This is just an observation, but I have noticed that the ram cache seems to get aggressively paged out on my system. To illustrate the point. Eboostr with no ram cache on my system reports Page file usage as 486 MB. Adding a 512 MB ram cache takes the page file usage to 997 MB. So 511 MB appears to have gone straight back on to disk.

If the 512MB where to be pegged in ram, and not allowed to page out, performance should be substantially better. To illustrate the impact of paging on performance I attempted the following tests.
1. Using Alacrity PC starting a game, allowing it to load and then exiting immediately.
– First with eboostr disabled
– Second with Alacrity PC set to force programs in memory to swap to disk
– Third with the force to swap option turned off.
– Finally I have included the time for 2x2GB flash disks

Note that the game (simulator) in question has a significant start up time, partly bound by disk but also by processor.

2. Results

Test 1. Basic load – eboostr off
11:39:20 Monitored program(s) have stopped
11:37:58 Starting auto start program
00:01:22 Duration

Test 2. Eboostr Ram cache 1632MB filled to 96% – Memory optimized = forced to swap
12:26:50 Monitored program(s) have stopped
12:24:29 Starting auto start program
00:02:21 Duration

Test 3. Eboostr Ram cache 1632MB filled to 96% – Memory not optimized = not forced to swap
12:34:57 Monitored program(s) have stopped
12:32:52 Starting auto start program
00:02:05 Duration

Test 4. Eboostr No Ram cache, 2x2GB UFD 56% – Memory not optimized = not forced to swap
13:30:53 Monitored program(s) have stopped
13:28:43 Starting auto start program
00:02:10 Duration

3. Conclusion and thoughts

– The ram cache is fastest in this basic test provided it has not been forced to swap to page file.
– If ram cache is forced to swap, performance was pretty much what one would expect for a start without eboostr.