Jump to content
seppa

LOGIC 8 BENCHMARK

133 posts in this topic

Recommended Posts

only 10.5.2 with logic 8.02

 

will try with 10.5.3 next week

 

So I guess it might be worth trying to uppgrade to 10.5.2 then. I mean 7-8 tracks more that´s like round 40% more than I can run and it just can´t be that an Asus mobo with a q6600 cpu is so much better than a Gigabyte with the same cpu on and 8 gig memory. This is more a test to se how much your cpu can take so mobos should´nt make any big difference. I´m also running Logic 8.02. Are using your onboard sound? I´m using a Saffire Pro 10 i/o FW card.

Any way thanks man I´ll have to take a look at the 10.5.2 looks promising for sure.

Share this post


Link to post
Share on other sites
Advertisement

Ok sorry but slow down with that eye smiley, I think I read somewhere people getting good scores with the onboard stuff that´s why I asked. I see you overklocked your cpu could be that´s it.

Share this post


Link to post
Share on other sites

you´re right............I didn´t realize that it´s not in my specs.

 

maybe better benches with onboard sound (didn´t try) - but never would use it because of the sound quality.

Share this post


Link to post
Share on other sites

27 tracks here as well. Logic 8.0.2, OS 10.5.3, Kalyway on an Asus P5K, Q6600, 4gb RAM. 512 and 1024 buffers.

 

I actually had to disable my onboard firewire to get my Siig pci fw400 card to work properly (TI chipset).

Share this post


Link to post
Share on other sites

Not sure quite how to work this benchmark. In continual loop I got 35 tracks to play for a period of time. Didn't play with any of the buffers or other settings.

 

This was through the onboard sound of my P5K-E WIFI mobo. QX9650 quad core at 3.6GHz

 

I also have a MOTU 24i, with PCI 424 card, but its not hooked up right now (in the process of moving).

 

This benchmark isn't exactly the right measuring stick for me because I only use real audio, and I have found that I can run as many plugins and tracks as I ever could unreasonably expect to use. A better test would be how many tracks of fragmented audio can I play with how many plugins, while recording x number of additional tracks. In other words, a combination of CPU and disk performance.

 

In any case, my hack is working well with the MOTU and Logic, and I can record and process all my typical projects with plenty of CPU and disk performance to spare.

Share this post


Link to post
Share on other sites

I have a max of 23 tracks playing on a q6600 @ 2.8ghz running at a 256 buffer on a {censored} soundblaster card.

 

Two things i noticed though was that i had to disable the I/0 safety buffer in the audio prefs also in the activity monitor my most maxed out core is only about 75% where as in logic its touching the top. Tried with the logic script fix posted on here same results as without.

 

Leo4all v2 10.5.2 and logic 8.0.2

Share this post


Link to post
Share on other sites

I spent a some time using this test in a q6600 (p35) + digi002, and well...at 256 samples, and logic 8.0.3 buffer setting in medium I can get about 25 tracks at the same time...but it stops after 1-2 minutes. The longest time without stopping were 8 minutes with 16 tracks. With the buffer set to "big" It never stopped playing, but I don't know if it's increasing latency too much.

 

On the other hand, I made the same test in a E6600 with AL883 in 10.4.11 and Logic 8.0.2, and at default settings (256 samples, medium buffer) It never stopped, so...it seems 10.5.3 still causes some problems in audio .

 

(spotlight was disabled in both computers)

 

With my own test for protools (almost 100% cpu usage) the longest time without stopping were 17 minutes.

Share this post


Link to post
Share on other sites

I managed to install retail Leo and the Apple update to 10.5.3 on my old P4 3Ghz hack, incredible but it worked runs really OK. Had to test it on the old hack before moving over to the quad core. I´ll come back with new testresults when I´ve fired up my q6600 hack on retail Leo 10.5.3. People talk about getting pretty much better performance in 10.5.3 so it´s going to be interesting to see.

Share this post


Link to post
Share on other sites

Tested with a Line6 Toneport UX8 @ 256 samples.

Same results as the others with similar specs.

 

Overclocking the CPU from 2.4GHz to 3GHz gave 3 more tracks.

