Page 1 of 2

Sad CPU Load issue on a new PowerBook G4 1 ghz 17" 512 ddr

Posted: Tue Apr 08, 2003 11:37 pm
by CPU man
Hi !

I just tried Live 2.0.3 on a new PowerBook G4 17" 1 ghz 512 ddr ram
and compared it with a stationary P4 1,4 amd 1,4 ghz 256 sd ram
and got VERY scared:

After playing 8 tracks simultaniosly, each containing different 14 seconds
samples in 1 slot each in session view and adds this effect scheme:
Track 1 - Three Live Effects
Track 2 - Three Live Effects
Track 3 - Three Live Effects
Track 4 - Four Live Effects
Track 5 - Five Live Effects
Track 6 - Five Live Effects
Track 7 - Two Live Effects
Track 8 - Two Live Effects

The CPU Load meter on the PowerBook G4 1 ghz shows 78%
and can´t playback this short piece of music properly.
The CPU Load meter on the stationary amd 1,4 ghz shows only 22%

I feel fooled here, how come the PowerBook be so slow?????

Please give me a feedback asap as I´m considering bringing the slowbook back to the store. :evil:

Posted: Wed Apr 09, 2003 1:13 am
by jdg

Posted: Wed Apr 09, 2003 3:46 am
by Guest
i understand the grain of salt theory but what this guy needs to know is that the new G4 he got is not as powerful as the amd he compared it too.
it's really that simple and if he wants better performance, and it seems clear that he does, then he will need to take that power book back.

live OSX doesn't perform as well as OS9

Posted: Wed Apr 09, 2003 5:58 am
by yuval
but check out the latency - no PC could match that.

Posted: Wed Apr 09, 2003 6:09 am
by Guest
:lol: pc's are capable of getting latencies of 1.5-3ms these days. first thing to do is stop using pc's that are 10yrs old for your refference point.

Posted: Wed Apr 09, 2003 2:34 pm
by esoterica
Sigh. Yes, PCs are faster than Macs with Live. This does *not* have any bearing on whether PCs are faster than Macs on any other task, however. For a prime example, check out these statistics pages from distributed.net:

PowerPC statistics:
http://n0cgi.distributed.net/speed/quer ... test=rc564

x86 (Intel/AMD) statistics:
http://n0cgi.distributed.net/speed/quer ... test=rc564

Yes, a 1GHz G4 7450 gets a speed rating of 10.5M, while a 2.2GHz Athlon gets a speed rating of 6.7M, and a 2.8GHz P4 gets a speed rating of 5.2M (higher numbers are better). And yes, this means that Mhz for Mhz, with this application a G4 is 5.7X as fast as a P4, and 3.4X as fast as an Athlon XP.

So, while right now (and maybe forever, if Ableton never optimizes Live for PowerPC and AltiVec) a PC is faster for Live, the speed advantage doesn't extrapolate to any other application.

Posted: Wed Apr 09, 2003 3:42 pm
by esoterica
Anonymous wrote::lol: pc's are capable of getting latencies of 1.5-3ms these days. first thing to do is stop using pc's that are 10yrs old for your refference point.
My 2-month-old Dell 4550 desktop has out-of-the-box has latency of 110ms. My 2-year-old iBook has out-of-the-box latency of 4ms. I have yet to see a PC audio device that consistently provides latencies as low as you describe under Windows (this is one of the reasons Final Scratch is only for Linux and, shortly, Mac OS X).

There are certainly many compelling reasons to use Macs in audio production, if utmost speed in Live isn't your singular requirement.

Posted: Thu Apr 10, 2003 12:49 am
by Guest
Esoterica, I have a suspicion that your Dell might be set up incorrectly. I see latencies of ~15ms with no problems with my Dell P3/866 that I bought three years ago, with a SoundBlaster Live!Platinum card, and 256 MB RDRAM.

-0

Posted: Thu Apr 10, 2003 12:52 am
by ZeroAltitude
Oh ack. Stupid Mozilla alpha. I'm 'Guest' who has a Dell P3/866.

-0

Posted: Thu Apr 10, 2003 1:05 am
by Guest
You've been able to get less than 3ms with 16 I\O on a PC, plus onboard effects, for over four years on a PC with a Mixtreme card along with it's multi-client ASIO drivers. Two cards will give you less than 3ms with 32 I|O!

Next...

Posted: Thu Apr 10, 2003 1:49 am
by esoterica
Anonymous wrote:You've been able to get less than 3ms with 16 I\O on a PC, plus onboard effects, for over four years on a PC with a Mixtreme card along with it's multi-client ASIO drivers. Two cards will give you less than 3ms with 32 I|O!

Next...
That's fascinating, because according to Mixtreme's manufacturer, it only supports latencies down to 12ms. http://www.soundscape-digital.com/Suppo ... e/General/

It also retailed for $700 per card, and didn't even include any analog I/O. And in fact, according to the article, ASIO by itself doesn't even have a latency of less than 1.5ms.

By contrast, Apple's Core Audio with hardware allows for sub-1ms latencies. http://www.apple.com/macosx/technologies/audio.html Next...

Posted: Thu Apr 10, 2003 1:56 am
by esoterica
Anonymous wrote:Esoterica, I have a suspicion that your Dell might be set up incorrectly. I see latencies of ~15ms with no problems with my Dell P3/866 that I bought three years ago, with a SoundBlaster Live!Platinum card, and 256 MB RDRAM.

-0
Like I said, it's out-of-the-box with the standard (read: "junky") onboard SoundMax audio hardware. It's awful, awful sound quality (I didn't buy it for audio production; I just tested it on a whim), so I have no doubt that there's better available, but it still comes down to the out-of-the-box iBook still has significantly better latency than the aftermarket card you mention.

Posted: Thu Apr 10, 2003 2:12 am
by rpsx
esoterica wrote:
By contrast, Apple's Core Audio with hardware allows for sub-1ms latencies. http://www.apple.com/macosx/technologies/audio.html Next...
yes, that may be true, but i definatly am not getting those figures with live. though, i HAVE heard of people getting sub 1ms latencies with nuendo.

i rushed out and bought the mlan option for my kurzweil ksp8 hoping for the same numbers as reported on a demonstration through nuendo, and sadly, live comes nowhere close. i could use it at 3-4 ms, not the sub 1ms used in nuendo. wah wah.

why does different software perform differently with coreaudio? i'm seriously bummed with live now. too many variables in this world of computer audio... makes for less fun.

Posted: Thu Apr 10, 2003 2:14 am
by Guest
>That's fascinating, because according to Mixtreme's manufacturer, it only supports latencies down to 12ms.<

Well, they are underselling themselves then as the drivers go down to 64 samples with the ASIO drivers and the plugins work internally with a 2 sample latency.

Fascinating indeed ;)

Posted: Thu Apr 10, 2003 3:45 pm
by esoterica
Anonymous wrote:>That's fascinating, because according to Mixtreme's manufacturer, it only supports latencies down to 12ms.<

Well, they are underselling themselves then as the drivers go down to 64 samples with the ASIO drivers and the plugins work internally with a 2 sample latency.

Fascinating indeed ;)
The drivers will go down to 64 samples because that's ASIO's limit, not the card's. On my sound card I can set the latency to anything I want, but that doesn't mean it will work correctly.