mobile. performance?

My brother said that the mobiles perform fewer instructions per clock so that a Mobile at 2.2Ghz would perform worse than an Athlon XP at 2.2. I can't imaginw how/why this would be because they all come off the same factory line, don't they? And they are just really low heat chips. A mobile barton does have the same core as a barton xp, Right? except for the clock and multi lock. Help me convince him.

You are right, the Mobile XP (and Mobile 64) processors are rated to operate at a lower voltage (and thus lower power) than the desktop parts. No change is made regarding IPC (Instructions Per Cycle).

One difference is "Power Now" which allows the Mobile part to slow down when on battery power and under low workloads. This feature is why the mobile parts lack the multiplier lock.

There are some differences in the product line at 2400+ rating though. The Mobile is a Barton at a lower MHz than the Desktop Thoroughbred/Thorton but the extra cache of the Barton gives it the performance boost.

Mobile Bartons are basically normal Bartons spec'd to run at much lower wattage. They should perform just the same clock for clock as a desktop Barton.

they're the same chip, but unlocked and marked as a mobile.

however there was this discussion about how they're not recongized properly and lose their optimizations. not sure how that turned out.

the common mobile chips are bartons, and do run at a lower clock speed at the same rating. however, if you run them at the same ghz speed, the mobile barton should do much better than a non-barton.

Graphic67: PowerNow has nothing to do with whether the chip is locked or not. You can have a chip with every bridge locked, and PowerNow still functioning correctly.

Which cpus that have powernow technology also have locked multipliers?

emboss said:

Graphic67: PowerNow has nothing to do with whether the chip is locked or not. You can have a chip with every bridge locked, and PowerNow still functioning correctly.

Click to expand...

hmm.. any clue why they didn't do that?

I think that it is would have been more correct for me to say that the implementation of PowerNow BYPASSES the otherwise absolute control of the bridges allowing the multiplier to "float" as needed allowing the cpu to slow down without disrupting the fsb frequency.

yep, basically powernow doesn't require the multiplier to be unlocked. this is proven in the case where people mod regular superlocked bartons into mobile processors to adjust the multipliers using powernow capability. (cpumsr)