Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[outofspace7]

version 12 has a cpu usage flaw

Just look at this:


version. 12.0.1.30

[Alex B]

Perhaps it is creating thumbnails or running Auto-Import.

Mine uses typically 0-2 % during standard audio playback. (I don't run Auto-Import in the background)

What do you actually do with MC when the CPU usage is high?

As a test you could try to disable all third party plugins, visualizations and possibly running automatic tasks. Also post your system info from MC's Help menu.

[outofspace7]

Auto-Import is off,  i dont have any plugin.

Output format is :


When playing music cpu= 50-90% in stop mode cpu = normal usage

I got sempron 3400+ and x-fi creative music card with latest drivers

[Matt]

Resampling 6 channels of audio is quite CPU expensive using MC's high-quality resampler.

For example, a 3.2 Core Duo normally uses 0% when playing APE files even with some DSPs enabled.  It uses closer to 8% when doing 5.1 @ 96 kHz.

We'll take a look on our end, but with an older CPU this may simply be too much heavy lifting.  Using 2 channels or 44.1 kHz will substantially reduce the CPU load.

[outofspace7]

ok i will try, but is it strange that in Media Center 11 everything is ok?

btw amd sempron 3400+ isn't at all a old cpu

[ADDiCT]

Matt, are all JRiver staff members using Dual Core machines with 2Gig of RAM and the latest ATI cards? If this is true, then i want to work for you! Testing should be done on a "slow" machine, IMHO.

[Matt]

ok i will try, but is it strange that in Media Center 11 everything is ok?

We'll take a look.  CPU usage of resampling (and 5.1 conversion) shouldn't have changed from MC 11. (except to get faster if anything)

btw amd sempron 3400+ isn't at all a old cpu

Agreed.  I didn't mean to imply otherwise -- only that resampling 5.1 sound is CPU hungry on any machine.

[Matt]

Matt, are all JRiver staff members using Dual Core machines with 2Gig of RAM and the latest ATI cards? If this is true, then i want to work for you! Testing should be done on a "slow" machine, IMHO.

I agree that speed and efficiency is crucial.  We spend a lot of time optimizing.  Hopefully you'll agree that MC's performance bares this out.

However, it's also important to keep an eye on what's coming.  For example, multi-core computing is the (immediate) future for CPU speed increases.  For this reason, MC 12 uses threaded parallelism wherever possible. (equalization, list drawing, handheld operations, etc.)

[outofspace7]

so at the end, this problem should be fixed if nothing was new from MC11 in 5.1  (that is also a huge reason to sorrow) ,so the problem is somewhere else....hope so to fix that in next build

[Alex B]

On my 2.8 GHz Hyperthreaded P4 (over three years old by now) the CPU usage varies between 15% and 30 % when a 2-channel/44.1 KHz/16-bit source is outputted as 5.1/96 kHz/24-bit .

MC11.1 and MC12 are not different. The CPU usage looks identical.


outofspace7,

What is your source sample rate? Are you aware that resampling cannot increase quality? High quality resamplers can only keep the quality reduction inaudible.

Your X-fi card should be able to playback bit-perfectly without any software resampling if you have selected the correct driver options. I recall that the bit-perfect output mode is called "creation mode" or something like that.

With lesser soundcards that can use only 48 kHz it may be beneficial to resample to 48 kHz instead of letting the other output stages do mediocre quality resampling. 96 kHz just uses more resources without making the audio quality any better.

[Matt]

I think Alex is correct that CPU usage of an apples-to-apples comparison should show similar CPU usage between MC 11 and MC 12.

However, the next build of MC 12 will have some tuning to improve performance when resampling AND converting to 5.1.

Thanks.

[MGD_King]

Matt, for what it's worth, CPU usage is significantly higher when I crank up the sample rate to anything past 44100 on my Athlon 2500. When set at the recommended setting, CPU usage is peaks about 6% regardless of what I set anything else to.