Path: utzoo!mnetor!uunet!steinmetz!vdsvax!barnett From: barnett@vdsvax.steinmetz.ge.com (Bruce G. Barnett) Newsgroups: comp.arch Subject: Re: Tasting of Dhrystone 2.0 Results Message-ID: <4074@vdsvax.steinmetz.ge.com> Date: 15 Mar 88 17:30:57 GMT References: <503@pcrat.UUCP> Reply-To: barnett@vdsvax.steinmetz.ge.com (Bruce G. Barnett) Organization: General Electric CRD, Schenectady, NY Lines: 21 In article <503@pcrat.UUCP> rick@pcrat.UUCP (Rick Richardson) writes: |(Actually, the whole thing needs 160 columns if I include |date and submitter fields; I figured the 132 column format |would be printable by more people. I can be convinced to |post the 160 column format by a preponderance of the net) Maybe you should post the submitter. Some of the entries don't make sense. There is one case when a Sun 3/50 is faster than a Sun 3/280. There are also entries for Sun models that don't exist. I am also confused by the 5 entries for the Sun 3/50 going from 2344 to 3271. None of those entries mention whether the 68881 chip was used or not. Not to knock what Rick is doing, if there is that much variance without explanation, how useful is this data? Can anything be done to remove this variance? -- Bruce G. Barnett uunet!steinmetz!barnett