LIVE FIVE PERFORMANCE TEST - results here

Discuss music production with Ableton Live.
Post Reply
AdamJay
Posts: 4757
Joined: Thu Mar 11, 2004 7:17 pm
Location: Indianapolis, USA

Post by AdamJay » Sun Aug 07, 2005 11:48 pm

subterFUSE wrote:Seems like we have a pretty decent sized compilaton of results....

Are we going to see a consolidated results list?
yes, eventually.

ethios4
Posts: 5377
Joined: Tue Dec 02, 2003 6:28 am

Post by ethios4 » Mon Aug 08, 2005 12:17 am

Live 5.0.1

1) CPU - 40%
2) Desktop - Shuttle XPC SN41G2
3) OS - Windows XP Home sp1
4) CPU info - AMD AthlonXP 2600+ (2.08 GHz)
5) RAM info - 512 MB PC2700 DDR333
6) Soundcard - M-Audio Audiophile 2496(pci) (M-Audio drivers v5.10)
7) HD - 7200 Maxtor 8MB

Thanks Adam!

drush
Posts: 1282
Joined: Tue Jul 29, 2003 4:40 am
Location: Venice, CA
Contact:

Post by drush » Mon Aug 08, 2005 7:08 am

Poster wrote:
drush wrote:1) Ableton Live CPU Meter 43%
2) Apple G5 Desktop
3) OSX Tiger 10.4.2
4) CPU; dual 2.0 ghz
5) Ram; 2.5 GB
6) Soundcard; RME Multiface/PCI
7) Hard Disk; 7200 rpm SATA


ooo, a whole percentage point! ;)
haaaah..
so you beat me by 1%?
damn..

but very strange;
you've got double my RAM and a slightly faster CPU..
or is it the RME that slows things down?
unless you or i have a bunch of other things open i really doubt there is any difference between 1.25 and 2.5gb ram for purposes of this test. the RME card definitely does not slow anything down, so i'd say that 1% difference is just CPU and then the mysterious difference that always exists between any two computers that are supposedly set up the same.

Danji
Posts: 16
Joined: Fri Oct 01, 2004 8:54 am
Location: Berlin, Friedrichshain, KochhannStr.
Contact:

Post by Danji » Mon Aug 08, 2005 7:28 am

FaX-01 wrote:
Danji wrote:On Live 5.01 Demo
1) Ableton Live CPU Meter - 36-39% 8O :(
2) Laptop- Fujitsu-Siemens Amilo D8830
3) Operating System - WinXP Home sp1
4) CPU Make Model and Speed -P4 3.06HT
5) Amount of Ram / Speed of Ram - 2048 DDR 2700
6) Soundcard - Maudio 410Firewire
7) Hard Disk Drive Speed - ??4200-5400??

That's not bad at all why the sad face :) ?
I'm on a 3.2ghz Machine and get 32% so i wouldn't be all that worried.
This test was meant to be far more CPU intensive than the last one.
In relative terms you're still doin' better than a G5 1.8-2.0ghz system by a few %.
And none of our fellow Mac buddies are finding Live unproductive at those levels of CPU use.
I get more than enough juice on my PC IMHO.
You do on yours also.
Seriously are you finding that level of CPU functionality is not enough ?
Just curious.
Jes its tru. I can good work with my system. load a lot of VST(i) and to play many tracks is no problem. But there a Centrinos and Pentium M´s with about 1.7 Ghz and a performance with about 20-25% in the "Live 5 performance test". Thats what i don´t understand. 1.7Ghz "LATRINO" is better than a PENTIUM ROCK´N ROLL 3.06H/T. Maybe i have a wrong chipset or a silly motherboard.?
Win XP Home, P4 3.06 H/T, 2048Ram, 4200rpm HD, M-Audio FW410, Faderfoxes LV1-LX1
Live 4.13, Asio4All2.6. Audiorealism Basline, Effectrix, Massive
My Music: http://danji.bplaced.net/wordpress/

subterFUSE
Posts: 1557
Joined: Tue May 31, 2005 11:04 pm
Location: Winter Park, FL

Post by subterFUSE » Mon Aug 08, 2005 11:56 am

The Pentium M is a different CPU than the Pentium 4.