Upgrading from 10.5.2 to 10.5.3 gave 1 more track.

 

/js

Share this post


Link to post
Share on other sites

After I upgraded to retail Leo 10.5.3 I got 23-24 tracks so big performanceincrease well not exactly :( hopefully I have a more stable system though. Lost dual display though cause I had to go for natit for my video card, Nvinject couldn´t give me QE in Leo 10.5.3 and that´s needed to be able to install Logic. Maby I´ll se if I can change to Nvinject anyway now that I´ve got it installed.

Share this post


Link to post
Share on other sites

I manged to get 38 tracks with default project settings on logic 8.2

 

my system is the following:

 

Ausus P5k Premium

8GB Kingston Memory @ 800mhz

Intel QX9650@ 3.6ghz (stock voltage, zelman cooler)

500GB 7k200rpm SATA drive

 

Retail DVD install, vanilla system with PC_EFI

Share this post


Link to post
Share on other sites
I manged to get 38 tracks with default project settings on logic 8.2

 

my system is the following:

 

Ausus P5k Premium

8GB Kingston Memory @ 800mhz

Intel QX9650@ 3.6ghz (stock voltage, zelman cooler)

500GB 7k200rpm SATA drive

 

Retail DVD install, vanilla system with PC_EFI

 

That´s amazing you can get that amount of tracks out of you hack, got to be cause of the CPU

you´re using. Checked out what they charge over here for it and it showed the CPU alone costs more than what I paid for my complete hack and I have a hack based on the Q6600 CPU.

Share this post


Link to post
Share on other sites

looks like that logic 8.02 is designed for new intel cpus.

 

i tried to overclock my q6600 up to 3550 mhz (worked stable) but cant get more then 32 tracks.

 

 

some {censored} things i found out, and i hope this is a logic bug, not a hack mac problem.

 

lets begin:

on my macbook pro i can run 18 tracks (both meters of cpu are full loaded, in activity monitor i have 185% cpu use for logic, sounds ok and normal)

 

same tracks on my hackmac at 3.55 ghz give me only 160% cpu load in activity monitor (quite ok i think)

but now i put more tracks, so in logic the meters got higher and higher... 32 is maximum without stop. when i look now in actity monitor

i see only 245% cpu use from logic.

 

is that normal? i tested some cpu burn tools and i can fully load all cpus without problems around 370% one application.

but in logic my maximum is 280% ( i got this with my own tests by using some minimoogs ) but more is not possible.

 

i understand that quad core at same frequenz is not twice as fast then dual core... but at the end i have 2x2,2 ghz vs 4x3,55 ghz. i thought that this can be handle minimum twice performance which means minimum 36 tracks for me. :)

Share this post


Link to post
Share on other sites
looks like that logic 8.02 is designed for new intel cpus.

 

i tried to overclock my q6600 up to 3550 mhz (worked stable) but cant get more then 32 tracks.

some {censored} things i found out, and i hope this is a logic bug, not a hack mac problem.

 

lets begin:

on my macbook pro i can run 18 tracks (both meters of cpu are full loaded, in activity monitor i have 185% cpu use for logic, sounds ok and normal)

 

same tracks on my hackmac at 3.55 ghz give me only 160% cpu load in activity monitor (quite ok i think)

but now i put more tracks, so in logic the meters got higher and higher... 32 is maximum without stop. when i look now in actity monitor

i see only 245% cpu use from logic.

 

is that normal? i tested some cpu burn tools and i can fully load all cpus without problems around 370% one application.

but in logic my maximum is 280% ( i got this with my own tests by using some minimoogs ) but more is not possible.

 

i understand that quad core at same frequenz is not twice as fast then dual core... but at the end i have 2x2,2 ghz vs 4x3,55 ghz. i thought that this can be handle minimum twice performance which means minimum 36 tracks for me. :lol:

 

It would be interesting to hear which version of OSX you are running. 10.5.3 was a huge improvement for multicore Logic users. 8 core MacPro users that before 10.5.3 had a maximum of around 50 tracks now report up to 70 tracks. So it is not the CPU:s but something about 10.5.3 and Logic that makes a huge improvement. Prior 10.5.3 a Quad 2.66 macPro had the same benchmarkresults as a 2.8 Ghz duo iMac.

