Path: utzoo!censor!geac!jtsv16!uunet!tut.cis.ohio-state.edu!ucbvax!janus!kennish From: kennish@janus.uucp (Ken A. Nishimura) Newsgroups: comp.sys.dec Subject: Re: Questions on 5400 systems Keywords: 5400, vax, risc, ultrix, 3000, dec Message-ID: <30464@ucbvax.BERKELEY.EDU> Date: 8 Aug 89 06:49:40 GMT References: <106@pserv.UUCP> Reply-To: kennish@janus.UUCP (Ken A. Nishimura) Distribution: usa Organization: University of California, Berkeley Lines: 36 We have a 5400 in the cabinet (not floor pedestal) type case. I have run some i/o benchmarks against the DecStation 3100. Preliminary results indicate that the RA90 with the DecSystem 5400 is 4.87 times as fast as a DecStation 3100. Now, the DecStation 3100 is about 4 times as fast as a uVAX II/GPX with RD54 drives, or about 0.8 times as fast as a VAX8800 with RA81 drives. Therefore, the 5400 with RA90's can be considered to be quite fast. As far as porting software for the 5400 - it is 100% binary compatible (DEC claim, yet to be disproven here) with the 3100 except for the OS (device drivers). Hence, any DS3100 application should run on the 5400. As far as the OS, well, it's quite similar to the OS for Vaxen, but there are some differences. As we are field testing Ultrix for both Vax and Risc both seem about the same number of problems, although the bugs are not identical. Overall, I must say that Ultrix 3.1 is pretty good. (please no flames!) As far as CPU performance of the 5400 vs. the 3100, it is about 15% faster,o the 5400 is about 17 times a uVaxII. FP performance is about 18% faster. My opinion -- for CPU, the best bang for your buck is the 2100 (list prices), 3100 (university prices). If I/O performance is y (i.e. Ingres/Databases) then the 5400 is worth considering since its so much faster than the 3100 for I/O. -ken I usually don't post a disclaimer, but since this message contains numerical data and opinions....... "All data and statements made above are unofficial and strictly the opinion of the author. No connection with the University of California is expressed or implied."