More > JRiver Media Center 27 for Linux

Docker Image for MC27

<< < (6/32) > >>

HaWi:

--- Quote from: sjhilton on September 25, 2020, 08:52:04 pm ---
I'd be interested to know what everyone thinks of the performance of the types of docker images. The JRMark benchmark numbers are almost the same for jatzoo vs shiomax. I think the network performance seems to be a bit faster on shiomax?


--- End quote ---

I am using the same library on my iMac MC27 and for the Docker container, sharing the same media files. Although the JRMark of the Docker container is much lower than the iMac's, I still find the WebUI of the Docker container more responsive and snappier, especially when scrolling a large pane.  That may have something to do with the fact that both, the media and the library, are located on the NAS.

max096:

--- Quote from: HaWi on September 26, 2020, 11:39:49 am ---I am using the same library on my iMac MC27 and for the Docker container, sharing the same media files. Although the JRMark of the Docker container is much lower than the iMac's, I still find the WebUI of the Docker container more responsive and snappier, especially when scrolling a large pane.  That may have something to do with the fact that both, the media and the library, are located on the NAS.

--- End quote ---

Well... your prebuilt NAS probably has half the cores and much lower clock speeds. If you wanted to make a comparison you should be running both on your MAC or even better both on some Linux Box (as I'm not sure how consistent JRMark is cross platforms).

That's what I'm getting on my desktop on Fedora 32

--- Code: ---=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 1.597 seconds
    Single-threaded floating point math... 2.159 seconds
    Multi-threaded integer math... 0.329 seconds
    Multi-threaded mixed math... 0.282 seconds
Score: 4350

Running 'Image' benchmark...
    Image creation / destruction... 0.295 seconds
    Flood filling... 0.352 seconds
    Direct copying... 0.345 seconds
    Small renders... 0.671 seconds
    Bilinear rendering... 0.179 seconds
    Bicubic rendering... 0.114 seconds
Score: 11255

Running 'Database' benchmark...
    Create database... 0.261 seconds
    Populate database... 1.028 seconds
    Save database... 0.111 seconds
    Reload database... 0.043 seconds
    Search database... 1.580 seconds
    Sort database... 0.596 seconds
    Group database... 0.887 seconds
Score: 4771

JRMark (version 27.0.15 64 bit): 6792

--- End code ---

And in docker on that same machine


--- Code: ---=== Running Benchmarks (please do not interrupt) ===

Running 'Math' benchmark...
    Single-threaded integer math... 1.602 seconds
    Single-threaded floating point math... 2.160 seconds
    Multi-threaded integer math... 0.327 seconds
    Multi-threaded mixed math... 0.284 seconds
Score: 4346

Running 'Image' benchmark...
    Image creation / destruction... 0.313 seconds
    Flood filling... 0.354 seconds
    Direct copying... 0.348 seconds
    Small renders... 0.685 seconds
    Bilinear rendering... 0.186 seconds
    Bicubic rendering... 0.114 seconds
Score: 10998

Running 'Database' benchmark...
    Create database... 0.263 seconds
    Populate database... 1.058 seconds
    Save database... 0.123 seconds
    Reload database... 0.057 seconds
    Search database... 1.750 seconds
    Sort database... 0.581 seconds
    Group database... 0.850 seconds
Score: 4592

JRMark (version 27.0.15 64 bit): 6645

--- End code ---

So... really it does not matter much.

In the container I run on my NAS I also only get 3400'ish. Again much slower box overall and the container is in a VM I only gave 4 vcores of a 4790k and 8gb ram vs a dekstop with 3900x and 32gb ddr4. It's basically my old desktop recycled as a NAS. Interestingly enough it gets 5099 in the database benchmark, which is more than the desktop. Both the desktop and the NAS have similar NVME ssds, just the one on the NAS is bigger as it's the boot drive for 5+ VMs. Guess bigger really is faster. It does make little sence to me though, because I thought that only makes a difference for max threwput and that a database load won't care about that.

HaWi:
I agree, my comparison is not a great way of doing it. Your benchmarks are impressive!

HaWi:
Max, thanks for the update to MC27.0.21
With docker compose it only took a minute to update. So cool!

RvdZ:
Hi,
I'm running MC25 on docker, and this is working great. I tested MC27 and this also works fine. Great job!
But now I'm trying to get scrobbling to Last.fm working. I haven't done that before on JRiver. I did this like 10 years ago on squeezecenter, and I'm trying to get it working now on JRiver again.
I would like to get it working on MC25, but if it needs MC27, I just have to upgrade.

But now I'm running into the following problem:
When I go to tools-options-services-last.fm-...manage last.fm account...-connect to last.fm account, then nothing happens (MC27).
In MC25 it says a webpage should launch to authorize access, but this webpage is NOT launched.
The same happens when I try to manage the accounts for Twitter or Youtube (no webpage launched).

If I go to the main-menu - streaming - Last.fm, then the Last.fm page is launched correct.
So the docker is able to connect to the web, but it's not able to open a web page from the tools-menu.

Any idea what's going wrong?

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version