I see a difference between FETCH/STORE and db_put/db_get. All this confirms for me is that BerkeleyDB is fast enough. I'm just glad it competes nicely with the file system (which as you said has all sorts of in-kernel advantages). My own system is OpenBSD 3.2 using the GENERIC kernel on a 233 MMX pentium using ATA-100 discs in "PIO mode 4, Ultra-DMA mode 5" (whatever that means).
In reply to Re^3: BerkeleyDB vs. Linux file system
by diotalevi
in thread BerkeleyDB vs. Linux file system
by perrin
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |