Experience thus far
Posted: Fri Feb 17, 2012 9:44 pm
I’ve run FancyCache for about three weeks.
Hardware
Core i5
8GB RAM
500GB Hitachi z5k500 HDD
24GB SSD
FancyCache Configuration
Volume Edition
2048 MB lv 1
4kB Block Size LFU-R
Read/Write
600 seconds Release After Write with Averaging Write Amount
Readyboost enabled on 24GB SSD
My daily usage profile (Microsoft Office, Outlook, iTunes, Firefox) results in about 35% Read Hit Rate and 80% Write Hit Rate. Reducing lv1 cache size to 1024MB results in approximately 20% Read Hit Rate with almost no change to Write Hit Rate.
These statistics suggest that the Read Cache is marginally effective (35% hit rate), although much faster, compared to the 24GB Readyboost. The Write Cache is highly effective.
What would be helpful for analysis purposes is a FancyCache datalogger that would record the Performance Monitor for post-analysis. With datalogging, we could quantify write flushes, write buffer performance, read hits, etc. This is what “real” software testers do, no?
Hardware
Core i5
8GB RAM
500GB Hitachi z5k500 HDD
24GB SSD
FancyCache Configuration
Volume Edition
2048 MB lv 1
4kB Block Size LFU-R
Read/Write
600 seconds Release After Write with Averaging Write Amount
Readyboost enabled on 24GB SSD
My daily usage profile (Microsoft Office, Outlook, iTunes, Firefox) results in about 35% Read Hit Rate and 80% Write Hit Rate. Reducing lv1 cache size to 1024MB results in approximately 20% Read Hit Rate with almost no change to Write Hit Rate.
These statistics suggest that the Read Cache is marginally effective (35% hit rate), although much faster, compared to the 24GB Readyboost. The Write Cache is highly effective.
What would be helpful for analysis purposes is a FancyCache datalogger that would record the Performance Monitor for post-analysis. With datalogging, we could quantify write flushes, write buffer performance, read hits, etc. This is what “real” software testers do, no?