This is the mail archive of the
mailing list for the GCC project.
Re: [patch] Privatize gimplify_ctx structure.
- From: Jeff Law <law at redhat dot com>
- To: Jakub Jelinek <jakub at redhat dot com>, Andrew MacLeod <amacleod at redhat dot com>
- Cc: gcc-patches <gcc-patches at gcc dot gnu dot org>, Richard Biener <richard dot guenther at gmail dot com>, Diego Novillo <dnovillo at google dot com>
- Date: Wed, 20 Nov 2013 09:51:54 -0700
- Subject: Re: [patch] Privatize gimplify_ctx structure.
- Authentication-results: sourceware.org; auth=none
- References: <528CBAA0 dot 5060801 at redhat dot com> <20131120134434 dot GH892 at tucnak dot redhat dot com>
On 11/20/13 06:44, Jakub Jelinek wrote:
I was going to suggest hacking things so that he only had 2, then do a
bootstrap/comparison to test the fallback. Obviously that would be for
testing purposes only.
On Wed, Nov 20, 2013 at 08:35:28AM -0500, Andrew MacLeod wrote:
I also hacked up the compiler to report what the 'top' of the stack
was for a compilation unit. I then ran it through a bootstrap and
full testsuite run of all languages, and looked for the maximum
number of context structs in use at one time. Normally only 1 or 2
were ever in use, but its topped out at 7 during some of the openmp
tests. So my current limit of 30 will likely never been reached.
??? You can trivially create a testcase where it will be reached
(and you certainly should, so that you actually test the fallback).
However, a test we can put into the suite is even better :-)