Intel Dynamic Acceleration in MBP

[brettanderson]

One of the features of santa rosa, is intel dynamic acceleration. When a program is only using one core and the other core is idle, IDA will overclock the active core. In the case of the MBP, the 2.4 GHz will overclock to 2.6 GHz and the 2.2 GHz will overclock to 2.4 GHz.

Is there a real world situation where this actually occurs? Has anyone observed IDA in action on a SR MBP?

[opticalserenity]

Thats pretty neat. I'd like to see it demonstrated as well.

I saw a video once where one of the Intel founders said the software world is always several years behind them in technology...I believe it.

[illustratorDavi]

Don't supposed it will be used until Leopard. Wouldn't it use software to trigger it ?

[DoFoT9]

Thats pretty neat. I'd like to see it demonstrated as well.

I saw a video once where one of the Intel founders said the software world is always several years behind them in technology...I believe it.

yea there are so many things that we dont know about yet. 3D burning disks holding terabytes of info being one of those.. i read that somewhere. seems pretty dooable to me.

[otispunkmeyer]

Don't supposed it will be used until Leopard. Wouldn't it use software to trigger it ?

i believe this feature should work regardless of OS or software.

the logic (either in the cpu or chipset) is supposed to be able to detect when one cores being thrashed whilst the other remains idle....and overclock the busy core. its a hardware feature not a software one.

[LeviG]

well in windows (yes I know) my dual cores shunt all the lesser work (overheads) onto the other core when one core is being thrashed (well it actually sets both cores at around 50% but you get the idea). The overall speed of the single core encoded program is about 5-10% higher than if it was on a single core cpu of the same clock.

So say a 2.2Ghz dual core would be similar to a 2.4Ghz single core in single core operations.

[Kilamite]

Is this related to how the MBP only uses one core on battery?

[DoFoT9]

Is this related to how the MBP only uses one core on battery?

the MBP does that!?!?!?!?! is that from SR onwards?? ive never heard of this before...

[Kilamite]

the MBP does that!?!?!?!?! is that from SR onwards?? ive never heard of this before...

Not sure if it only uses 1 core, but I heard something along those lines for the SR ones.

If it's plugged in, Expose is smooth and animations are smooth, but on battery Expose is sluggish etc.

[DoFoT9]

Not sure if it only uses 1 core, but I heard something along those lines for the SR ones.

If it's plugged in, Expose is smooth and animations are smooth, but on battery Expose is sluggish etc.

that reminds me of a program.. it tells you your GHZ thingo, mines currently at 1GHZ... but im not sure how accurate it is. im running off power (battery unplugged) and CPU is at 100% constantly (folding)... interesting..

[NJuul]

Not sure if it only uses 1 core, but I heard something along those lines for the SR ones.

If it's plugged in, Expose is smooth and animations are smooth, but on battery Expose is sluggish etc.

The older models definitely does not. It's easy to test though, someone with an SR MBP simply unplug it and run activity monitor, and open the "CPU usage" window. It shows the per core utilization.


that reminds me of a program.. it tells you your GHZ thingo, mines currently at 1GHZ... but im not sure how accurate it is. im running off power (battery unplugged) and CPU is at 100% constantly (folding)... interesting..


If you take out the battery, your 'Book will only run at 1 ghz... There is some logic behind it, in that it prevents the computer from running at a higher speed than the powersupply can deliver. Sounds far fetched, but there you go.

[Kilamite]

The older models definitely does not. It's easy to test though, someone with an SR MBP simply unplug it and run activity monitor, and open the "CPU usage" window. It shows the per core utilization.

I loaded up Activity Monitor (I assume its the 2 black boxes with the red and green squares that are the cores?). Both boxes have red and green squares moving across, so I guess that means both cores are still active..?

Anyhow, Expose is really sluggish (although it can be smooth occasionally) compared to being on mains power.

[DoFoT9]

If you take out the battery, your 'Book will only run at 1 ghz... There is some logic behind it, in that it prevents the computer from running at a higher speed than the powersupply can deliver. Sounds far fetched, but there you go.

no sorry i muddled it up.. running off the power is like the same as running with your battery fully charged/charging.. i was just wondering why mine says its going at 1GHZ. i dont think it includes NICE processes in this calculation tho.
i really have very little idea about this tho.

[DoFoT9]

I loaded up Activity Monitor (I assume its the 2 black boxes with the red and green squares that are the cores?). Both boxes have red and green squares moving across, so I guess that means both cores are still active..?

Anyhow, Expose is really sluggish (although it can be smooth occasionally) compared to being on mains power.

yes still active, but not necessarily running at full capacity/clock

[NJuul]

I loaded up Activity Monitor (I assume its the 2 black boxes with the red and green squares that are the cores?). Both boxes have red and green squares moving across, so I guess that means both cores are still active..?

Anyhow, Expose is really sluggish (although it can be smooth occasionally) compared to being on mains power.

Strange... Exposé runs fluently on my CD MBP on or off battery. But yes, both your cores should be running. You can check the frequency using among others the ministat2 (http://shockwidgets.com/?widget=miniStat2) widget.

