home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!stanford.edu!agate!ucbvax!SDD.HP.COM!ken
- From: ken@SDD.HP.COM (Ken Stone)
- Newsgroups: comp.protocols.time.ntp
- Subject: Re: Reducing distribution clutter
- Message-ID: <9208171915.AA20931@hpsdlz.sdd.hp.com>
- Date: 17 Aug 92 19:15:55 GMT
- References: <9208161609.aa21356@huey.udel.edu>
- Sender: daemon@ucbvax.BERKELEY.EDU
- Distribution: inet
- Organization: The Internet
- Lines: 27
-
-
- > Splitting the milliseconds is easy with the '240 birds. Send me your
- > additions to this list (use: authspeed auth.samplekeys in the authstuff
- > subdirectory) and I'll award prizes.
-
- Ok, first question is ... has anyone ever thought about cleaning up this
- code to use nice thing like "volatile" so that at leats semi-intelligent
- optimizers and make semi-intelligent decisions :-) ??? If I take the
- standard conservative approach and compile using +O1, I get numbers like
-
- authentication delays (us) DES MD5
- HP 730 running 8.07 (+O1) 18 87
-
- but If I get brave and crank the optimizer up to +O3, I can clean up the
- MD5 number a bit to match the DES
-
- authentication delays (us) DES MD5
- HP 730 running 8.07 (+O3) 16 50
-
- Now have I tried building and running xntpd at +O3 ? ... nope. Do I think
- it would work ? ... nope. At least with the v2 code, funny things seemed
- to start happening at +O2.
-
- So I guess 18/87 needs to be my entry for now.
-
- -- Ken
-
-