Article View: cu.slug
Article #536lucid on the SUN Ultra
From: Paul Martin - Sun Microsystems L...
Date: Tue, 30 Apr 1996 00:00
Date: Tue, 30 Apr 1996 00:00
24 lines
1571 bytes
1571 bytes
I just finally got a chance to run some code on a SUN Ultra Sparc 1, the most modest member of the new 64-bit line. I'm in SUN Labs, and you all know that the shoemaker's kids go barefoot.... The Lisp I'm running is Lucid from before their demise. It crashes randomly and pretty early in the load-up. I suspect it is because the instruction cache and data cache of the processor don't talk to each other; from the point of view of c-based hardware builders, Lisp is committing the sin of self-modifying code every time it loads a bin file. So I've handled this on the Sparc5 by using a patch that flushes the instruction cache after every fasload, but I'm unclear on whether this patch will do the trick for the Ultra. It appears to crash before the patch could load (or sometimes during loading it), and I wonder if the op code used to flush the cache is even the same. Any experience with the Ultra and Lucid lurking out there? Thanks. Paul Martin
Technical Details
Message-ID:
<830875497.10610@news.Colorado.EDU>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10077e,f5f9ca1ab9f3e6f4,start X-Google-Attributes: gid10077e,public From: Paul.Martin@East.Sun.COM (Paul Martin - Sun Microsystems Labs BOS) Subject: lucid on the SUN Ultra Date: 1996/04/30 Message-ID: <830875497.10610@news.Colorado.EDU> X-Deja-AN: 152279397 distribution: cu sender: daemon@lace.colorado.edu x-note1: message-id generated by recnews x-note2: mail msgid was <199604301419.KAA02024@juniper.East.Sun.COM> x-sun-charset: US-ASCII organization: University of Colorado at Boulder newsgroups: cu.slug