The Pentium M is more closely based on the Pentium 3 processor (probably one of Intel's most successful products ever).... but optimized for speed, and heat dissipation.

So you can't just compare the clock speeds... because the Pentium M is obviously a more efficient CPU than the P4.

anti-banausic
Posts: 1609
Joined: Fri Jun 11, 2004 9:15 pm
Location: NYC

Post by anti-banausic » Mon Aug 08, 2005 8:47 pm

My setup is very similar to Fax's:

1) Ableton Live CPU Meter - 33%
2) Laptop- Sager8790
3) Operating System - WinXP Home sp1
4) CPU Make Model and Speed -P4 3.2 GHz H/T
5) Amount of Ram / Speed of Ram - 1024 Mb DDR3200
6) Soundcard - Echo Indigo DJ
7) Hard Disk Drive Speed - 7200

Pretty pleased with that since it is 16 months old now.
Macbook c2d 2.0, 2G RAM, 160G HD 5400 RPM, OSX(10.5.5), XP Home, LIVE6, BCR 2000, UC33e, Yamaha P-200, Logic Studio, KRK V6 II

FaX-01
Posts: 1483
Joined: Sat Jun 26, 2004 3:58 am

Post by FaX-01 » Tue Aug 09, 2005 1:48 am

anti-banausic wrote:My setup is very similar to Fax's:

1) Ableton Live CPU Meter - 33%
2) Laptop- Sager8790
3) Operating System - WinXP Home sp1
4) CPU Make Model and Speed -P4 3.2 GHz H/T
5) Amount of Ram / Speed of Ram - 1024 Mb DDR3200
6) Soundcard - Echo Indigo DJ
7) Hard Disk Drive Speed - 7200

Pretty pleased with that since it is 16 months old now.
And identical CPU use :D .........
My aren't the wings of butterflies beautiful and do they not make wonderful perturbations.....

anti-banausic
Posts: 1609
Joined: Fri Jun 11, 2004 9:15 pm
Location: NYC

Post by anti-banausic » Tue Aug 09, 2005 2:10 am

Odd init?

Cheers Fax.
Macbook c2d 2.0, 2G RAM, 160G HD 5400 RPM, OSX(10.5.5), XP Home, LIVE6, BCR 2000, UC33e, Yamaha P-200, Logic Studio, KRK V6 II

smart1123
Posts: 447
Joined: Thu Mar 10, 2005 10:39 pm
Location: Vancouver, Canada

Post by smart1123 » Tue Aug 09, 2005 5:05 am

Ableton Live CPU Meter 64-68%
1.5 gHz Powerbook G4 (15" Combo)
OSX.4.2
1 GIG DDR PC2700 RAM
MOTU Traveller
5400 RPM Internal

Enabling All I/O's on the Traveller added 4%
15" TiBook 1.5 GHz 1Gig RAM, MOTU Traveller, Live 5, Reaktor 5, Alesis Micron, Yamaha EX-5, UC-33e, BCR2000, Lexicon MPX-1, Orbit, Event 20/20's

handojin
Posts: 50
Joined: Sun May 29, 2005 12:21 am
Location: Vancouver BC

Post by handojin » Tue Aug 09, 2005 5:24 am

"So you can't just compare the clock speeds... because the Pentium M is obviously a more efficient CPU than the P4."
totally true... the thing that im seeing lots of about the centrino's is that using them at low latency settings (64k to 256 k or so buffersize) they perform spot on with dual xeon's workstations, kid you not, the xeon's speed up once the buffer goes past 12 or 24 milliseconds but live works best at low latencies and thats what makes it so different then other apps.. also another thing me and many other cubase/nuendo users notice with the centrino processors is when using extremely low latency settings, 1.5- 3 ms you can also pin the peformance and get it higher than anything else prior be it amd64 opteron or xeon. with my socket 754 desktop @ 1.5 it only takes about 75% performance to cause a click or drop out in audio where as with a centrino you can get it up to nearly 95% performance before you notice anything.. im not sure with live but this is very evident with cubase sx.. maybe someone here can confirm, live is definatly much more elegant with a high cpu meter so this might not be the case with ableton software but these sort of results have been noticed by many other people aswell and im pretty sure ive seen someone else mentioning this.

