it's much cheaper and still very fast so spend that $50 (?) 'saved' on more ram or a proper graphics chip and you'll end up with a very nice mini/ibook...
In terms of bang per buck, a Celeron D is even better. That's not the type of metric Apple wants to use in this context, however. They want energy-efficient machines. They stated quite clearly that this was the primary reason to switch to Intel.
Aside from the fact that there isn't officially a "line-up" of Yonah Celeron Ms anyway (at least I still have yet to see one single bit of confirmation), the data I've heard is 1 MB L2 cache and 133/533 MHz bus, which sounds like a Dothan Pentium M with half the cache.
Compare that to a Core Solo with 2 MB cache and a 167/667 MHz bus, and I don't see how this mystical Yonah Celeron M could in any way be "better". Higher clock rate options? Perhaps, but unless they're really very significantly higher, I still don't see how they could compete. Not to mention it's probably not even pin-compatible.
It's a very, very low-cost CPU, and I don't see any reason for Apple to use it. They might as well put used Pentium Ms in there.
It's NOT a lower clock rate, it's higher. 233Mhz higher than the Core Solo in the Mini. Higher than the Core Duo model even.
1MB or 2MB cache probably makes no difference at all on a single core chip, even on an Intel which need larger caches. PPC970(FX) G5's have 512K L2, the 970MP 1MB. 2MB on a single core CPU is good for willy waving and little else.
133Mhz drop in FSB makes almost no difference either.
We're talking iBooks right? Not the sort of thing that has to shove huge amounts of data about. Saving half your CPU budget for about the same performance is pretty useful.
Check this thread out on the Monty Hall benchmark. Granted it's only one test but perhaps it's not farfetched to see ICD machines keeping up with quad g5s in many tasks or uses.
It's NOT a lower clock rate, it's higher. 233Mhz higher than the Core Solo in the Mini. Higher than the Core Duo model even.
I was talking about the FSB.
In fact, you're providing a very good reason for Apple not to do it: marketing-wise, it's not good to have a low-end model with clock rates higher than more expensive models.
Quote:
1MB or 2MB cache probably makes no difference at all on a single core chip,
Er, you tell Intel that. And have a good look at the Pentium M, too.
Quote:
133Mhz drop in FSB makes almost no difference either.
More to the point, which Intel chips have the performance of the G5.
I really don't care about watts, I want performance!
Amen!
But, the Yonah is closer than we ever thought it would be. Merom will probably equal it, or better. Conroe will surpass it, and Woodcrest will race by.
Er, you tell Intel that. And have a good look at the Pentium M, too.
Oh, I dunno, it's only a third.
Have you seen tests that showed that the FSB or cache differences in the "-M" Intel chips made a significant performance difference? Except for maybe heavy compute stuff, the difference doesn't seem to matter:
This test showed a 1.3GHz Celeron-M with 1MB L2/400MHz FSB just about matching the compute performance and memory bandwidth of a 1.4GHz Pentium-M with 2MB L2 / 533MHz FSB.
The only significant down side in real-world use for Celeron-M was the greatly reduced flexibility in the power savings measures.
Have you seen tests that showed that the FSB or cache differences in the "-M" Intel chips made a significant performance difference? Except for maybe heavy compute stuff, the difference doesn't seem to matter:
Shhhh. The chip snobs will have you against the wall if you start disseminating non-geek opinions like that.
Check this thread out on the Monty Hall benchmark. Granted it's only one test but perhaps it's not farfetched to see ICD machines keeping up with quad g5s in many tasks or uses.
ICD keeping up with quad g5s is quite feasible, given the right optimisations, I reckon. But Apple/others are probably still working on it. According to Anandtech's benchmarks, one Core performs as well as a G4. The dual core probably does g5-level performance. (although Anandtech has not done g5-level comparisons )
ICD keeping up with quad g5s is quite feasible, given the right optimisations, I reckon. But Apple/others are probably still working on it. According to Anandtech's benchmarks, one Core performs as well as a G4. The dual core probably does g5-level performance. (although Anandtech has not done g5-level comparisons )
Well one thing that MArvin mentioned in another thread is that the Quad is rarely using all it's resources. Most the time it's only using 50%. Therefore it behaves like a dualie. Makes me wonder how great the Woodcrest will actually be. It may not be twice as fast as Conroe.
