Re: Oprofile and Compaq DCPI Comparison

[John L. <le...@mo...>]

On Tue, Aug 24, 2004 at 02:18:52PM -0400, Balle, Susanne wrote:

> Could you get me some information about the tests you guys did that 
> proved it to be useless compared to careful coding of the fast path?
> Do you have a white paper or a memo that describe the results?

This is the closest you're going to get that's on record:

http://marc.theaimsgroup.com/?l=oprofile-list&m=105840481716960&w=2

819.09 (0.00%)  | 60.30 (0.00%)         | 506.39 (0.00%)        | 50000.cvs.out
816.10 (-0.37%) | 59.65 (-1.08%)        | 505.91 (-0.10%)       | 50000.mod.out

This is CPU_CLK_UNHALTED at 50,000. cvs.out is with the module hash table in
the interrupt handler. mod.out is without.

regards
john