| From | Sent On | Attachments |
|---|---|---|
| Dima Dorfman | Jul 13, 2001 6:55 am | |
| Terry Lambert | Jul 13, 2001 11:39 am | |
| Dima Dorfman | Jul 14, 2001 1:52 am | |
| Peter Wemm | Jul 14, 2001 6:06 am | |
| Dima Dorfman | Jul 14, 2001 6:23 am | |
| Steve Kargl | Jul 14, 2001 7:47 am | |
| Terry Lambert | Jul 16, 2001 8:15 am | |
| Peter Pentchev | Jul 16, 2001 8:40 am | |
| Kris Kennaway | Jul 16, 2001 12:29 pm |
| Subject: | Re: Getting rid of libgmp | |
|---|---|---|
| From: | Terry Lambert (tlam...@mindspring.com) | |
| Date: | Jul 16, 2001 8:15:07 am | |
| List: | org.freebsd.freebsd-arch | |
Peter Wemm wrote:
Comments? Suggestions?
Benchmarks, proving that you increased, or at least did not injure performance with this change?
This isn't really relevant. There are only a couple of things that use it. Namely the secure rpc key generators, the secure diffie hellman rpc key exchange, and telnet SRA key exchange at startup. None of these use it more than once (or once per connection).
telnet is already linked against libcrypto. It should be using that for bignum support instead of libmp.
libmp is dead. libcrypto is the interface of choice to use these days, or libgmp. Nothing in our tree uses libgmp.
We currently can do 600 1024 bit SSL connections a second, and expect to double that via interface changes.
So performance _is_ relevent.
-- Terry
To Unsubscribe: send mail to majo...@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message





