Home » Mailing list archives » linux-kernel » 2007 » October » 25

Re: Is gcc thread-unsafe?

Previous thread: Re: - mmconfig-validate-against-acpi-motherboard-resources.patch removed from -mm tree by Robert Hancock on Thursday, October 25, 2007 - 7:20 pm. (35 messages)

Next thread: [PATCH 1/3] ide: remove IRQF_DISABLED from IRQ flags for IDE IRQ handler by Bartlomiej Zolnierkiewicz on Thursday, October 25, 2007 - 7:35 pm. (1 message)
[view original message]
From: Robert Hancock <hancockr@...>
To: Arjan van de Ven <arjan@...>
Cc: <davids@...>, Linux-Kernel@Vger. Kernel. Org <linux-kernel@...>
Subject: Re: Is gcc thread-unsafe?
Date: Thursday, October 25, 2007 - 7:27 pm

It is a win if the branch cannot be effectively predicted, i.e. if the

outcome is essentially random, as may occur with data-dependent

conditionals. I've seen a doubling of performance on one workload using

a predicated instruction instead of a branch on newer Xeons in such a case.


I suspect that if branch prediction fails often, the data dependency

created by the cmov, etc. is less expensive than the pipeline flush

required by mispredicts..


--

Robert Hancock      Saskatoon, SK, Canada

To email, remove "nospam" from hancockr@nospamshaw.ca

Home Page: http://www.roberthancock.com/


-
Previous thread: Re: - mmconfig-validate-against-acpi-motherboard-resources.patch removed from -mm tree by Robert Hancock on Thursday, October 25, 2007 - 7:20 pm. (35 messages)

Next thread: [PATCH 1/3] ide: remove IRQF_DISABLED from IRQ flags for IDE IRQ handler by Bartlomiej Zolnierkiewicz on Thursday, October 25, 2007 - 7:35 pm. (1 message)