Geoff Thorpe wrote:
>
> On Fri, 30 Jun 2000, Ben Laurie wrote:
>
> > The point about using the user time is that it works when the machine is
> > loaded. I agree about hardware accelerators - my thought when testing
> > the Atalla is that we should provide two figures, a real-time and a
> > user-time one, and the reader of the figures can decide which to use.
>
> Even when using hardware accelerators, the CPU time can be useful. Eg. if
> doing speed for 10 seconds shows 0.3 seconds of CPU time used, it gives a
> measure of how free the CPU is to do other things while the card is
> performing the crypto - in particular how many other hardware accelerators
> it might be able to drive at the same time before it became an issue. If
> 10 seconds of hardware accelerated crypto requires 4 seconds of CPU, then
> it shows the CPU is not 99-100 % free to do other things, and that the
> host system itself probably can't be pushed to do much more than it
> already is.
Agreed - which is why you need both figures and a brain to interpret
them.
Cheers,
Ben.
--
http://www.apache-ssl.org/ben.html
Coming to ApacheCon Europe 2000? http://apachecon.com/
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [EMAIL PROTECTED]
Automated List Manager [EMAIL PROTECTED]
