This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: The nvptx port

From: Jeff Law <law at redhat dot com>
To: Jakub Jelinek <jakub at redhat dot com>, Bernd Schmidt <bernds at codesourcery dot com>
Cc: Richard Henderson <rth at redhat dot com>, gcc-patches at gcc dot gnu dot org, Thomas Schwinge <thomas_schwinge at mentor dot com>, Cesar Philippidis <cesar_philippidis at mentor dot com>
Date: Fri, 14 Nov 2014 10:04:43 -0700
Subject: Re: The nvptx port
Authentication-results: sourceware.org; auth=none
References: <20141114082948 dot GH5026 at tucnak dot redhat dot com> <5465E2CF dot 6000105 at codesourcery dot com> <20141114113945 dot GM5026 at tucnak dot redhat dot com> <5465F1B8 dot 80205 at codesourcery dot com> <20141114123653 dot GS5026 at tucnak dot redhat dot com>

On 11/14/14 05:36, Jakub Jelinek wrote:


So, for a warp, if some threads perform one branch of an if and other
threads another one, all threads perform the first one first (with some
maybe not doing anything), then all the threads the others (again, other
threads not doing anything)?

Nobody ever specified exactly what happens in this case to me, but Igathered from reading the docs that once you have some threads in onepath and others in a different path, things slow down to a horrid crawl.So you try to avoid that :-)



jeff

Follow-Ups:
- Re: The nvptx port
  - From: Nathan Sidwell

References:
- The nvptx port
  - From: Jakub Jelinek
- Re: The nvptx port
  - From: Bernd Schmidt
- Re: The nvptx port
  - From: Jakub Jelinek
- Re: The nvptx port
  - From: Bernd Schmidt
- Re: The nvptx port
  - From: Jakub Jelinek

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]