INTERACT FORUM
More => Old Versions => Media Center 17 => Topic started by: Matt on November 08, 2011, 05:34:15 pm
-
I spent some time today benchmarking Media Center compared to iTunes:
http://jriver.com/speed.html
-
Matt and the team deserve a big hand for all the optimization work that's been done.
-
Yes, its quite amazing how fast and robust MC is.
-
That's amazing Matt. You guys rock!
-
Completely agree. Thanks guys!
-
The Big Guy.
-
I can definitely see it, skipping a couple versions in my upgrades.
FWIW, there's a typo in your second paragraph, "The versions of the programs 'tester' (sic) were Media Center 17 and iTunes 10.5." Let's not get cocky here!
-
I just can't believe you got iTunes to import 100k files without crashing, honestly.
It seems like it crashes on me when I breathe on it wrong.
-
Before somebody else beats me to it, I have the % faster wrong. It should be written like 6.59x faster, not 659% faster.
I was a little surprised just how slow some things got, so the testing took longer than expected. I originally wanted to use 500k tracks, but iTunes beat the idea out of me.
They use a strange database, because it looks like there are two copies -- a native one and an XML one. Both have huge overhead for seemingly simple things. For example, making a few 100k song playlists cause the database on disk and the memory usage while running to bloat by over 100 MB.
They do a pretty good job at panes and search. We're faster, but not by orders of magnitude. However, I think we're doing a much deeper search so that it's not an apples-to-apples (or apples-to-jriver) comparison. We also have to support flexible grouping and expression based grouping, which isn't possible with iTunes.
One final thing that struck me strange was that to change a value I see in a list, I have to right-click and pick "Get Info" and then find the value in a popup dialog inside several tabs. Editing with a command called "Get Info" is wonky.
I may give WMP 12 a shake next.
-
Some others to consider at some point:
The infamous XBMC and MediaPortal.
I know from experience that MediaPortal and it's relational db backend are slower than dirt. Back then I only had about 9000 media items (mostly audio) and looking at the audio view there took minutes (literally) to appear.
Oddly enough my brother and I were discussing the merits of the backends of XBMC and Media Center yesterday. I havne't tested that one either recently but it was slow too in the past. I would love to see how it stacks up...
-
I might add that this isn't just academic.
I've been able to get MP3 import running about twice as fast since starting testing yesterday. I updated the page to reflect this.
-
I tried to drag race WMP 12 today, but it looks like importing 100,000 MP3 files with it will take around 12 hours and I started it after lunch.
Ironically, I had its database on an SSD by accident which is making it faster than it should be.
I'll move the DB to the rotating drive used for the other benchmarks (this requires junction points with WMP or iTunes), and test again over night.
-
You know, whenever I feel like I can't do anything right, I open up WMP, and then I don't feel quite so useless.
-
14.5 hours later and WMP has 57332 of 100000 imported.
-
14.5 hours later and WMP has 57332 of 100000 imported.
:o ;D
-
You know, whenever I feel like I can't do anything right, I open up WMP, and then I don't feel quite so useless.
Hah!
I missed you, Tab. ;)
-
WMP is at 70% and counting.
-
WMP is at 70% and counting.
Microsoft engineers hard at work on WiMP (http://quietube2.com/v.php/http://www.youtube.com/watch?v=_flgoDc9igs).
-
One thing I've always wondered (as I'm not a Mac user).... does iTunes perform so badly just on Windows or is it a lot better on a Mac? Is this a relevant factor when doing such speed comparisons?
-
It is much better on OSX. Though still not "good", and I suspect many of the ways Matt is testing would come out basically the same. UI responsiveness is better and it isn't as flaky, though.
Of course, Macs do tend to have better hardware than your average generic Wintel PC. So, to lots of people, it seems even "more better" than it actually is.
-
WMP is at 93% after a few days.
Guess I have to start over, let it run through the weekend, and then check again probably Tuesday.
Did I mention we take 14 minutes to do the same thing?
-
glynor,
How about a movie of this? Maybe with little plastic race cars.
I wish we could get the Mac guy and the PC guy to narrate. Cardboard cutouts?
-
Wednesday 5:09:30 pm - 0%
Thursday 7:43 am - 57%
Thursday 4:12 pm - 70%
Friday 4:55 pm - 93%
It's slowing down, doing a less than 1% an hour. I think I'll just say "55 hours" and not bother to run it again.
-
How about 54? They could use a little help.
-
I think this info is pretty important, and ought to be set up with a simple graph on the JR website.
FWIW, I suspect MC could also have analyzed the files imported before iTunes got around to just importing them.
Against the blessed WMP, analyzing and also cover art, easily.
I suspect the reason WMP is taking so long is that it's still harvesting personal information from the test PC and phoning it home to Redmond.
-
glynor,
How about a movie of this? Maybe with little plastic race cars.
I wish we could get the Mac guy and the PC guy to narrate. Cardboard cutouts?
If you make poor Matt run it again, and screen capture it (full res, Animation codec), something could certainly be done.
I think it would need clown music.
-
I'm wondering if that WMP import has finished yet.
-
Might it be that WMP is doing a lookup on the internet for the tags at the same time?
This can slowdown things dramatically.
-
Might it be that WMP is doing a lookup on the internet for the tags at the same time?
This can slowdown things dramatically.
No. There was no network activity.
-
I'm testing WMP 12 import inside a Parallels Win 7 VM, and am getting about 500 tracks/minute imported.
So, at 100,000 tracks, that would be 3hrs 20minutes. How did yours take 2 days, 6 hours?
edit: it chugged along at 500/minute, completing my 17500 tracks in 35 minutes.
-
MrC, it is entirely possible (likely even, based on what I've seen) that the process slows down dramatically as the database size increases.
In other words, it might very well import the first 25k items at 500/min, but as the database design hits memory limits, it starts slowing down until near the 90k mark it is doing more like 2/min.
Based on what Matt reported while his tests were running, this seems to be the case.
-
That's probably it. I can test/verify if its worthwhile.
-
That's probably it. I can test/verify if its worthwhile.
To try to put a positive sounding spin on it, every file was faster than the next :P
-
A word of warning.
Anyone who intends to test WMP's MP3 import performance should be aware that WMP may resize & hide the folder.jpg cover art files (to 200x200 & set the hidden and system attributes: http://www.hydrogenaudio.org/forums/index.php?showtopic=75461) and ruin ID3v2 tags (any tags that it does not support may be in danger: http://www.hydrogenaudio.org/forums/index.php?showtopic=92813).
I don't know if anything triggers WMP's tag writing / cover art manipulation automatically, but I would test it only with copies of my files.
-
Yup, it does upon import. It also creates an additional AlbumArtSmall.jpg, so benchmarking is also being hit by excessive file creations.
I didn't test the registery settings to disable this, so don't know if they'll work (do you know?).
Best to probably stay away from the WiMP.
-
(do you know?).
No, I don't.
Best to probably stay away from the WiMP.
Yes.
In addition to these problems, in my experience, WMP12 is extremely awkward to use, even more so than WMP11. But perhaps I am biased...