Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <8442171.1181726546305.JavaMail.jira@brutus>
Date: Wed, 13 Jun 2007 02:22:26 -0700 (PDT)
From: "Igor V Chebykin (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Commented: (HARMONY-2873) [drlvm][jit][opt] frem and drem
 instructions return incorrect result when divisor is Float/Double MIN_VALUE
In-Reply-To: <8494986.1167121941078.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HARMONY-2873?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12504174 ] 

Igor V Chebykin commented on HARMONY-2873:
------------------------------------------

The problem is due to improper algorithm in library functions:
  double fmod(double,double)
  float fmodf(float,float)

The suggested fix is:
  1. replace usage fmodf by fmod
  2. replace usage fmod by our own function (classlib already contains those function)


> [drlvm][jit][opt] frem and drem instructions return incorrect result when divisor is Float/Double MIN_VALUE
> -----------------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-2873
>                 URL: https://issues.apache.org/jira/browse/HARMONY-2873
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: Linux and Windows
>            Reporter: Vera Petrashkova
>            Assignee: Alexey Varlamov
>
> The following test demonstrates that frem and drem instructions return incorrect result when divisor
> is correspondently Float or Double MIN_VALUE. The result equals to dividend.
> This issue is reproducible only on Jitrino/OPT.
> On  interpreter and on Jitrino/JET result is equal to 0.0.
> ---------------------test.java---------------
> public class test {
>     public static void main(String[] args) {
>         float f1 = 5.5f;
>         float f2 = Float.MIN_VALUE;
>         float ff = f1 % f2;
>         System.out.println( ff == 0.0f ? 
>                        "Float test passed. "+f1+" % "+f2+" = "+ff : 
>                        "Float test failed. "+f1+" % "+f2+" = "+ff +" instead of 0.0" );
>         double d1 = 3.3d;
>         double d2 = Double.MIN_VALUE;
>         double dd = d1 % d2;
>         System.out.println( dd == 0.0d ? 
>                        "Double test passed. "+d1+" % "+d2+" = "+dd : 
>                        "Double test failed. "+d1+" % "+d2+" = "+dd +" instead of 0.0" );
>     }
> }
> ------------------------
> Output is:
> Float test failed. 5.5 % 1.4E-45 = 5.5 instead of 0.0
> Double test failed. 3.3 % 4.9E-324 = 3.3 instead of 0.0

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.