[flang-dev] [EXTERNAL] Re: arm + denormals

Jost, Gabriele (ARC-TNC)[Supersmith] via flang-dev flang-dev at lists.llvm.org
Thu Apr 23 11:49:30 PDT 2020


Many thanks. I will give this a try. Yes, indeed, it is for the arm platform.


Gabriele Jost

************************************************************************
Gabriele Jost       NASA Ames Research Center
Email: Gabriele.Jost at nasa.gov      Mail Stop 258-6
Work Phone: (650) 604-0468         Bldg. 258, Rm. 232-4
Scientific & HPC Application           P.O. Box 1
Development/Optimization            Moffett Field, CA 94035-0001
************************************************************************

From: Kiran Chandramohan <Kiran.Chandramohan at arm.com>
Date: Thursday, April 23, 2020 at 11:37 AM
To: "Jost, Gabriele (ARC-TNC)[Supersmith]" <gabriele.jost at nasa.gov>
Cc: "flang-dev at lists.llvm.org" <flang-dev at lists.llvm.org>
Subject: Re: [EXTERNAL] Re: arm + denormals

Ahh.

What about Mflushz?
Also, Is this question specifically for the Arm platform?

--Kiran
________________________________
From: Jost, Gabriele (ARC-TNC)[Supersmith] <gabriele.jost at nasa.gov>
Sent: 23 April 2020 19:14
To: Kiran Chandramohan <Kiran.Chandramohan at arm.com>
Subject: Re: [EXTERNAL] Re: arm + denormals


Thanks for the reply, Kiran. So, the message comes from the code itself, not from the compiler. There are denormals, but if the compiler could flush those to 0, the code would not detect them, I think. The denormals don’t impact the results, just the timings, because the code checks for them.





Gabriele Jost



************************************************************************

Gabriele Jost       NASA Ames Research Center

Email: Gabriele.Jost at nasa.gov      Mail Stop 258-6

Work Phone: (650) 604-0468         Bldg. 258, Rm. 232-4

Scientific & HPC Application           P.O. Box 1

Development/Optimization            Moffett Field, CA 94035-0001

************************************************************************



From: Kiran Chandramohan <Kiran.Chandramohan at arm.com>
Date: Thursday, April 23, 2020 at 11:07 AM
To: "flang-dev at lists.llvm.org" <flang-dev at lists.llvm.org>, "Jost, Gabriele (ARC-TNC)[Supersmith]" <gabriele.jost at nasa.gov>
Subject: [EXTERNAL] Re: arm + denormals



Hello Gabriele,

Does setting the environment variable NO_STOP_MESSAGE help?
export NO_STOP_MESSAGE=1

--Kiran

________________________________

From: flang-dev <flang-dev-bounces at lists.llvm.org> on behalf of Jost, Gabriele (ARC-TNC)[Supersmith] via flang-dev <flang-dev at lists.llvm.org>
Sent: 18 April 2020 22:56
To: flang-dev at lists.llvm.org <flang-dev at lists.llvm.org>
Subject: [flang-dev] arm + denormals



Hello,

I am running into the problem that flang checks to see if they are denormals. If it thinks they are denormals, it prints out a warning. That slows down the runtime a lot.

Other folks have compiled the same code with ifort using -ftz, probably because they are not concerned about underflows. Is there something similar available for flang, to avoid the checking/printing warnings?

Thanks and greetings,





Gabriele Jost



************************************************************************

Gabriele Jost       NASA Ames Research Center

Email: Gabriele.Jost at nasa.gov      Mail Stop 258-6

Work Phone: (650) 604-0468         Bldg. 258, Rm. 232-4

Scientific & HPC Application           P.O. Box 1

Development/Optimization            Moffett Field, CA 94035-0001

************************************************************************
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/flang-dev/attachments/20200423/a7ddd559/attachment.html>


More information about the flang-dev mailing list