FaX-01
Posts: 1483
Joined: Sat Jun 26, 2004 3:58 am

Post by FaX-01 » Tue Aug 09, 2005 6:50 am

handojin wrote:"So you can't just compare the clock speeds... because the Pentium M is obviously a more efficient CPU than the P4."
totally true... the thing that im seeing lots of about the centrino's is that using them at low latency settings (64k to 256 k or so buffersize) they perform spot on with dual xeon's workstations, kid you not, the xeon's speed up once the buffer goes past 12 or 24 milliseconds but live works best at low latencies and thats what makes it so different then other apps.. also another thing me and many other cubase/nuendo users notice with the centrino processors is when using extremely low latency settings, 1.5- 3 ms you can also pin the peformance and get it higher than anything else prior be it amd64 opteron or xeon. with my socket 754 desktop @ 1.5 it only takes about 75% performance to cause a click or drop out in audio where as with a centrino you can get it up to nearly 95% performance before you notice anything.. im not sure with live but this is very evident with cubase sx.. maybe someone here can confirm, live is definatly much more elegant with a high cpu meter so this might not be the case with ableton software but these sort of results have been noticed by many other people aswell and im pretty sure ive seen someone else mentioning this.

I'm sure it's due too the 2meg cache and bussing architecture.
My 3.2ghz P4 Toshiba is kick ass when you a)have larger latency settings and b)absolutely shit loads of plug-ins going.
The dothan/sonoma/centrino chipset really do shine though in low buffer/latency tests and I suspect that Live relies on CPU caching alot more than similar apps/daws.
Most people forget that all the time stretching is based on relatime granular algo's on multiple tracks so it pretty much uses CPU effectively for the amount of realtime CPU calculations that are need IMHO.
Coming from hardware back in the late 80's it's pretty mind boggling what this app can pull off ,particularly from a non destructive realtime sample editing perspective.
Still marvelling at what software can do having only been using it since Live 3.0.
Truly great piece of application coding IMHO :D .
My aren't the wings of butterflies beautiful and do they not make wonderful perturbations.....

rikhyray
Posts: 3644
Joined: Wed Aug 25, 2004 12:13 pm
Contact:

Post by rikhyray » Wed Aug 10, 2005 11:04 pm

CPU 24-27%
1.73 Sonoma ( I think) Vaio FS215 M
XP Home
512 Ram
Indigo i/o
7200 Hitachi
i checked it first without any audio tweaks, absolutely no difference, I keep the original 512 Ram because there comes FS 285 with 1G and have the deal that I can exchange the notebooks ( well it is just the law in Germany i can give it ack and get full refund within 14 days)
First ever pc that looks good, like Mac (or better) but hell to set it up right. Costed me some 20 hours to have it my way.

subterFUSE
Posts: 1557
Joined: Tue May 31, 2005 11:04 pm
Location: Winter Park, FL

Post by subterFUSE » Wed Aug 10, 2005 11:10 pm

I'm sure it's due too the 2meg cache and bussing architecture.
I don't think it's the 2 MB cache.... but maybe it is the bussing architecture.

I say this because I just got a Pentium 4 3.4 GHz with 2 MB cache.... and it only runs the test at about 35%. If the 2 MB cache were such a big help, I think my results would be better.

handojin
Posts: 50
Joined: Sun May 29, 2005 12:21 am
Location: Vancouver BC

Post by handojin » Wed Aug 10, 2005 11:26 pm

that extra 2 mb of cache should make your the rendering process faster and thats about it... when i say render i mean bounce down file.. encoding.. that sort of stuff.. shouldnt make any difference with real time performance. not just in audio appz.. im pretty sure ive seen other computer enthusiasts saying that. if you where to time the mixdown process you probably would have great results. fpu performance is what matters for software synths, so amd rule the desktops, and intel rule the laptops, memory performance is why i like the amd setups, memory controller built into the cpu, therefore memory performance/speed is not tied to the fsb.

hellospiral
Posts: 7
Joined: Wed Jan 19, 2005 8:14 pm

Post by hellospiral » Thu Aug 11, 2005 11:39 am

1) Ableton Live CPU Meter 68 - 70%
2) Apple macmini
3) OS X 10.3.7
4) 1.25
5) 256MB
6) Built in Soundcard

I'm pretty happy.

Post Reply