Coolbook on 2010 Macbook air 13"

[aberrero]

Reducing frequency and maintaining the same voltage DOES reduce power consumption, just not to the same degree that reducing voltage does. Leaving throttling active will save you battery, although you might want to get rid of the intermediate frequencies and only have 800MHz B/2 and 1.86Ghz without anything in between.

I just tested mine at 1.6Ghz and 0.925V and it works fine. Going for 1.8Ghz now.

Anyway, can someone please explain to me what the throttling level dropdown box means?

Also, would it be wrong for someone to post the beta version here for us, or should I email the dev and ask him for it?

 

[Mechinyun]

Read my post a few back, explains what i believe is the levels of throttling.

Beta version people refer too was before the current version was released for testing on the air.

 

[mrmister]

I disabled throttling because I just picked one frequency--the highest one, at the lowest voltage. Thus no need for a process to dynamically adjust. If it does use more power, for whatever reason, it is not reflected in my (admittedly ad hoc) testing.

 

[Cerano]

you can use coolbook to overclock too right?

 

[kamui]

My MBA 2010 13.3" settings:

Frequency Voltage
1862 MHz 0.9250 V (highest speed that was stable at 0.9250 V, the lowest voltage)
2128 MHz 0.9500 V (at 2.13 GHz, any lower voltage was unstable)

Throttling level Medium Temp Limit Off

Here's my settings:

Frequency Voltage
1862 MHz 0.9250 V
2128 MHz 0.9650 V

2.13 GHz at 0.9500 for me would kernel panic.

 

[Mr. Zarniwoop]

Curious, how did you find out that .9250 was unstable at 2128 ghz?

I actually worked well into the night using my MBA 2010 13.3" and had to get a little more conservative with the voltages to keep it stable:

Frequency Voltage
1596 MHz 0.9250 V
1862 MHz 0.9375 V
2128 MHz 0.9625 V

 

[abriwin]

My MBA 11" is cool.

Talking of cool MacBookAirs, has anyone noticed when picking up their MBA when it's not been used for a while how cold it is to the touch? I wonder whether Apple have quietly used a new aluminium alloy that has a higher thermal conductivity than the old ones, to help keep it cool with less use of the fan?

 

[abriwin]

My MBA 12" is cool

duplicate
how to delete?

 

[Mechinyun]

Talking of cool MacBookAirs, has anyone noticed when picking up their MBA when it's not been used for a while how cold it is to the touch? I wonder whether Apple have quietly used a new aluminium alloy that has a higher thermal conductivity than the old ones, to help keep it cool with less use of the fan?

I noticed this too, but since its my first mac I thought they all did this.

 

[petterihiisila]

I turned off throttling so that it can just stay locked on 1.6ghz.

For me the undervolt settings were gone after a reboot, when Throttling active was unchecked. I left it checked with just one pair: 1.6 GHz with the lowest voltage.

Frequency: 1600 MHz
Voltage: 0.8750 V
Throttling level: Low
Temp limit: Off
[✓] Throttling active
[✓] CoolBook active

With these settings, Activity Monitor states that CoolBookSwitcher takes 0.0 % CPU cycles, and Real Mem consumption is 660 kB. All stable. I'll leave it like this.

--
Macbook Air 11.6" | 1.6 GHz | 4 GB | 128 GB

 

[petterihiisila]

Aftering tinkering whether it would be better to lock the frequency to 1.6 GHz or let it fluctuate, I decided to email CoolBook maker Magnus L. and ask: "Does CPU MHz setting affect power consumption and heat? Or is it just voltage that matters? Intuitively both could make sense."

Here's Magnus, copy-pasted with a permission:

"The consumption is close to equal for the same number of instructions performed by the CPU. This means that when you edit text files and similar "static" tasks, the difference should be minimal.

However, when you run at maximum or high loads, "holding" back the CPU will decrease the consumption a lot. The computer will of course be a lot slower too, so your work may also take twice as long, resulting in the same battery drainage.

You can use CPUTest and System information in OS X to experiment. For my 11" Air I get a consumption of around 2080 mA when running CPUTest at full frequency and 0.8750 V. I only get a consumption of around 1350 mA at 600MHz and the same voltage.

For some tasks this can be used to increase the battery times. Web browsing flash sites could be an example where this may give you some gains."

Based on this, the best bet is probably one pair, or max two pairs. I put mine to 800 MHz (B/2) and 1600 MHz. Both at the lowest voltage.