2015-02-19, 06:01
Thanks. I may not respond to every reply posting results - I'll probably collate them for a comment in PR6484, or simply post a link to this thread so the relevant parties can determine the effectiveness of the change themselves.
Is anyone using a higher-end MySQL server (ie. i3, i5 etc.)? I think low-end-ish servers are likely to add a noticeable delay with the extra compression stage, but this may not be nearly so bad with higher end gear.
Also, comparing the results over WiFi would be interesting (I'm guessing there may be a noticeable benefit when accessing a low-end server over WiFi).
It may be that compression is only of benefit when the client-to-server connection is bandwidth limited (so not wired, and irrespective of server spec).
Edit: Just to add - when reporting results please include the hardware spec of your MySQL server (Make/Model/CPU/RAM), Kodi client (Pi1/Pi2, ARM frequency), and network connection (WiFi 11g/n/ac etc., wired 10/100/1000Mbit/s etc.)
Is anyone using a higher-end MySQL server (ie. i3, i5 etc.)? I think low-end-ish servers are likely to add a noticeable delay with the extra compression stage, but this may not be nearly so bad with higher end gear.
Also, comparing the results over WiFi would be interesting (I'm guessing there may be a noticeable benefit when accessing a low-end server over WiFi).
It may be that compression is only of benefit when the client-to-server connection is bandwidth limited (so not wired, and irrespective of server spec).
Edit: Just to add - when reporting results please include the hardware spec of your MySQL server (Make/Model/CPU/RAM), Kodi client (Pi1/Pi2, ARM frequency), and network connection (WiFi 11g/n/ac etc., wired 10/100/1000Mbit/s etc.)