Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-ID: <413397BB.2020006@melbourne.sgi.com>
Date: Tue, 31 Aug 2004 07:10:19 +1000
From: David Chatterton <chatz AT melbourne DOT sgi DOT com>
Reply-To: chatz AT melbourne DOT sgi DOT com
Organization: SGI
User-Agent: Mozilla Thunderbird 0.6 (Windows/20040502)
MIME-Version: 1.0
To: Chuck McDevitt <cmcdevitt AT ABINITIO DOT COM>
CC: cygwin AT cygwin DOT com
Subject: Re: Cygwin1.dll problem with Hyperthreaded machines.
References: <OFDBD11567 DOT 72234441-ON85256F00 DOT 00560135-85256F00 DOT 00564CBC AT abinitio DOT com>
In-Reply-To: <OFDBD11567.72234441-ON85256F00.00560135-85256F00.00564CBC@abinitio.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Chuck,

We have a build environment for a reasonably large piece of code
that several people could be building at the same time. On a dual
800MHz P3 and even a quad 550MHz box we had no problems, but upgrading
to a dual 1.8GHz AMD box improved build times dramatically but
the vast majority of builds fail with:

- Microsoft's compiler reporting missing source file. I suspect
   the environment and/or current working directory is not
   setup correctly everytime a prcess is created. Rerun make and
   it works. Very strange.

- Thread activation errors which kill the build and the running
   shell.

rsync is also unreliable as it regularly hangs, but its unclear if
that is related.

I will also investigate providing a box, but that might be the
easy part. Providing a test case that does not rely on our code
base might be more of a challenge.

David

Chuck McDevitt wrote:

> Here at the company I work for, we have one of our developers spending 
> half his time trying to track down this problem, but so far with no luck.
> He did find that forcing processor affinity helps avoid the problem, but 
> that's not much of a solution.
> 
> I keep trying to convince management to send a hyperthreaded machine that 
> shows the problem to Christopher or Corinna, but they keep nixing the idea 
> (I don't understand why).


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/