valgrind-users

[Valgrind-users] POW() function behaving different when running in valgrind?

[Wilfried G. <dot...@ci...>]

Hi Everyone,  

Duplicating the discussion from the V8 user group:  
>  
>
>https://groups.google.com/forum/#!topic/v8-users/qRorKvYNq24    On Fri, Dec
>4, 2015 at 3:26 PM, Wilfried Gösgens <doth...@gmailcom> wrote: 
>> Hi everyone 
>> Using the regular d8 shell: 
>> 
>>     # ./d8 
>>     V8 version 4.8.271.5 
>>     d8> Math.pow(0, 0.1); 
>>     0 
>>     d8> 
>> 
>> gives proper results. 
>> But using d8 in valgrind gives: 
>> 
>>     /usr/bin/valgrind.bin --log-file=/tmp/valgrindlog.%p 
>> --show-reachable=yes --leak-check=full ./d8 
>>     Math.pow(0, 0.1); 
>>     V8 version 4.8.271.5 
>>     d8> NaN 
>>     d8> 
>> 
>> Is this a bug in V8 or in Valgrind? 
>
>  
>
>  
>
>I can reproduce what you describe and it doesn't go away when you pass 
>--smc-check=all to Valgrind or --nofast_math to V8. 
> 
>I'm inclined to suspect Valgrind but I'm honestly not sure.  `pow(0, 
>0.1)` from C code or a Perl script doesn't exhibit the same behavior. 
>Those may not be issuing the same instructions, of course.  

  

d8 is the sample shell provided by the V8 project for testing the V8
javascript engine.  

It will be built alongside the V8 engine:
https://developers.google.com/v8/build?hl=en  

Does valgrind replace mathematical functions?  

   

Cheers,  

Willi

Re: [Valgrind-users] POW() function behaving different when running in valgrind?

[Tom H. <to...@co...>]

On 07/12/15 11:40, Wilfried Goesgens wrote:

> Does valgrind replace mathematical functions?

No, but see the section on x86/AMD64 floating point support in the manual:

http://valgrind.org/docs/manual/manual-core.html#manual-core.limits

Tom


-- 
Tom Hughes (to...@co...)
http://compton.nu/

Re: [Valgrind-users] POW() function behaving different when running in valgrind?

[John R. <jr...@bi...>]

> Duplicating the discussion from the V8 user group:
>
>     https://groups.google.com/forum/#!topic/v8-users/qRorKvYNq24
>
>     On Fri, Dec 4, 2015 at 3:26 PM, Wilfried Gösgens <do...@gm...> wrote:
>      > Hi everyone
>      > Using the regular d8 shell:
>      >
>      >     # ./d8
>      >     V8 version 4.8.271.5
>      >     d8> Math.pow(0, 0.1);
>      >     0
>      >     d8>
>      >
>      > gives proper results.
>      > But using d8 in valgrind gives:
>      >
>      >     /usr/bin/valgrind.bin --log-file=/tmp/valgrindlog.%p
>      > --show-reachable=yes --leak-check=full ./d8
>      >     Math.pow(0, 0.1);
>      >     V8 version 4.8.271.5
>      >     d8> NaN
>      >     d8>

> Does valgrind replace mathematical functions?

No, but valgrind forces round-to-nearest 64-bit result
for every floating-point operation: no extra precision
nor exponent range for 80-bit extended opcodes,
and no directed rounding (truncate, ceil, floor).  This is
well documented.  Thus if the user code relies on 80-bit mode
or rounding other than round-to-nearest, then "bad results"
may ensue.  The bug is in V8 not diagnosing the actual environment
before proceeding.  Never run a laboratory experiment
for quantitative results without first calibrating the instruments.