Paul Blackburn <[EMAIL PROTECTED]> writes: > It is interesting that you believe "A memory-based cache does not > usually improve performance". Having used both disk and memory based > caches. My experience is that a RAM cache is considerably faster. Modern OSes actually maintain a memory-based cache of the disk, so even when you use a nominally disk-based cache, you are actually caching the hottest data in memory as well. In some cases (probably heavily write-biased workloads, with poor locality of reference), memory-based caches will outperform disk-based caches. On the other, no cache at all would perform even better.
- Question about optimum cache size. Brian Buhrow
- Re: Question about optimum cache size. Lyle Seaman
- Re: Question about optimum cache size. Paul Blackburn
- Re: Question about optimum cache size. Lyle Seaman
- Re: Question about optimum cache size. Mitch Collinsworth
- Re: Question about optimum cache size. Lisa Giacchetti
- Re: Question about optimum cache size. Lyle Seaman
- Re: Question about optimum cache size. Mitch Collinsworth
- Re: Question about optimum cache ... John Hascall
- Re: Question about optimum cache size. Lisa Giacchetti
- Re: Question about optimum cache size. Michael Niksch
- Re: Question about optimum cache size. Lyle Seaman
- Re: Question about optimum cache size. Chuck Lever
- Re: Question about optimum cache size. Lyle Seaman
