Ableton Forum

ejectorset wrote:adamja,
but couldnt you get the load up high enough if live supported multi-threading and multiple procressors/cores ???

it seems like then when you needed the extra power it would be used.

im just curious if you are sliding the blame (or frustration, whatever you want to call it) towards apple or ableton?

i put the blame on Apple.
why take away this option that was in Powerbooks for 6 years, and remove it starting with the last G4 PB?

AdamJay wrote:

ejectorset wrote:adamja,
but couldnt you get the load up high enough if live supported multi-threading and multiple procressors/cores ???

it seems like then when you needed the extra power it would be used.

im just curious if you are sliding the blame (or frustration, whatever you want to call it) towards apple or ableton?

i put the blame on Apple.
why take away this option that was in Powerbooks for 6 years, and remove it starting with the last G4 PB?

Correct me if I'm wrong, but it would seem that Jobs might have implimented this "flaw" to keep heat or battery life down. Possibly for PR's sake. Most users (including myself) wouldn't run performance tests or so much as read into a published performance test deeply enough (fortunately I did), to notice that yet again Apple shipped crippleware.

Like "The iBucket will hold 4 gallons of water (but we filled it half way with rocks) because it leaks if it has more than 2 gallons. I hope no one looks inside iBucket)."

It could simply be that Apple has managed to refine their CPU speed algorithms, making the Automatic vs. Highest option completely unnecessary - the only people that would find this to be less then satisfactory are people doing benchmarks!

My guess is that Apple have improved the CPU speed algorithms to some extent, and decided to combine the improved CPU management with the marketing requirement for longer battery life, thus removing the option altogether.

They may have done tests and came to the conclusion that "Reduced" was unnecessary too, since the extra time taken for the processing of most tasks means that the overall power/time saved is negligible. I'm just throwing ideas around here....

that could be true forgie.

Just so you see how far we've come, my TiBook G4 goes up to 156% with buffer set to 2048

lol

forgie wrote:It could simply be that Apple has managed to refine their CPU speed algorithms, making the Automatic vs. Highest option completely unnecessary - the only people that would find this to be less then satisfactory are people doing benchmarks!

My guess is that Apple have improved the CPU speed algorithms to some extent, and decided to combine the improved CPU management with the marketing requirement for longer battery life, thus removing the option altogether.

They may have done tests and came to the conclusion that "Reduced" was unnecessary too, since the extra time taken for the processing of most tasks means that the overall power/time saved is negligible. I'm just throwing ideas around here....

i thought this at first... but then when i start getting drop outs when i've got live 5.2 going at 85%... and firing up Xbench was the only thing to get it to drop down (and it went down to 62%)... it becomes painfully clear that for realtime audio their speedstepping algorhythm is flawed. i cannot get Live to push it up far enough to turn off the speedstepping. and the more i push it, the less stable Live is.

just give us the fucking option back Apple.
i paid for a dual 2ghz, not a dual 1.2ghz.

AdamJay wrote:
i thought this at first... but then when i start getting drop outs when i've got live 5.2 going at 85%... and firing up Xbench was the only thing to get it to drop down (and it went down to 62%)... it becomes painfully clear that for realtime audio their speedstepping algorhythm is flawed. i cannot get Live to push it up far enough to turn off the speedstepping. and the more i push it, the less stable Live is.

just give us the fucking option back Apple.
i paid for a dual 2ghz, not a dual 1.2ghz.

someone should do a similar test with Logic Pro
I'll bet it works perfectly

gomi wrote:

AdamJay wrote:
i thought this at first... but then when i start getting drop outs when i've got live 5.2 going at 85%... and firing up Xbench was the only thing to get it to drop down (and it went down to 62%)... it becomes painfully clear that for realtime audio their speedstepping algorhythm is flawed. i cannot get Live to push it up far enough to turn off the speedstepping. and the more i push it, the less stable Live is.

just give us the fucking option back Apple.
i paid for a dual 2ghz, not a dual 1.2ghz.

someone should do a similar test with Logic Pro
I'll bet it works perfectly

since logic pro has a multi-threaded audio engine, yes i am sure once you start to push it - it actually well give up the rest of the cpu for you to use.

AdamJay wrote:

gomi wrote:

AdamJay wrote:
i thought this at first... but then when i start getting drop outs when i've got live 5.2 going at 85%... and firing up Xbench was the only thing to get it to drop down (and it went down to 62%)... it becomes painfully clear that for realtime audio their speedstepping algorhythm is flawed. i cannot get Live to push it up far enough to turn off the speedstepping. and the more i push it, the less stable Live is.

just give us the fucking option back Apple.
i paid for a dual 2ghz, not a dual 1.2ghz.

someone should do a similar test with Logic Pro
I'll bet it works perfectly

since logic pro has a multi-threaded audio engine, yes i am sure once you start to push it - it actually well give up the rest of the cpu for you to use.

Has anyone done such a test elsewhere on the net?

OK, just did the Live 5 performance test again on my 2.16Ghz MacBook Pro (built-in audio, 2GB RAM, 7200 rpm HD). But this time I had iTunes running in the background and Im getting 29-34%. 10 points faster than with Live running by itself.

i think that might have been what i did to warpus.


i think i had one more thing running as well when i got those low numbers.

i gotta test it again, i just need to download the new beta.

i've been talkign with some tech heads over at the mac forums.

basically they say all this is necessary in order to keep the laptop cool and lasting on battery supply.

so whats the main issue here? that we would have liked apple to allow us to turn the stepping feature off?


part of the problem is ableton not using both cores so once it does will we still see this problem?

Digi V wrote:i've been talkign with some tech heads over at the mac forums.

basically they say all this is necessary in order to keep the laptop cool and lasting on battery supply.

i think that that is a default answer.

i just ran UnrealTournament 2004 for 2 hours on all high settings (utilizing all the cpu) and it got no hotter than usual.

i also monitored the battery usage. screen brightness seems to have more of an effect on battery than the cpu utilization. folks at macrumors.com have been experimenting with CHUD tools. when you install CHUD tools from the OSX DVD, you get a new pref pane for disabling one of the cores. and they've found it has a marginal effect on battery saving.

also i tried disabling the 2nd core to get live to "push" into high cpu usage and i have had some luck...

i was able to bring one .als that was hitting 74% down to 62%
so as far as "usability" goes with actually getting work done in Live - this is one workaround (though it kills your 2nd core for graphics processing). this is not great for benchmarking though.

But really we just need Apple to give us the option back.

yeh i dont understand why they wouldn't give us the option either.

AdamJay, this is a known issue. The last rev of the PPC powerbook have the same problem!

Do you have reaktor?