I am running Leo4all 10.5.2 myself on a q6600@3.01 Ghz and can only play 25 tracks, as many as a MacPro quad 2.66Ghz at the same OS. If I look at the activity monitor I can see that Logic is only capable to use up to 50 percent of each core which is a shame but I have some reports from 10.5.3 users that indicate that 10.5.3 gives Logic the power to use up to 80 percent of each core which is really good actually. In my case I have some standaloneapps running at the same time as Logic, like BFD2,Kontakt2 and EWQL PLAY engine, so for me the poor multicore usage from Logic prior 10.5.3 is a good feature... leaving CPU headroom for other apps... Of course I will update soon to 10.5.4 but I will wait to see if Eddie (leo4all) comes up with an EASY upgradeway from earlier Leo4all versions first.

Share this post


Link to post
Share on other sites
It would be interesting to hear which version of OSX you are running. 10.5.3 was a huge improvement for multicore Logic users. 8 core MacPro users that before 10.5.3 had a maximum of around 50 tracks now report up to 70 tracks. So it is not the CPU:s but something about 10.5.3 and Logic that makes a huge improvement. Prior 10.5.3 a Quad 2.66 macPro had the same benchmarkresults as a 2.8 Ghz duo iMac.

I am running Leo4all 10.5.2 myself on a q6600@3.01 Ghz and can only play 25 tracks, as many as a MacPro quad 2.66Ghz at the same OS. If I look at the activity monitor I can see that Logic is only capable to use up to 50 percent of each core which is a shame but I have some reports from 10.5.3 users that indicate that 10.5.3 gives Logic the power to use up to 80 percent of each core which is really good actually. In my case I have some standaloneapps running at the same time as Logic, like BFD2,Kontakt2 and EWQL PLAY engine, so for me the poor multicore usage from Logic prior 10.5.3 is a good feature... leaving CPU headroom for other apps... Of course I will update soon to 10.5.4 but I will wait to see if Eddie (leo4all) comes up with an EASY upgradeway from earlier Leo4all versions first.

 

I have just upgraded to 10.5.4 and it seems my performance has dropped. im getting huge CPU spikes in my projects pushing one core into red from about 20% utillisation into the red, on just one core at a time.

My benchmark score has fallen to 31 tracks from 38

Its got to be something with the update as nothiing else has changed. Its very frustrating.

The thing that makes it even stranger is that when I push the AUDIO counter in logic to nearly 100% on all four cores, my Activity monitor only shows 46% cpu usage. Why cant I use the other 54%??? Strange. I didnt do this test when I got 38 tracks playing so I have no idea if there was as much headroom on the CPU from Activity monitots point of view or not,

 

I have tried ditching PC_EFIv8 for the Chameleon boot loader, replaced all vanilla .kext back from the upgrade including AppleSMBIOS.kext and AppleIntelCPUPowerManagement.kext

 

No ansers yet.

 

I might try putting the 10.5.3 kernel back

Share this post


Link to post
Share on other sites
It would be interesting to hear which version of OSX you are running. 10.5.3 was a huge improvement for multicore Logic users. 8 core MacPro users that before 10.5.3 had a maximum of around 50 tracks now report up to 70 tracks. So it is not the CPU:s but something about 10.5.3 and Logic that makes a huge improvement. Prior 10.5.3 a Quad 2.66 macPro had the same benchmarkresults as a 2.8 Ghz duo iMac.

I am running Leo4all 10.5.2 myself on a q6600@3.01 Ghz and can only play 25 tracks, as many as a MacPro quad 2.66Ghz at the same OS. If I look at the activity monitor I can see that Logic is only capable to use up to 50 percent of each core which is a shame but I have some reports from 10.5.3 users that indicate that 10.5.3 gives Logic the power to use up to 80 percent of each core which is really good actually. In my case I have some standaloneapps running at the same time as Logic, like BFD2,Kontakt2 and EWQL PLAY engine, so for me the poor multicore usage from Logic prior 10.5.3 is a good feature... leaving CPU headroom for other apps... Of course I will update soon to 10.5.4 but I will wait to see if Eddie (leo4all) comes up with an EASY upgradeway from earlier Leo4all versions first.

