Date: Sat, 8 Jan 2000 13:38:28 -0500 (EST) From: Garrett Wollman <wollman@khavrinen.lcs.mit.edu> To: Mohit Aron <aron@cs.rice.edu> Cc: wollman@khavrinen.lcs.mit.edu (Garrett Wollman), freebsd-current@freebsd.org Subject: Re: performance of FreeBSD-current as SMP Message-ID: <200001081838.NAA51522@khavrinen.lcs.mit.edu> In-Reply-To: <200001081834.MAA14332@cs.rice.edu> References: <200001081832.NAA51499@khavrinen.lcs.mit.edu> <200001081834.MAA14332@cs.rice.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
<<On Sat, 8 Jan 2000 12:34:24 -0600 (CST), Mohit Aron <aron@cs.rice.edu> said: > Well, some difference is to be expected. But 22% ???? Flaws in the Intel SMP design make it more expensive than it ought to be. Things like atomic RMW cycles are very, very expensive. (We had a discussion about six months ago about how expensive something like a locked xchg instruction actually is.) -GAWollman -- Garrett A. Wollman | O Siem / We are all family / O Siem / We're all the same wollman@lcs.mit.edu | O Siem / The fires of freedom Opinions not those of| Dance in the burning flame MIT, LCS, CRS, or NSA| - Susan Aglukark and Chad Irschick To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200001081838.NAA51522>