Path: utzoo!utgpu!watmath!clyde!ima!spdcc!dyer From: dyer@spdcc.COM (Steve Dyer) Newsgroups: comp.sys.ibm.pc.rt Subject: Re: Why the original RT seemed/was slow (was ...) Message-ID: <2213@spdcc.COM> Date: 5 Dec 88 00:17:47 GMT References: <5046@polya.Stanford.EDU> <1287@auschs.UUCP> <1309@auschs.UUCP> <3736@pt.cs.cmu.edu> <447@scifi.UUCP> <3764@pt.cs.cmu.edu> Reply-To: dyer@spdcc.COM (Steve Dyer) Organization: S.P. Dyer Computer Consulting, Cambridge MA Lines: 27 Just FYI, in running Dhrystone 2.1, here are a few figures for the model 25 and model 125 under the latest September AOS 4.3 with the stock "cc" (some rev of "hc" which I am not sure of), another version of "hc" (older I hope): "hc1.4", and "pcc". The same binary images were run on both. Model 25 Model 125 hc hc1.4 pcc hc hc1.4 pcc No Reg 3075 2201 1623 7264 --- 3333 Reg 3075 2201 1831 7281 3685 I didn't bother with rerunning the hc1.4 on the APC. Just more evidence that a good compiler is critical for this machine. If Sauer's numbers are to be believed, it's a shame that the AIX C compiler technology is not available to the AOS group in Palo Alto (and therefore to us). You can compare these results with an Intel Inboard 386/AT at 16mhz & 64kb cache running under SCO XENIX 386 2.2.3: 4587 (no reg), 4893 (reg). -- Steve Dyer dyer@harvard.harvard.edu dyer@spdcc.COM aka {harvard,husc6,linus,ima,bbn,m2c,mipseast}!spdcc!dyer