 

Well...after upgrading to 10.5.3 (using JaS Combo updater) I ran the benchmarktest again and I only managed to run 25 tracks ( maybe 2-3 more tracks than in 10.5.2 ) and the activity monitor only reported logic using 175% of the CPU before overload in Logic...( 400% is maximum with a quad CPU) showing that Logic has a looooong way to go before the term "multicore optimized" could be used...

It would be really interesting to hear some results from real Quad MacPro users. it is kind of strange that Logic is better optimized for 8 core than quad core computers...

Share this post


Link to post
Share on other sites

YES !!!! After updating to 10.5.4 using the Apple delta update and installing modbin 9.4.0 kernel before restarting I can now play 31 tracks in Logic 8, a HUGE improvement compared to the 22 tracks with 10.5.2. The same great improvement in a similar test in Logic 7.

Share this post


Link to post
Share on other sites

Hey FC this is BS, oh man first of all upgrading to 10.5.4 broke my 10.5.3 retail install but I managed to fix it. Still the same amount of tracks though so no succes story here :) . Seems the puter runs a little better though but it could be me just imagining things. I´ll take a look at the OC thing do´nt know nothing about it and I don´t know if I really need more cpu power since I´m basicly a wav guy but if I figure it out I´ll give it a shot.

 

I made sure to install the new mach_kernel and the right system.kext but it seems it made now diffrence in my case. Got my USb back though. Maby I should try the modbin kernel where did you get it.

Share this post


Link to post
Share on other sites

Wow, what's with the crappy performance nowadays? My old Athlon X2 system can do a lot more with XP + Nuendo. And I use quality plugins, not the built in ones. How does Cubase or Nuendo compare to Logic on a quad core Leopard system?

Share this post


Link to post
Share on other sites

Uhmm so you are able to run the Logic benchmark test mentioned in the first post on your Windows system with Cubase/Nuendo, congrats then.... :(

I don't doubt I'm able to playback way over 100 audio tracks with logic8 or cubase for that matter.... That is just not as interesting as running tracks with a heavy reverb plugin :-P

 

However I run Cubase4 on my Hackintosh system and did a little test a while ago with the Albino3 synth. Just a 4 note chord over 1 bar, same buffer settings etc.. Cubase use VST and Logic8 AU so it's not exactly the same setup but as close as it gets I think.

I was able to do 10 more stero tracks within Cubase4 compared to Logic8. Or 40 more notes if you like.

I believe it was about 30 Albino3 steroe tracks with logic and 40 in Cubase....

Share this post


Link to post
Share on other sites

Ah, I thought you were talking about tracks in general. :P Sorry.

 

I tried Logic 8 on my current setup and it was a lot hungrier than Nuendo in XP. I also thought the editing was clumsy (even though I was a Logic 4.8 user back in the day), so I probably won't bother with Logic again (no matter how great the plugins are).

Share this post


Link to post
Share on other sites

i began with kalway 10.5.2 updated to kalway 10.5.3 and now

i am running 10.5.4 (directly downladed in system update) with new kernel.

for me gave that update no improvements to 10.5.3 in logic. (maybe i get lower result in 10.5.2 anyway)

 

i was asking some friends with 4 core mac pros (3 ghz) they can open only 25-35 tracks with latest update.

 

what i really dont understand is that issue between actitivy monitor and logic cpu show. same on mac pro too.

 

the next thing what i really want to know, is how its possible that we users here, have so big differences in our results.

 

can somebody tell if he/ she has the same issue like me in following thing:

 

when i do a big project, for example:

20 audio tracks + 10 synth (cpu usement max 100% in actitvy monitor)

then i view to automations on all tracks... and my grafik stocks. its not anymore in nice framerates.

same happens when i show some eqs / compressors (when they are in front to see)... the grafic begin to become really slow.

 

anyone can tell me the same issue?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

×