Subject: Re: ip_flow.c
To: Guenther Grau <Guenther.Grau@bk.bosch.de>
From: Bill Sommerfeld <email@example.com>
Date: 01/28/1999 10:03:55
> Besides, Charles, did you do any testing on how much computing
> power/network throughput you change would yield?
I have this gut feeling that, given the cost of global varible
accesses and mispredicted branches on some architectures, that testing
to see if we should skip the check will cost about as much as just
checking the checksum all the time.