or simply because they do not have either of the products in stock. Let's hope for the first one.
On my Apple retailer's pricing list, they haven't listed the new updated iMac or the iPod on it I was wondering if this was because they knew that powerbooks would be coming at the Apple Expo, laziness, or simply because they do not have either of the products in stock. Let's hope for the first one.
Re: Re: Re: ????
Actually, that is the argument here. Looking at the CPU in isolation give no indication of a systems feasibility.
For example, if you look at the G4 alone you might think it performs much better than it does because you might assume a higher performance memory subsystem.
If you neglect the power of the other G5 system components you'll either wind up with a box that's too hot or too slow...
Actually, that is the argument here. Looking at the CPU in isolation give no indication of a systems feasibility.
For example, if you look at the G4 alone you might think it performs much better than it does because you might assume a higher performance memory subsystem.
If you neglect the power of the other G5 system components you'll either wind up with a box that's too hot or too slow...
Re: 7457 Specs
Finally, someone with a reasonable attitude. The G4 is NOT a slouch when compared to other processors running at the same speed. It consistently rates about 50% faster than a P4 (although I don't know how it compares to the new M series chips). Anyway, I will be very surprised if it is more than 20% slower than a G5 at the same clock speed.
Furthermore, the other elements of the laptop, like disk speed, etc., tend to moderate the effect of differences in chip processing power.
You know, it's like the cost of gasoline. The price of crude oil goes up to $30 a barrel, doesn't mean the price at the pump goes up exactly that much; because the truth is, the cost of the gasoline itself is only a small percentage of the total cost (check out how much you are paying in taxes, subtract profit and refining costs, etc.)
So what I'm saying is a 1.25 mhz Powerbook G4, especially with the new low power Moto chip, is likely to be every bit as good as a 1.0 or maybe even 1.2 ghz G5, and probably will hold its own very well against those new Pentium laptops with the M chip, what are they these days, 1.4 ghz?
Finally, someone with a reasonable attitude. The G4 is NOT a slouch when compared to other processors running at the same speed. It consistently rates about 50% faster than a P4 (although I don't know how it compares to the new M series chips). Anyway, I will be very surprised if it is more than 20% slower than a G5 at the same clock speed.
Furthermore, the other elements of the laptop, like disk speed, etc., tend to moderate the effect of differences in chip processing power.
You know, it's like the cost of gasoline. The price of crude oil goes up to $30 a barrel, doesn't mean the price at the pump goes up exactly that much; because the truth is, the cost of the gasoline itself is only a small percentage of the total cost (check out how much you are paying in taxes, subtract profit and refining costs, etc.)
So what I'm saying is a 1.25 mhz Powerbook G4, especially with the new low power Moto chip, is likely to be every bit as good as a 1.0 or maybe even 1.2 ghz G5, and probably will hold its own very well against those new Pentium laptops with the M chip, what are they these days, 1.4 ghz?