Re: More disc cache improvements

On 8 May 2015 as I do recall,
Vincent Sanders wrote:

>
> I am interested in the persistent disc cache write performance on
> different platforms, especially RISC OS, to see if my recent changes
> improve the situation there..
>

[snip]

> I would like these two lines from the logfile along with the OS and
> hardware spec of the system. E.g. "RISC OS 5 on Iyonix with FAT
> formatted hard drive" or "ROOL beta on Raspberry Pi 2 with FAT
> formatted SD card"
>
Here are three examples from recent browsing sessions on using RISC OS
5 on Iyonix - no idea what format the hard drive is, I'm afraid, but it is a
normal hard disc and not a card-based system.


(13649.720000) content/fs_backing_store.c finalise 1613: Cache total/hit/miss/fail (counts) 1376/258/1118/0 (100%/18%/81%/0%)
(13649.720000) content/llcache.c llcache_finalise 3361: Backing store wrote 35071872 bytes in 0 ms average 23816 bytes/second


(7479.800000) content/fs_backing_store.c finalise 1613: Cache total/hit/miss/fail (counts) 656/168/488/0 (100%/25%/74%/0%)
(7479.800000) content/llcache.c llcache_finalise 3361: Backing store wrote 3034863 bytes in 0 ms average 5796 bytes/second


21213.750000) content/fs_backing_store.c finalise 1613: Cache total/hit/miss/fail (counts) 2061/418/1643/0 (100%/20%/79%/0%)
(21213.750000) content/llcache.c llcache_finalise 3361: Backing store wrote 10209144 bytes in 0 ms average 22666 bytes/second

There's clearly massive variation involved; I notice that the 'fast'
examples above (1 and 3) list over a hundred images never
rendered while the 'slow' example lists only 41, and example 1 lists 83
'excessive conversions' from 59 images converted more than once while the
second two runs both list zero excessive conversions....
--
Harriet Bazley == Loyaulte me lie ==

An atheist is a man with no invisible means of support.