Paradoxxx
12-19-2005, 12:13 PM
INTEL HAS A naming problem again, and it is not one that is easy to pigeonhole. The problem is that if you buy a chip, it theoretically runs at a distinct clock rate, and that is that. If you want to save power, you can run it slower also, but that is an option. What happens when a chip automatically runs faster, briefly? How do you market it?
Lets take the auto-overclocking in Yonah for example. If you buy a CPU at 2.16GHz, you expect it to act like the rest of the 2.16GHz CPUs. If that CPU can ramp up to 2.33 when circumstances permit with the laptop equivalent of Foxton, how fast is it? 2.16 or 2.33? Before you write to say it is a model T243-7aR//763T243-7aR//763, please choke on something before you put your fingers on the keyboard.
Now, to make matters worse, what do you do if production variances mean that CPU1 will spend 10 per cent of its life OC'd, and CPU2 will be there 20% of the time? Is CPU 1 binned at a higher spec than 2? Is it marketed as faster? What about when you add in a laptop with a more efficient cooling system, are the CPUs in each one still the 'same' since the environment affects the cores in a measurable way?
This one gets ugly quickly, and if Intel doesn't deem its sales people smart enough to trust them with GHz figures, how are they going to explain this? The funky sales card that doesn't exist decoding model numbers won't help here. Yup, this one degrades into a royal mess in no time at all.
There are a few things that Intel is doing differently this time. It is changing the way it bins CPUs. Instead of sheer clock rate, it is moving to clock rate and power. If a CPU can hit 2GHz on the tester, it is then subjected to power use tests. The next phase of the binning is one where average power usages get taken into account as well. If you have a 2GHz 1W average chip, and another that is 2W, they will be in different lots. The 2W part will probably be downbinned to a 1.8GHz model, and it will definitely be a lesser part in the real world.
Rumour has it that the top two speed bins of Yonah will be notably lesser in real world power draw than their lower brethren, leading to 'faster' chips with real world longer battery life. This may explain why Yonah went from a planned 2.5GHz +/- a bin to 2.16GHz, the extra clock was sacrificed on the altar of power efficiency.
This is actually a really good thing for the consumer, and we hope it catches on, it will extend battery life, for a price. The problem is, how do you tell the proles? Do you tell them it is a set speed, and put 12 little footnotes after every use of a model name, between the numbers and the ™ symbol? Do you just give up and dangle a shiny thing in their faces?
Whatever it ends up doing, there is real science behind it, and due to the current US war on anything intellectual, that message will most assuredly be lost. µ
Nothing new, just interesting reading.
Lets take the auto-overclocking in Yonah for example. If you buy a CPU at 2.16GHz, you expect it to act like the rest of the 2.16GHz CPUs. If that CPU can ramp up to 2.33 when circumstances permit with the laptop equivalent of Foxton, how fast is it? 2.16 or 2.33? Before you write to say it is a model T243-7aR//763T243-7aR//763, please choke on something before you put your fingers on the keyboard.
Now, to make matters worse, what do you do if production variances mean that CPU1 will spend 10 per cent of its life OC'd, and CPU2 will be there 20% of the time? Is CPU 1 binned at a higher spec than 2? Is it marketed as faster? What about when you add in a laptop with a more efficient cooling system, are the CPUs in each one still the 'same' since the environment affects the cores in a measurable way?
This one gets ugly quickly, and if Intel doesn't deem its sales people smart enough to trust them with GHz figures, how are they going to explain this? The funky sales card that doesn't exist decoding model numbers won't help here. Yup, this one degrades into a royal mess in no time at all.
There are a few things that Intel is doing differently this time. It is changing the way it bins CPUs. Instead of sheer clock rate, it is moving to clock rate and power. If a CPU can hit 2GHz on the tester, it is then subjected to power use tests. The next phase of the binning is one where average power usages get taken into account as well. If you have a 2GHz 1W average chip, and another that is 2W, they will be in different lots. The 2W part will probably be downbinned to a 1.8GHz model, and it will definitely be a lesser part in the real world.
Rumour has it that the top two speed bins of Yonah will be notably lesser in real world power draw than their lower brethren, leading to 'faster' chips with real world longer battery life. This may explain why Yonah went from a planned 2.5GHz +/- a bin to 2.16GHz, the extra clock was sacrificed on the altar of power efficiency.
This is actually a really good thing for the consumer, and we hope it catches on, it will extend battery life, for a price. The problem is, how do you tell the proles? Do you tell them it is a set speed, and put 12 little footnotes after every use of a model name, between the numbers and the ™ symbol? Do you just give up and dangle a shiny thing in their faces?
Whatever it ends up doing, there is real science behind it, and due to the current US war on anything intellectual, that message will most assuredly be lost. µ
Nothing new, just interesting reading.