Hey, I've seen a number of detailed Primocache posts from the AT forums show up on my Google searches so I figured I had a better shot here than anywhere else of ironing out this issue, as Romex software's support seems to be a little lackluster.
To really test Primocache out, I assigned a huge amount of RAM to my Steam games drive, like 56GB (I've also tested this with 8GB, 16GB, and 32GB). Cache was set up as a read-only 4kB cluster size with prefetch enabled and start on boot. I loaded up Total War Warhammer 2, launched my campaign, ended turn, and quit the game. I repeated this process 3-5 times before rebooting the PC and doing it all again. The performance I got from the cache varied wildly.
Mostly, the cache hit rate was less than 1%. Sometimes it would get to around 20-25%, others 35-45%, and the big kahuna 90%+ happened on one reboot, but has not happened since. Load times in this way were sharply affected. Loading with a 1% cache hit rate vs a 90% cache hit rate feels like two completely different computers. So, I know that Primocache CAN work, but I don't know why it often does not.
I adjusted the RAM cache size to see if there was any issue there. I made a read+write cache. I changed the cluster size. I cached all my storage devices and not just my Steam library SSD. My Windows recovery partition often got its RAM cache hit rate higher more than my Steam library's cache. Bummer!
Hopefully someone can help me out here to dial this in. I really like the performance when the software works, and I'd like to be able to support them. Thanks for any help!
To really test Primocache out, I assigned a huge amount of RAM to my Steam games drive, like 56GB (I've also tested this with 8GB, 16GB, and 32GB). Cache was set up as a read-only 4kB cluster size with prefetch enabled and start on boot. I loaded up Total War Warhammer 2, launched my campaign, ended turn, and quit the game. I repeated this process 3-5 times before rebooting the PC and doing it all again. The performance I got from the cache varied wildly.
Mostly, the cache hit rate was less than 1%. Sometimes it would get to around 20-25%, others 35-45%, and the big kahuna 90%+ happened on one reboot, but has not happened since. Load times in this way were sharply affected. Loading with a 1% cache hit rate vs a 90% cache hit rate feels like two completely different computers. So, I know that Primocache CAN work, but I don't know why it often does not.
I adjusted the RAM cache size to see if there was any issue there. I made a read+write cache. I changed the cluster size. I cached all my storage devices and not just my Steam library SSD. My Windows recovery partition often got its RAM cache hit rate higher more than my Steam library's cache. Bummer!
Hopefully someone can help me out here to dial this in. I really like the performance when the software works, and I'd like to be able to support them. Thanks for any help!