Date: Thu, 13 Feb 1997 16:35:45 -0800 (PST) From: Jake Hamby <hamby@aris.jpl.nasa.gov> To: Satoshi Asami <asami@vader.cs.berkeley.edu> Cc: hackers@FreeBSD.ORG Subject: Re: Sun Workshop compiler vs. GCC? Message-ID: <Pine.GSO.3.95.970213161306.10210C-100000@aris> In-Reply-To: <199702132125.NAA18583@vader.cs.berkeley.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 13 Feb 1997, Satoshi Asami wrote: > I'm optimistic about this. My understanding is that the slowness and > number of bugs of Solaris is intrinsic to its complexity of design > (and also the fact that it was designed for workstations in mind, > initially). You just can't make a huge mammoth run fast, no matter > how much cash you sink into the compiler. Compiler optimization is no > cure for bad design. I didn't mean to imply that Sun is *only* tuning the compiler. The biggest improvements I know of will come through extensive use of "doors" (an IPC mechanism from Spring, their concept OS, and much faster than SYSV messaging, pipes, RPC, or any other UNIX IPC), which can pass messages in a couple of microseconds, and kernel sockets (and other TCP/IP improvements bringing the API in compliance with 4.3BSD), which provides up to 25% better Web server performance, or up to 2-3X improvement with tuning. Furthermore, they're planning to release their own Webserver for free, which should be multithreaded and VERY performance tuned, since they have Apache and Netscape server to learn from. I don't want to sound like an advertisement for Sun, it's just that they're so Buzzword Enriched, they will make people believe they have the best server platform. And it actually benefits them that the current version of Solaris has relatively nasty TCP/IP performance, because Solaris 2.6 will just look that much better in comparison. > Besides, Solaris x86 is such a festering piece of crap I can't believe > anyone actually using it for "serious" work. I'm now having sooooo > much fun trying to connect a bunch of disks to it (I need to yank and > reinsert cables at the right moment during boot, can't have an IDE disk > in the system if there is a SCSI disk with ID > 7, can't have three > SCSI adapters active at the same time or the system will hang, etc.). Solaris x86 is a piece of crap for device configuration, I agree. The autotuning kernel is a wonderful concept, but non-Plug-and-Play hardware completely thwarts it. I had the worst time getting an NE2000 card to be recognized, but that's probably the _worst_ card to detect so I can cut them a little slack. Again, *if* Solaris 2.6 gets this area fixed, a lot of people are going to give it a second look, especially if they combine it with a big ad campaign. > FreeBSD beats Solaris hands down in every aspect (reliability, ease of > configuration (/devices/pci@0,0/pci1011,1@f/pci9004,7278@4/cmdk@0,0:q,raw > is not my idea of simplicity), performance). I'd say Solaris is potentially more reliable because Sun has been working on it for 14 years (if you include time spent developing SunOS). You just wouldn't know that if you have buggy x86 drivers. I know there are SPARCs with uptimes in the _years_. Solaris is *potentially* easier to configure if you have all Plug-and-Play hardware, again the x86 version needs lots of improvement. Also, the Netra systems are prebundled with Web server, POP/IMAP, etc., although that's not available for x86, it shows the direction they'll be moving. Finally, there are a few areas where Solaris beats FreeBSD in performance, and vice versa, depending on who you talk to (for example, I remember on the NetBSD list, some users said NetBSD/SPARC was twice as fast as Solaris on certain math computations because of some register window effects, while Solaris had a significantly faster filesystem). Obviously, you should choose whichever OS best fits your needs. I didn't bring up Solaris to start a flamewar, just to point out that it offers some compelling advantages for an academic user, such as myself, who uses Suns at work and school, and their new compiler sounded intriguing. They are definitely the UNIX vendor to beat. Also, both FreeBSD and Linux have _lousy_ Java support (actually FreeBSD is better than Linux because we have a working JDK 1.0.2 that doesn't require shared library hell to set up), compared to Solaris, obviously. That's why it'd be nice if we could get SVR4 emulation support some day. ------------------------------------------------------------------------------ |Jake Hamby| APT Engineer at JPL, CS student at Cal Poly, and BeOS developer!| ------------------------------------------------------------------------------ "Life is hard..."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.GSO.3.95.970213161306.10210C-100000>