eric@plus5.UUCP (Eric W. Kiebler) (05/26/85)
I have heard it said that the relative performance of the 3b machines can be summed up as follows: the number following the '3b' part is the number of users. While this smacks of nonsense, it has been my limited experience that this may in fact be correct. Has anyone any experience with the 3b2, 3b5, and 3b20 processors running the following types of loads: development -- concurrent compilation, editting, linking, etc. word proc. -- no development, nroff-troff word proc. -- no development, WhatYouSeeIsWhatYouGet programs comm appls -- uucp server *EDUCATED* guesses are welcome, though hard facts are more welcome. Estimates in terms of "... its faster than our Bambleweeny 57 ..." are most welcome. eric PS. This is a personal investigation, not a corporate one. -- ..!ihnp4!wucs!plus5!eric ..!ihnp4!plus5!eric (314) 725-9492
amc@mhuxl.UUCP (COHILL) (05/28/85)
> Has anyone any > experience with the 3b2, 3b5, and 3b20 processors running the following > types of loads: > > development -- concurrent compilation, editting, linking, etc. > word proc. -- no development, nroff-troff > word proc. -- no development, WhatYouSeeIsWhatYouGet programs > comm appls -- uucp server > We have fifteen 3B2s in daily use here. One machine is used for UNIX training, and we frequently have ten to twelve people using *vi* or *20/20* (a very good spreadsheet). Response time is not blinding, but it is not irritating. In terms of price/performance, the 3B2s win hands down over the 3B5 and the 3B20. We also have those units here too. Andy Cohill AT&T Technology Systems Springfield Operations {allegra,ihnp4}mhuxl!amc
shelby@rtech.UUCP (Shelby Thornton) (05/30/85)
> I have heard it said that the relative performance of the 3b machines > can be summed up as follows: the number following the '3b' part is > the number of users. The person who came up with that must have been an optimist :-) > development -- concurrent compilation, editting, linking, etc. On a 2 or 5, don't bother. > word proc. -- no development, nroff-troff > word proc. -- no development, WhatYouSeeIsWhatYouGet programs The 2 is suppose to be o.k., forget the 5. > comm appls -- uucp server If you don't mind it being slow... Notice that I made no mention of a 3b20. It's suppose to be a decent machine but I have had no exposure to it. The 3b5 makes a great paper- weight and the 3b2 makes a cute little toy. ucbvax!mtxinu!rtech!shelby Shelby Thornton amdahl!rtech!shelby Relational Technology Inc.
sean@ukma.UUCP (Sean Casey) (06/02/85)
We have 2 3b2s. My experience is that they are about 2/3 of a Vax 11/750 speedwise, excepting floating point. A simple floating point benchmark that did the basic 4 (+ - / *) operations in a loop gave an average speed of less than 200 floating point operations per second. -- - Sean Casey UUCP: {cbosgd,anlams,hasmed}!ukma!sean - Department of Mathematics ARPA: ukma!sean@ANL-MCS.ARPA - University of Kentucky
hoffman@pitt.UUCP (Bob Hoffman) (06/07/85)
> I have heard it said that the relative performance of the 3b machines > can be summed up as follows: the number following the '3b' part is > the number of users. According to our local AT&T-IS rep, the number stands for the number of thousands of telephone circuits the processor was designed to switch, e.g. 3b2 = 2000 lines, 3b5 = 5000 lines, etc. Remember that these things have been in use in COs for quite a while and have only recently been sold as general purpose computers. -- Bob Hoffman, N3CVL Pitt Computer Science