1.) I don't have mathlab.
2.) I never said you made a mistake (someone else did).
I'm just saying that the CPU difference doesn't make sense for a 45x difference. I have no idea why you are seeing this difference.
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.
Thinking to buy the 11.6'', is it enough for my needs?
- Thread starter cili0
- Start date
- Sort by reaction score
1.) I don't have mathlab.
2.) I never said you made a mistake (someone else did).
I'm just saying that the CPU difference doesn't make sense for a 45x difference. I have no idea why you are seeing this difference.
I don't have Matlab either, I'm just commenting on basic expectations. Theoretically speaking, there is no task that the 2.53Ghz MBP(I'm assuming it is the C2D) will do more quickly than what can be predicted by just the linear difference in cpu speed. Realistically, you can be limited by the hard driv,e by ram, by cpu cache, and possibly some other things depending on the test, and these limitations will prevent the MBA CPU from running at its full potential.
please list the full specs of both machines and maybe describe the simulation a bit more if you want to get some better feedback of what went right/wrong. My guess is that one of the following things happened:
1) The simulation was huge and you ran out of ram on the MBA and started swapping
2) The simulation was tiny and small enough to fit in the L2 cache of the MBP but not small enough to fit in the L2 cache of the MBA CPU.
3) You were running some apps in the background on the MBA that slowed it down.
4) You had a Core i5 and not a C2D in the MBP and the simulation you ran was accelerated by one of the new extensions(like I described in another thread with the AES acceleration, which is several orders of magnitude faster on a Core i5 than a C2D)
I'm sorry, but I think that you are wrong. The 11.6" MBA processor is ULV (Ultra-Low Voltage) while the 13.3" MBA processor is only LV (Low Voltage). In electronics, lower voltage systems are generally more power efficient (more computations per Joule of energy). You gain MHz with a higher voltage pricessir but you pay a penalty in power efficiency.
The TPD (Total Power Dissipation) of the ULV device is 10W while that of the LV device is 17W. That means that the 13.3" MBA processor will dissipate 70% more heat. It may provide more horse power (get things done a little faster)... but it DOES NOT provide a 70% improvement over the 11.6" MBA processor!
Result: The 13.3" CPU generates more heat any way you look at it.
Last edited:
That is all true, in theory. The TDP isn't a perfect estimate of the heat generated, especially with these CPUs. The 17W is just the benchmark for being classified as ULV, but I bet the real number is quite a bit lower now that the CPUs are so much more mature.
The other thing you aren't considering is the rest of hte system. If the 13" finishes 30% faster, that means you also use the display, memory, ssd, etc. 30% less as well.
Last edited:
Doesn't work that way.
The power used in the display is primarily the backlight, which is on regardless of how fast you complete a task assuming you want to see whats going on.
Memory, SSD are only used when being accessed. Their access is not appreciably slower on the 11" than the 13". They're both running the same bus speed.
Your comment is true for the CPU. When running processor constrained tasks, if the CPU needs to run at high speed longer, it will consume its max power for a longer period of time than the faster CPU. For some tasks, that means the 11" will run out of battery power much faster than the 13". For less CPU intensive tasks, the difference isn't as large. Anandtech's review included exactly these types of comparisons.
The display backlight doesn't use power when you shut the lid and put the system into standby. If you are running a simulation or something, that is exactly what you will be doing. Run it, and when it finishes, put the laptop in standby. You aren't going to just leave it sitting there waiting for the 11.6" model to finish its simulation.
Hello,
I liked your reply,
I did some more matlab today using old codes that I knew were very efficient on my Macbook pro 13'' 2.53Ghz c2d - same with those, I did notice a huge difference and it is sad because the computer feels like a good machine when doing anything else. my conclusion is that matlab (or say matematica) are the kind of programs that require a lot of calculation from the processor and at the end of the day, it is when pushed to limit like so that we feel the slower processor. But seriously, i doubt the average guy uses these programs on an every day basis and that being said, the average guy would never feel a difference and should go for it if that,s what he wants.
I purchased the computer knowing what to expect, and I'm aware now that my MBP is gone, that I will have to compensate eventually by buying a macmini for instance, which can keep up with what I do.
Cheers
Michael
Doesn't make sense.
You can't calmly explain away a 45x difference in speed based on a 1.7x increase in processor speed.
You should try to address some of the questions. Otherwise it just looks like a typo or you didn't understand what you were running.