prehistoric, via dongarra From estevax.uucp!weicker Tue May 29 08:16:05 +0200 1990 Received: by pyxis; Tue May 29 03:29 EDT 1990 Received: by inet.att.com; Tue May 29 03:29 EDT 1990 Received: from estevax by unido.informatik.uni-dortmund.de with UUCP (UNIDO-2.0.1.j) via EUnet for research.att.com id AJ27998; Tue, 29 May 90 09:02:04 +0100 Date: Tue, 29 May 90 08:16:05 +0200 From: weicker@estevax.uucp (Reinhold Weicker) Message-Id: <9005290616.AA17176@estevax.uucp> Received: by estevax.uucp (5.51/3.14) Siemens Erlangen id AA17176; Tue, 29 May 90 08:16:05 +0200 To: ehg@research.att.com, petter@nac.no Subject: Dhrystone benchmark mailing, and other netlib / benchmark questions Dear Eric, dear Petter, Following this mail, I'ld like to send to both of you the current version (2.1) of the Dhrystone benchmark. Thank you for your earlier mails of May 17 and May 18, and please excuse me for being late. ------m houxm!vaximile!rer uunet!pcrat!rick Tue Jun 5 17:03:16 EDT 1990 From-: Eric Grosse ehg@research.att.com 201-582-5828 To-: houxm!vaximile!rer uunet!pcrat!rick Date-: Tue Jun 5 16:58:41 EDT 1990 Subject: dhrystones in netlib Dear Dr. Richardson No doubt you know that your version of the dhrystones has been distributed by the netlib email server. Dr. Weicker has just sent me updated versions, and to be absolutely sure that there is no controversy, I thought I'd check with you as I install these. You can get the latest versions by saying mail research!netlib send dhry-c from benchmark send dhry-ada from benchmark send dhry-pascal from benchmark. Please let me know if you approve of these changes or if you otherwise see anything I ought to fix before installing in the other copies of netlib. As the index for benchmark indicates, the old benchmark files are still available for those who need to reproduce old tests. Best wishes Eric Grosse From coma!att!pcrat!rick Tue Jun 5 17:47 EDT 1990 Received: by pyxis; Wed Jun 6 07:57 EDT 1990 From: coma!att!pcrat!rick Date: Tue, 5 Jun 90 17:47 EDT To: att!research!ehg Subject: Stones Eric: I'm not a Dr., just a Mr. Sigh, these days you have to make economic choices, and my wife got the choice to go for a Dr. Haven't used that houxm!vaximile!rer address in, gosh, 3-4 years I guess. I wonder if there's this giant mail file growing there (what with sysadms not always being on top of things). uunet!pcrat!rick is best, but att!pcrat!rick also works, but slower. Anyway, I'm going to pull the C version to check it against what I've got (they should be the same, I think, modulo maybe some makefile tweaks). Reinhold and I spent some time making sure that there weren't two versions of 2.X. With 1.X, there wasn't a C version initially; I did one translation and distributed it widely on the net, while Reinhold did another translation. At this point, my only contributions to the effort are to collect results for 2.X. If you'd like to archive the 2.X results collection, I can send them to you as well. I haven't messed with Ada or Pascal, so whatever you got from Reinhold is the Gospel. Actually, *anything* you get from Reinhold is the Gospel, since it is his baby. I'll be getting the netlib copy tonight around 9PM, when att calls me, and I'll try to respond ASAP after that. -Rick Rick Richardson | Looking for FAX software for UNIX/386 ??? Ask About: |Mention PC Research,Inc.| FaxiX - UNIX Facsimile System (tm) |FAX# for uunet!pcrat!rick| FaxJet - HP LJ PCL to FAX (Send WP,Word,Pagemaker...)|Sample (201) 389-8963 | JetRoff - troff postprocessor for HP LaserJet and FAX|Output From coma!att!pcrat!rick Thu Jun 7 01:26 EDT 1990 Received: by pyxis; Thu Jun 7 01:42 EDT 1990 From: coma!att!pcrat!rick Date: Thu, 7 Jun 90 01:26 EDT To: att!coma!ehg uunet!mcsun!unido!estevax!weicker Subject: Dhyrstone Eric, Reinhold: 41,42c41,42 < * Phone: (201) 389-8963 (9-17 EST) < * Usenet: ...!uunet!pcrat!rick --- > * Phone: (201) 834-1378 (9-17 EST) > * Usenet: ...!seismo!uunet!pcrat!rick In dhry.h, it would be good of you to fix the address info. The first two lines in the above diff are correct. The only source code differences between the copy I have and the copy Reinhold and netlib has is that mine includes #ifdef'ed code for measuring the time under MS-DOS, #define Too_Small_Time is set to (2*HZ) rather than 120, and I also have a Makefile. Nothing is different in the actual stones code. -Rick ------m att!pcrat!rick uunet!mcsun!unido!estevax!weicker Sun Jun 10 13:28:41 EDT 1990 From-: Eric Grosse ehg@research.att.com 201-582-5828 To-: att!pcrat!rick uunet!mcsun!unido!estevax!weicker Date-: Sun Jun 10 13:22:45 EDT 1990 Subject: dhrystones in netlib Rick, I changed your phone number and usenet address in the files here. Too_Small_Time, however, I left alone because the name of HZ seems to have changed in ANSI C to CLK_TCK in most drafts and then CLOCKS_PER_SECOND in the final standard. It would have been nice to have your Makefile in dhry-c, but I imagine people can cook it up without help. Recently I noticed that NIST, (formerly the US National Bureau of Standards) is running a variant of netlib that also distributes the dhrystones, along with a zillion other benchmarks. I'll try to find out who runs this and alert them to your new version. Best wishes Eric .