From mboxrd@z Thu Jan 1 00:00:00 1970 Path: news.gmane.org!not-for-mail From: ludo@gnu.org (Ludovic =?iso-8859-1?Q?Court=E8s?=) Newsgroups: gmane.lisp.guile.devel Subject: Re: data-crunching in guile Date: Thu, 25 Jun 2009 10:04:42 +0200 Message-ID: <86r5x891t1.fsf@gnu.org> References: NNTP-Posting-Host: lo.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: ger.gmane.org 1245917303 23892 80.91.229.12 (25 Jun 2009 08:08:23 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Thu, 25 Jun 2009 08:08:23 +0000 (UTC) To: guile-devel@gnu.org Original-X-From: guile-devel-bounces+guile-devel=m.gmane.org@gnu.org Thu Jun 25 10:08:16 2009 Return-path: Envelope-to: guile-devel@m.gmane.org Original-Received: from lists.gnu.org ([199.232.76.165]) by lo.gmane.org with esmtp (Exim 4.50) id 1MJk0F-0006FN-My for guile-devel@m.gmane.org; Thu, 25 Jun 2009 10:08:16 +0200 Original-Received: from localhost ([127.0.0.1]:48765 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MJk0F-0005gQ-6Y for guile-devel@m.gmane.org; Thu, 25 Jun 2009 04:08:15 -0400 Original-Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1MJjzY-0005Os-4e for guile-devel@gnu.org; Thu, 25 Jun 2009 04:07:32 -0400 Original-Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1MJjzT-0005N6-7v for guile-devel@gnu.org; Thu, 25 Jun 2009 04:07:31 -0400 Original-Received: from [199.232.76.173] (port=38424 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MJjzS-0005Mv-WE for guile-devel@gnu.org; Thu, 25 Jun 2009 04:07:27 -0400 Original-Received: from main.gmane.org ([80.91.229.2]:51843 helo=ciao.gmane.org) by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1MJjzS-0000Ai-9P for guile-devel@gnu.org; Thu, 25 Jun 2009 04:07:26 -0400 Original-Received: from list by ciao.gmane.org with local (Exim 4.43) id 1MJjzP-0006CO-4V for guile-devel@gnu.org; Thu, 25 Jun 2009 08:07:23 +0000 Original-Received: from 193.50.110.50 ([193.50.110.50]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Thu, 25 Jun 2009 08:07:23 +0000 Original-Received: from ludo by 193.50.110.50 with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Thu, 25 Jun 2009 08:07:23 +0000 X-Injected-Via-Gmane: http://gmane.org/ Original-Lines: 28 Original-X-Complaints-To: usenet@ger.gmane.org X-Gmane-NNTP-Posting-Host: 193.50.110.50 X-URL: http://www.fdn.fr/~lcourtes/ X-Revolutionary-Date: 7 Messidor an 217 de la =?iso-8859-1?Q?R=E9volution?= X-PGP-Key-ID: 0xEA52ECF4 X-PGP-Key: http://www.fdn.fr/~lcourtes/ludovic.asc X-PGP-Fingerprint: 821D 815D 902A 7EAB 5CEE D120 7FBA 3D4F EB1F 5364 X-OS: i486-pc-linux-gnu User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.0.91 (gnu/linux) Cancel-Lock: sha1:CHOQWghosW3ZUiaPg8JkY5KMnTk= X-detected-operating-system: by monty-python.gnu.org: GNU/Linux 2.6, seldom 2.4 (older, 4) X-BeenThere: guile-devel@gnu.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "Developers list for Guile, the GNU extensibility library" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Original-Sender: guile-devel-bounces+guile-devel=m.gmane.org@gnu.org Errors-To: guile-devel-bounces+guile-devel=m.gmane.org@gnu.org Xref: news.gmane.org gmane.lisp.guile.devel:8765 Archived-At: Hi, Andy Wingo writes: > What I'm getting at is that I think we should have VM ops for working on > vectors -- both generic vectors, and specific ops for bytevectors, and > probably an op for string-ref as well, and possibly string-set!. Then a > native code backend could be effectively implemented to operate on the > GLIL or assembly level, relying on the Tree-IL compiler's previous > resolution of high-level operations (i.e., vector-set!) to low-level > instructions. I think we have the space in the VM, and if we group all > of the vector instructions at the end, we shouldn't affect the > instruction cache too much. Why not, but... It looks to me like a failure to optimize the general case, which we work around by specializing the instruction set. Maybe we could have a more generic approach to the problem. For instance, there could be a way to annotate primitive procedures with information useful to the compiler, such as "returns-single-value", "return-value-unspecified", etc., which would allow the generated code to take shortcuts. But maybe this is just unrealistic. What do you think? Thanks, Ludo'.