[Andrew275]

Both cores definitely continue to run when on battery power. I've assumed that the drivers are scaling back the GPU clock when on battery though, and that's why Expose is sluggish. If you activative Expose several times in a row, it becomes fast, which I assume is because it has scaled the GPU back up.

[DoFoT9]

Strange... Exposé runs fluently on my CD MBP on or off battery. But yes, both your cores should be running. You can check the frequency using among others the ministat2 (http://shockwidgets.com/?widget=miniStat2) widget.

i downloaded mini watsi.. pushed my mac to the limit and it still stayed on 1.0ghz the whole time. could this be because i dont have my battery in?? hhmm i dont see why it would be..

whats more important.. have i been missing out on power all along!!! hahaha :(

[thejadedmonkey]

could this be because i dont have my battery in?? hhmm i dont see why it would be..
He sees the light!

[Kilamite]

Put the battery in and let us know if you see a power boost ;)

[DoFoT9]

He sees the light!

i just dont see how it could be because the battery isnt in there.. or is that the power issue?? wait now i see....less power??????


nice yoda btw

[DoFoT9]

Put the battery in and let us know if you see a power boost ;)

i dont have me battery :( getting fixed

[NJuul]

i downloaded mini watsi.. pushed my mac to the limit and it still stayed on 1.0ghz the whole time. could this be because i dont have my battery in?? hhmm i dont see why it would be..

whats more important.. have i been missing out on power all along!!! hahaha :(

As I already said, if you take out the battery, your CPU is scaled back to 1 ghz. I read somewhere (don't remember where), that Apples reasoning for this is that it prevents the computer from running at a higher speed than the powersupply can deliver (does not sound plausible, I know). However, because of this build-in limitation, running your mbp without its battery in is not the same as running your mbp plugged-in with a charged battery.

[DoFoT9]

As I already said, if you take out the battery, your CPU is scaled back to 1 ghz. I read somewhere (don't remember where), that Apples reasoning for this is that it prevents the computer from running at a higher speed than the powersupply can deliver (does not sound plausible, I know). However, because of this build-in limitation, running your mbp without its battery in is not the same as running your mbp plugged-in with a charged battery.

farout ive been dying all along!!!! ok thanks so much for telling me this. ill just have to have my laptop running at 75 constantly :P

[CBJammin103]

So my 2.0GHz Macbook is more powerful - when on the battery - than any given Santa Rosa Macbook Pro?? :eek:

Bit of a turn off to a potential trader-upper. Hm...

[Kilamite]

So my 2.0GHz Macbook is more powerful - when on the battery - than any given Santa Rosa Macbook Pro?? :eek:

Bit of a turn off to a potential trader-upper. Hm...

Nah. I think as already said, it is to do with power consumption to make sure the MBP doesn't draw anymore power by accident from the battery that wouldn't matter if it was on mains power.

[brettanderson]

well in windows (yes I know) my dual cores shunt all the lesser work (overheads) onto the other core when one core is being thrashed (well it actually sets both cores at around 50% but you get the idea). The overall speed of the single core encoded program is about 5-10% higher than if it was on a single core cpu of the same clock.

So say a 2.2Ghz dual core would be similar to a 2.4Ghz single core in single core operations.

What programs did you run? I'd like to see the overclock happening on my machine. I've tried a few apps, like super pi but using cpu-z the CPU never goes above 2.2 GHz.

[Kilamite]

So say a 2.2Ghz dual core would be similar to a 2.4Ghz single core in single core operations.

A single core of a dual-core processor is not the same as the rated clockspeed for the over-all CPU. Nor is it half of it.

[LeviG]

A single core of a dual-core processor is not the same as the rated clockspeed for the over-all CPU. Nor is it half of it.

I never said it was. I know how multi cpu/cores work, I've been using them long enough :rolleyes:

I said
well in windows (yes I know) my dual cores shunt all the lesser work (overheads) onto the other core when one core is being thrashed (well it actually sets both cores at around 50% but you get the idea). The overall speed of the single core encoded program is about 5-10% higher than if it was on a single core cpu of the same clock.

So say a 2.2Ghz dual core would be similar to a 2.4Ghz single core in single core operations.

which is in regards to non multithreaded/smp aware applications.

On a single core/cpu system all aspects of the programs/os are running together and being spread out over a single cpu/core.
With 2 cpus/cores (or more) the load of programs/os is spread over the 2, meaning a single core program (assuming only 1 program making full use of core) would receive 50% of the overall processing power while the spare 50% would be used to cover the os overheads. As such this would give the program an advantage over a single core cpu clocked at the same speed because the overheads are not on the same cpu/core.

[Kilamite]

^ Sorry I wasn't meaning to say you were wrong - I was just trying to point out to people incase they took your post in a different dimension.

[DoFoT9]

this would also happen in xp im presuming?? it would be a hardware thing.. that might explain why my games have been running like DOGS!!!

[thefoo]

I thought I read somewhere that part of the power enhancements of SR was to switch off one core when the AC is off?

Seperately, surely the management of the cores is done by the CPU management system, thereby being OS agnositc? I think (again I don't remember details so feel free to correct) the OS and userland software can provide hints or overrides to this to optimise. For example some apps are programmed to make fuller use of mulitple CPUs where available.