Date: Wed, 17 Mar 1999 11:17:20 +1030 From: Greg Lehey <grog@lemis.com> To: Sascha Schumann <sas@schell.de>, freebsd-chat@FreeBSD.ORG Subject: Real performance comparisons (was: FreeBSD 3.1 SMP outperforms SuSE 6.0 SMP by factor 2.3 !!!) Message-ID: <19990317111720.O429@lemis.com> In-Reply-To: <19990316150715.A3316@schell.de>; from Sascha Schumann on Tue, Mar 16, 1999 at 03:07:15PM %2B0100 References: <19990316150715.A3316@schell.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday, 16 March 1999 at 15:07:15 +0100, Sascha Schumann wrote: > On Mon, Mar 15, 1999 at 05:13:20PM +1030, Greg Lehey wrote: >> I tend to agree. I could accept the statement "Linux SMP gives less >> performance than FreeBSD SMP". I can`t accept the statement "Linux >> SMP gives no performance improvement". > > Well, I could accept it, if it would be technically correct. But Mr Klemm's > claims are simply invalid. I probably do not need to reiterate on that fact. > > Alternatively, I will claim that Linux is 1200% faster, since I can calculate > DES keys at a rate of 10M/sec on my Linux box (Alpha PC164UX, 633MHz) and only > 0.8M/sec on my FreeBSD box (Intel Pentium, 166MHz) (yes, it is non sense). > >> The first thing that came into >> my mind was "he must have made a mistake". Other possibilities are >> that make -j doesn`t work right on Linux, or that Linux SMP doesn`t >> even the load as well as FreeBSD does (for example, some MP systems >> put all children in the same CPU). > > Linux distributes the load across the various CPUs. If you want to > know which system scales better for your needs, test it directly > with your application. I believe this is exactly what Andreas did. For various reasons, it didn't work. > Do not rely on someone who does not even tell you how he produced > his results. Andreas was quite specific about his results. The only mistake he made (and one about which I commented) was that he initially accepted the fact that he got *no* performance improvement from the SMP implementation. This is at least a documentation problem of the SuSE distribution. As somebody else said, it *would* be interesting to compare Linux and FreeBSD performance objectively and in detail. The only comparison of this nature that we have at the moment is from the Gartner group (http://advisor.gartner.com/n_inbox/hotcontent/hc_2121999_3.html#h8), which shows that FreeBSD outperforms Linux by in the order of 40% in certain server applications. While I don't doubt the validity of this information, it would be interesting to research where the differences lie, and the extent to which such measurements are typical. Greg -- See complete headers for address, home page and phone numbers finger grog@lemis.com for PGP public key To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990317111720.O429>