Well one thing that MArvin mentioned in another thread is that the Quad is rarely using all it's resources. Most the time it's only using 50%. Therefore it behaves like a dualie. Makes me wonder how great the Woodcrest will actually be. It may not be twice as fast as Conroe.
Unfortunately, many programs will use two processors, one core each, or possibly (I'm not sure about this) two cores on one processor.
They aren't really MULTI cpu aware, just two.
I was surprised to find that PS only went about twice as fast on the quad as it did on a single core single cpu machine. But some other programs went four times as fast.
Unfortunately, many programs will use two processors, one core each, or possibly (I'm not sure about this) two cores on one processor.
They aren't really MULTI cpu aware, just two.
I was surprised to find that PS only went about twice as fast on the quad as it did on a single core single cpu machine. But some other programs went four times as fast.
Perhaps with current software dual core is the best bang for the buck. After that you may be better off going for more ghz.
Comments
Originally posted by tubgirl
better in a bang/$ way...
it's much cheaper and still very fast so spend that $50 (?) 'saved' on more ram or a proper graphics chip and you'll end up with a very nice mini/ibook...
In terms of bang per buck, a Celeron D is even better. That's not the type of metric Apple wants to use in this context, however. They want energy-efficient machines. They stated quite clearly that this was the primary reason to switch to Intel.
Originally posted by Chucker
Aside from the fact that there isn't officially a "line-up" of Yonah Celeron Ms anyway (at least I still have yet to see one single bit of confirmation), the data I've heard is 1 MB L2 cache and 133/533 MHz bus, which sounds like a Dothan Pentium M with half the cache.
Compare that to a Core Solo with 2 MB cache and a 167/667 MHz bus, and I don't see how this mystical Yonah Celeron M could in any way be "better". Higher clock rate options? Perhaps, but unless they're really very significantly higher, I still don't see how they could compete. Not to mention it's probably not even pin-compatible.
See http://www.reghardware.co.uk/2006/04...leron_ms_ship/
The Celeron-M 430 is 1.73Ghz as opposed to the Core Solo in the Mini at 1.5Ghz.
The Celeron-M 440 ships end of this year at 1.86Ghz.
They're about half the price of the Core Solo.
Originally posted by aegisdesign
The Celeron-M 430 is 1.73Ghz as opposed to the Core Solo in the Mini at 1.5Ghz.
The Celeron-M 440 ships end of this year at 1.86Ghz.
They're about half the price of the Core Solo.
Again: half the cache, lower clockrate, etc.
It's a very, very low-cost CPU, and I don't see any reason for Apple to use it. They might as well put used Pentium Ms in there.
Originally posted by melgross
Performance per awtt is alive and well. Which PPC chips have the performance the dual core Yonahs have, at anyway near the same watt levels?
[/B]
More to the point, which Intel chips have the performance of the G5.
I really don't care about watts, I want performance!
Originally posted by aegisdesign
More to the point, which Intel chips have the performance of the G5.
I really don't care about watts, I want performance!
ICD. Some posters on threads here feel their ICD is faster than their Quad g5.
Originally posted by Chucker
Again: half the cache, lower clockrate, etc.
It's a very, very low-cost CPU, and I don't see any reason for Apple to use it. They might as well put used Pentium Ms in there.
It's NOT a lower clock rate, it's higher. 233Mhz higher than the Core Solo in the Mini. Higher than the Core Duo model even.
1MB or 2MB cache probably makes no difference at all on a single core chip, even on an Intel which need larger caches. PPC970(FX) G5's have 512K L2, the 970MP 1MB. 2MB on a single core CPU is good for willy waving and little else.
133Mhz drop in FSB makes almost no difference either.
We're talking iBooks right? Not the sort of thing that has to shove huge amounts of data about. Saving half your CPU budget for about the same performance is pretty useful.
Originally posted by backtomac
ICD. Some posters on threads here feel their ICD is faster than their Quad g5.
'feel'?
Originally posted by aegisdesign
'feel'?
Hey I'm only relaying what they have said. Speaker said he is a professional video editor who now is doing all his editing on his MBP not his quad g5.
http://forums.appleinsider.com/showt...threadid=62870
Check this thread out on the Monty Hall benchmark. Granted it's only one test but perhaps it's not farfetched to see ICD machines keeping up with quad g5s in many tasks or uses.
http://forums.appleinsider.com/showt...threadid=60684
Originally posted by aegisdesign
It's NOT a lower clock rate, it's higher. 233Mhz higher than the Core Solo in the Mini. Higher than the Core Duo model even.
I was talking about the FSB.
In fact, you're providing a very good reason for Apple not to do it: marketing-wise, it's not good to have a low-end model with clock rates higher than more expensive models.
1MB or 2MB cache probably makes no difference at all on a single core chip,
Er, you tell Intel that. And have a good look at the Pentium M, too.
133Mhz drop in FSB makes almost no difference either.
Oh, I dunno, it's only a third.
Originally posted by aegisdesign
More to the point, which Intel chips have the performance of the G5.
I really don't care about watts, I want performance!
Amen!
But, the Yonah is closer than we ever thought it would be. Merom will probably equal it, or better. Conroe will surpass it, and Woodcrest will race by.
Originally posted by Chucker
Er, you tell Intel that. And have a good look at the Pentium M, too.
Oh, I dunno, it's only a third.
Have you seen tests that showed that the FSB or cache differences in the "-M" Intel chips made a significant performance difference? Except for maybe heavy compute stuff, the difference doesn't seem to matter:
Celeron-M vs Pentium-M per-clock perf (non) difference
This test showed a 1.3GHz Celeron-M with 1MB L2/400MHz FSB just about matching the compute performance and memory bandwidth of a 1.4GHz Pentium-M with 2MB L2 / 533MHz FSB.
The only significant down side in real-world use for Celeron-M was the greatly reduced flexibility in the power savings measures.
Originally posted by JeffDM
Have you seen tests that showed that the FSB or cache differences in the "-M" Intel chips made a significant performance difference? Except for maybe heavy compute stuff, the difference doesn't seem to matter:
Shhhh. The chip snobs will have you against the wall if you start disseminating non-geek opinions like that.
Originally posted by JeffDM
The only significant down side in real-world use for Celeron-M was the greatly reduced flexibility in the power savings measures.
yes - and afiak the speed stepping feature is availible on the new yonah-based celeron-m. ('enhanced speedstep technology' or whatever...)
The playboy iBook, ideal for the educa.... OK NEVERMIND
Originally posted by fooey
maybe when they say fashionable hues they are referring to Hue Hefner.
The playboy iBook, ideal for the educa.... OK NEVERMIND
It'd have the bunny logo on the top.
Originally posted by backtomac
Hey I'm only relaying what they have said. Speaker said he is a professional video editor who now is doing all his editing on his MBP not his quad g5.
http://forums.appleinsider.com/showt...threadid=62870
Check this thread out on the Monty Hall benchmark. Granted it's only one test but perhaps it's not farfetched to see ICD machines keeping up with quad g5s in many tasks or uses.
http://forums.appleinsider.com/showt...threadid=60684
ICD keeping up with quad g5s is quite feasible, given the right optimisations, I reckon. But Apple/others are probably still working on it. According to Anandtech's benchmarks, one Core performs as well as a G4. The dual core probably does g5-level performance. (although Anandtech has not done g5-level comparisons )
Originally posted by sunilraman
ICD keeping up with quad g5s is quite feasible, given the right optimisations, I reckon. But Apple/others are probably still working on it. According to Anandtech's benchmarks, one Core performs as well as a G4. The dual core probably does g5-level performance. (although Anandtech has not done g5-level comparisons )
Well one thing that MArvin mentioned in another thread is that the Quad is rarely using all it's resources. Most the time it's only using 50%. Therefore it behaves like a dualie. Makes me wonder how great the Woodcrest will actually be. It may not be twice as fast as Conroe.
Originally posted by backtomac
Well one thing that MArvin mentioned in another thread is that the Quad is rarely using all it's resources. Most the time it's only using 50%. Therefore it behaves like a dualie. Makes me wonder how great the Woodcrest will actually be. It may not be twice as fast as Conroe.
Unfortunately, many programs will use two processors, one core each, or possibly (I'm not sure about this) two cores on one processor.
They aren't really MULTI cpu aware, just two.
I was surprised to find that PS only went about twice as fast on the quad as it did on a single core single cpu machine. But some other programs went four times as fast.
Originally posted by melgross
Unfortunately, many programs will use two processors, one core each, or possibly (I'm not sure about this) two cores on one processor.
They aren't really MULTI cpu aware, just two.
I was surprised to find that PS only went about twice as fast on the quad as it did on a single core single cpu machine. But some other programs went four times as fast.
Perhaps with current software dual core is the best bang for the buck. After that you may be better off going for more ghz.