Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
As I've told you, I've played with processors that use Turbo Boost for years. I've already SEEN it at work. I've SEEN it go from it's turbo frequencies down to it's base clocks (using CPUZ on a PC)
...

Terrific. This entire time, I've been arguing that the numbers are irrelevant specifically for current MBAs which have way more cooling than Intel specifies.

I don't see how your experience with some other computers from years ago with different CPUs and different cooling is relevant to this conversation.

You want Intel to advertise their speeds based on nothing more than your own anecdotal evidence instead of how the CPU actually works. That's really what this boils down to.

Never, ever, ever said that, because it would be ridiculous.

I want Apple (not Intel) to list speeds that customers are likely to see to avoid customer confusion, like the confusion that started this thread.

They are doing a little better now vs. before. I just checked the Apple Store page and the Turbo Boost speed is listed in the main list of specifications, which is good. This wasn't always the case.

If you go to "About this Mac," no mention of turbo boost is made.
 
Terrific. This entire time, I've been arguing that the numbers are irrelevant specifically for current MBAs which have way more cooling than Intel specifies.

I don't see how your experience with some other computers/processors/cooling from years ago with different CPUs and different cooling is relevant to this conversation.



Never, ever, ever said that, because it would be ridiculous.

I want Apple to list speeds that customers are likely to see to avoid customer confusion, like the confusion that started this thread.

They are doing a little better now vs. before. I just checked the Apple Store page and the Turbo Boost speed is listed in the main list of specifications, which is good. This wasn't always the case.

If you go to "About this Mac," no mention of turbo boost is made.

Put all the arguments aside and one [fact] remains

Turbo CANNOT be GAURANTEED indefinitely. It just can't. If Intel can't guarantee it, neither can Apple. Just because you personally haven't encountered a work load that causes the CPU to revert to base clocks doesn't mean such a work load doesn't exist.

If your argument is that turbo frequencies should be documented, I agree. If your argument is that base frequencies are unimportant, I disagree.
 
Put all the arguments aside and one [fact] remains

Turbo CANNOT be GAURANTEED indefinitely. It just can't. If Intel can't guarantee it, neither can Apple. Just because you personally haven't encountered a work load that causes the CPU to revert to base clocks doesn't mean such a work load doesn't exist.

If your argument is that turbo frequencies should be documented, I agree. If your argument is that base frequencies are unimportant, I disagree.

My argument was that Apple publishing the base frequencies for MBAs is irrelevant (and even misleading) because MBAs [now] have over-specified power and cooling.

Users of current MBAs will basically never see their laptops running at base frequencies. Even if you find the magical workload that stresses the CPU more than Prime95 or whatever, that doesn't mean the CPU will back off to its base frequency. It will just reduce the amount of turbo.

You seem to have misinterpreted or misunderstood my position as saying that Intel specifying base frequencies is irrelevant. If you want, we can just chalk this situation up to a misunderstanding and call it good.
 
We'll have to chalk it up to a combination of miscommunication and simple disagreement as I still don't think publishing base frequencies is irrelevant. They should however publish turbo frequencies along with it. Just as it's documented here under the specifications tab.
 
We'll have to chalk it up to a combination of miscommunication and simple disagreement as I still don't think publishing base frequencies is irrelevant. They should however publish turbo frequencies along with it. Just as it's documented here under the specifications tab.

I don't know how to solve the problem of advertising CPU speeds but the base frequency is essentially just one data point, i.e., how fast the chip is guaranteed to run if it has exactly the amount of cooling that Intel specifies (TDP) and presumably the ambient temperature is within a specific range.

If an Intel CPU is in danger of overheating at its base clock speed (e.g., you are using it outside on a hot day) it will SpeedStep itself down to a lower clock speed. So even the base clock speed is fairly irrelevant unless you are talking about a processor under load installed in a controlled environment, e.g., in a supercomputer.

Maybe Apple could put a laptop in a room at room temperature, load down its CPU for ... 2 minutes (?), and advertise the average clock speed over that time.

Then again, it would take a lot of effort to explain why their clock speeds are different than the ones Intel specifies for the same chip.
 
I don't know how to solve the problem of advertising CPU speeds but the base frequency is essentially just one data point, i.e., how fast the chip is guaranteed to run if it has exactly the amount of cooling that Intel specifies (TDP) and presumably the ambient temperature is within a specific range.

