On 11/2/07, robert engels <[EMAIL PROTECTED]> wrote:
> As expected it did not make much difference. Here is the code:

Yeah, might already be IO bound with FSindex.
Might see better improvement with a RAMDir since copying in bigger
chunks eliminates the double buffering so you go from 4 copies
OS -> IndexInput.buffer -> user.buffer -> IndexOutput.buffer -> OS
 to 2 copies
OS -> user.buffer -> OS

-Yonik

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to