Return to BSD News archive
Xref: sserve comp.unix.bsd:15542 comp.unix.solaris:28179 Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!newshost.marcam.com!news.mathworks.com!news.cic.net!locust.cic.net!pauls From: pauls@locust.cic.net (Paul Southworth) Newsgroups: comp.unix.bsd,comp.unix.solaris Subject: Re: Unix for PC Date: 11 Dec 1994 18:50:39 GMT Organization: CICNet, Inc. Lines: 66 Distribution: inet Message-ID: <3cfhlv$dbl@spruce.cic.net> References: <199411210319.TAA18133@nic.cerf.net> <3cd8fo$ns5@fido.asd.sgi.com> <3cdef1$q87@pdq.coe.montana.edu> <3cdkob$4t1@fido.asd.sgi.com> NNTP-Posting-Host: locust.cic.net In article <3cdkob$4t1@fido.asd.sgi.com>, Larry McVoy <lm@slovax.engr.sgi.com> wrote: >Go buy a copy of Solaris, install it on your >PC, and tell that it "just works". I'll bet you that Linux runs on >more hardware than Solaris by a long shot. And I'll bet if you plot >turn around time on bugs, Linux is outdoing Solaris by a factor of two >at least. Just a guess, do you have data to prove me wrong? I think Larry is right on this one. I wish that the data were available, but hey, I figure it's not my responsibility. Anyway, having taken two fairly generic PC's (one ISA, one EISA, DX2-50 clones, 16mb RAM and supported SCSI adapters with ATI VLB cards) and run Solaris 2.1 x86, 2.4 x86 EA1, EA2 and through the beta period to full release, including using OpenWindoze on ATI, CL and Orchid cards, running through building my full working environment of GNU tools, compilers, and popular freeware (plus building Mosaic since they gave me Motif) I was distinctly unimpressed with the performance and features of Solaris on Intel, especially with regard to filesystem speed, compared to Linux on the same hardware. Having set up NetBSD, FreeBSD and Linux on the same equipment, I found FreeBSD and Linux were both more stable than Solaris 2.4 (and faster). At the time NetBSD was in 1.0 beta and I had some troubles with it, but I haven't tried the 1.0 release and I wasn't using FreeBSD current so that's not really a comparison. With regard to filesystem I/O, I was so impressed by Linux (both on SCSI and IDE systems) that I'm tempted to reinstall Solaris 2.4 just to benchmark it. The Linux ext2fs is one of the fastest filesystems I've ever witnessed; prior to trying it I really had no clue that there was that kind of potential for speed in my hardware. I found it easier to do quick software ports to the *BSD systems, especially since it was easier to get knowledgeable answers from developers than it was for Linux. There are a lot of knowledgeable Linux people out there but getting their attention amidst the noise of the Linux newsgroups is nigh impossible -- I don't see that as a disparaging remark -- there is really a phenomenal amount of activity over there and a lot of it is focused on new user questions. By comparison, my interaction with the people responsible for Solaris beta 2.4 x86 bug handling (whose name escapes me but many of you probably remember...) was far from satisfactory. I got several "well it works for me" type responses, which are naturally less than helpful. I also submitted several bugs using the "approved" process that are still in the release version, even though I reported them before EA2 came out (anyone else tried using xdm from OpenWindows on 2.4 x86?). Additionally, when I reported bugs related to several SCSI adapters (not Adaptec) that I tried out with it (adapters that were listed as OK for Solaris 2.1 x86) I got no answers and now I see they were just pulled from the supported list (that's commitment, eh?). As opposed to the *BSD/Linux attitude which is more like "Oh it doesn't work? Well send me an adapter and I'll make it work". I see that as a much healthier attitude even though it places more of the responsibility on the user to help make it possible for the developer to advance the "product". Or alternately, the user says "Oh, it doesn't work? I'll fix it and send in the changes." The source, the source, the source. It really seems like there's a mental gap at Sun -- they're trying to reinvent the wheel and ending up with a horse-drawn carriage instead of a Ferrari. I'll take public transportation over that any day. -- Paul Southworth pauls@cic.net