If an Intel CPU is in danger of overheating at its base clock speed (e.g., you are using it outside on a hot day) it will SpeedStep itself down to a lower clock speed. So even the base clock speed is fairly irrelevant unless you are talking about a processor under load installed in a controlled environment, e.g., in a supercomputer.

Maybe Apple could put a laptop in a room at room temperature, load down its CPU for ... 2 minutes (?), and advertise the average clock speed over that time.

Then again, it would take a lot of effort to explain why their clock speeds are different than the ones Intel specifies for the same chip.

Except temperature isn't the only factor that will case turbo to stop working. I don't think I need to post all the different data points again do I?

What you're referring to is thermal throttling (not speedstep) speed step is what happens when the CPU is under no load or very light loads. Speed step is actually happening the vast majority of the time the computer is powered on. You can even turn speed step off and thermal throttling will remain functional.

You'll also revert to base clocks WELL before any type of thermal throttling occurs. In fact, thermal throttling will only happen in an over heating situation.
 
Except temperature isn't the only factor that will case turbo to stop working. I don't think I need to post all the different data points again do I?

What you're referring to is thermal throttling (not speedstep) speed step is what happens when the CPU is under no load or very light loads. Speed step is actually happening the vast majority of the time the computer is powered on. You can even turn speed step off and thermal throttling will remain functional.

You'll also revert to base clocks WELL before any type of thermal throttling occurs. In fact, thermal throttling will only happen in an over heating situation.

SpeedStep can and is used to control temperature. Remember the original MBA that would often SpeedStep down to 800MHz under load when it got hot?

Thermal throttling as introduced with the Pentium 4 multiplies the duty cycle of the clock signal, so it can get down to 200MHz and maybe slower.

But whatever. This goes back to my original point, that modern CPUs run at wildly varying clock speeds, at least under typical consumer workloads. In fact, I would argue that most consumer workloads cause the processor to be idle most of the time (800MHz) with short bursts of CPU load which will be at maximum turbo boost speed (~3GHz) because the processor is given a chance to cool off between bursts. Hence, my point that Apple publishing base clock speeds is pointless and misleading, if the chips aren't really ever going to run at those speeds.

You can post all the technical specifics you want, complete with marketing names (SpeedStep, Turbo Boost, Thermal Throttling, blah blah blah), but that doesn't refute my original point or any that I've made since then. Simply elaborating on the details of a point obviously doesn't prove that that point is incorrect.
 
But the technical specifics and "marketing names" does refute your original point. Speed step is a power saving feature which has the added benefit of allowing the processor to run cooler when not under load. Thermal throttling is a completely different animal and is there to protect the processor from physical damage. Base clock is the frequency that the CPU is guaranteed to maintain under any load condition provided its not over heating (the same cannot be said for turbo)

These are facts, you can try to dance your way around them, but they'll remain facts. Seems to me like you just want to have the last word, so I'll let you have it and simply leave you with the following:

5b1ffd37c9f5f13bf4c2a9ee5e27d13a7be09958eb62ac819b12f648104544e9.jpg
 
But the technical specifics and "marketing names" does refute your original point. Speed step is a power saving feature which has the added benefit of allowing the processor to run cooler when not under load.

Yes, this is partially correct. SpeedStep is used to reduce power consumption and ALSO to maintain temperature when under load. Refer to this Intel slide deck:

http://www.cs.virginia.edu/~skadron/tacs/rotem_slides.pdf

But this is a case of you just providing more detail about something and making it sound like you're refuting a point I made. I say these CPUs often/usually run at 800MHz, and you say this is because of SpeedStep as if you're disagreeing. WTF?

Thermal throttling is a completely different animal and is there to protect the processor from physical damage.

Again, you just saying stuff and not disagreeing with me or proving me wrong about anything.

Base clock is the frequency that the CPU is guaranteed to maintain under any load condition provided its not over heating (the same cannot be said for turbo)

Sure, as long as the chip has Intel's suggested amount of cooling (TDP), ambient temperature is within a certain range, etc., which isn't always the case.

These are facts, you can try to dance your way around them, but they'll remain facts. ...

Again, how am I "dancing around" anything? You are simply providing some marketing names and details with the end result being essentially that you agree with me, except for some reason your tone indicates that you're disagreeing. I simply don't understand this conversation.
 
Last edited:
Why don't you two get a room. You've completely highjacked the OP's thread and it's not a pretty sight.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.