VM builders:
Recent updates in Squeak trunk expose an issue with the FloatMathPlugin. This plugin should be built with compiler optimization turned off, otherwise the plugin can crash the VM.
The Mantis report is at http://bugs.squeak.org/view.php?id=7592 and the discussion on squeak-dev is at http://lists.squeakfoundation.org/pipermail/squeak-dev/2010-December/156132....
Although this has not been verified on all platforms, the working assumption should be that optimization must be turned off for FloatMathPlugin when using gcc.
Dave
Yep.
On 26 December 2010 20:23, David T. Lewis lewis@mail.msen.com wrote:
VM builders:
Recent updates in Squeak trunk expose an issue with the FloatMathPlugin. This plugin should be built with compiler optimization turned off, otherwise the plugin can crash the VM.
The Mantis report is at http://bugs.squeak.org/view.php?id=7592 and the discussion on squeak-dev is at http://lists.squeakfoundation.org/pipermail/squeak-dev/2010-December/156132....
Although this has not been verified on all platforms, the working assumption should be that optimization must be turned off for FloatMathPlugin when using gcc.
Yep.. i noticed that when extracting the knowledge from makefiles and putting it into ST code:
configureFloatMathPlugin: maker "extra rules for FloatMathPlugin"
maker addCrossSources: #( 'acos.c' 'acosh.c' 'asin.c' 'asinh.c' 'atan.c' 'atan2.c' 'atanh.c' 'copysign.c' 'cos.c' 'cosh.c' 'exp.c' 'expm1.c' 'finite.c' 'fmod.c' 'hypot.c' 'isnan.c' 'k_cos.c' 'k_rem_pio2.c' 'k_sin.c' 'k_tan.c' 'ldexp.c' 'log.c' 'log10.c' 'log1p.c' 'modf.c' 'pow.c' 'rem_pio2.c' 'rint.c' 'scalb.c' 'scalbn.c' 'sin.c' 'sinh.c' 'sqrt.c' 'tan.c' 'tanh.c' ).
"On windoze, lib should be built without optimizations? On *nix & mac with -DNO_ISNAN " maker puts:' if (${WIN32}) add_definitions(-Werror-implicit-function-declaration -O0) else(${WIN32}) add_definitions(-DNO_ISNAN) endif (${WIN32})'.
Note, that on unix and mac it defines -DNO_ISNAN while on win32, it compiles with -O0 flag, which turns off all optimizations.
It would be good to know why optimization(s) causing crash, and how we could fix that in code, because i'm not in favor having VM which stability depends on compiler peculiarities.
Dave
On 12/26/2010 8:40 PM, Igor Stasenko wrote:
Note, that on unix and mac it defines -DNO_ISNAN while on win32, it compiles with -O0 flag, which turns off all optimizations.
It would be good to know why optimization(s) causing crash, and how we could fix that in code, because i'm not in favor having VM which stability depends on compiler peculiarities.
Because fdlibm aliases integer pointers with floating point values. Compilers generally don't assume that, i.e.,
double foo = 1.0; #ifdef BIGENDIAN unsigned int *highWord = (unsigned int*)&foo; #else unsigned int *highWord = ((unsigned int*)&foo)++; #endif *highWord |= 0x8000000;
might change the sign of foo and that it would invalidate a prior value in an fpu register. But fdlibm does this kind of stuff *all the time*.
Cheers, - andreas
On 26 December 2010 21:57, Andreas Raab andreas.raab@gmx.de wrote:
On 12/26/2010 8:40 PM, Igor Stasenko wrote:
Note, that on unix and mac it defines -DNO_ISNAN while on win32, it compiles with -O0 flag, which turns off all optimizations.
It would be good to know why optimization(s) causing crash, and how we could fix that in code, because i'm not in favor having VM which stability depends on compiler peculiarities.
Because fdlibm aliases integer pointers with floating point values. Compilers generally don't assume that, i.e.,
double foo = 1.0; #ifdef BIGENDIAN unsigned int *highWord = (unsigned int*)&foo; #else unsigned int *highWord = ((unsigned int*)&foo)++; #endif *highWord |= 0x8000000;
might change the sign of foo and that it would invalidate a prior value in an fpu register. But fdlibm does this kind of stuff *all the time*.
omg..
low level C code which pretending to know how to do it 'right' + C compiler , who also knowing how to do it 'right' ===>> mess in result :)
Cheers, - andreas
btw, how old that fdlibm portions there? maybe updating it will be a simple solution to that issue?
Oh, and please, can you shed some light on the history, why for floating-point math in Squeak, VM using non-standard math lib, which comes with any C compiler, but instead, a portions of separate 3rd party library?
On 12/26/2010 10:24 PM, Igor Stasenko wrote:
btw, how old that fdlibm portions there? maybe updating it will be a simple solution to that issue?
I've never found any newer release than here:
Oh, and please, can you shed some light on the history, why for floating-point math in Squeak, VM using non-standard math lib, which comes with any C compiler, but instead, a portions of separate 3rd party library?
You would be surprised to find out just how much the results of the C compiler libraries differ when it comes to edge cases. I've run the experiment in the past:
Using Python 2.4:
import math math.cos(1.0e32)
WinXP: -0.39929634612021897 LinuxX86: -0.49093671143542561
In short, the results *dramatically* vary depending on processor family, processor version, C compiler libraries, and so on. None of the libraries produce cross-platform bit-identical results. fdlibm does.
Cheers, - Andreas
On 26 December 2010 22:52, Andreas Raab andreas.raab@gmx.de wrote:
On 12/26/2010 10:24 PM, Igor Stasenko wrote:
btw, how old that fdlibm portions there? maybe updating it will be a simple solution to that issue?
I've never found any newer release than here:
oh , yeah.. http://www.netlib.org/fdlibm/readme actually says about that:
NOT FIXED YET
3. Compiler failure on non-standard code Statements like *(1+(int*)&t1) = 0; are not standard C and cause some optimizing compilers (e.g. GCC) to generate bad code under optimization. These cases are to be addressed in the next release.
Oh, and please, can you shed some light on the history, why for floating-point math in Squeak, VM using non-standard math lib, which comes with any C compiler, but instead, a portions of separate 3rd party library?
You would be surprised to find out just how much the results of the C compiler libraries differ when it comes to edge cases. I've run the experiment in the past:
Using Python 2.4:
import math math.cos(1.0e32)
WinXP: -0.39929634612021897 LinuxX86: -0.49093671143542561
In short, the results *dramatically* vary depending on processor family, processor version, C compiler libraries, and so on. None of the libraries produce cross-platform bit-identical results. fdlibm does.
i see. And it looks like this library used by java vm. It is very strange that Oracle having no interest in maintaining it anymore:
http://mailman.oakapple.net/pipermail/numeric-interest/2010-September/002054...
Cheers, - Andreas
Oh, and please, can you shed some light on the history, why for floating-point math in Squeak, VM using non-standard math lib, which comes with any C compiler, but instead, a portions of separate 3rd party library?
You would be surprised to find out just how much the results of the C compiler libraries differ when it comes to edge cases. I've run the experiment in the past:
Using Python 2.4:
import math math.cos(1.0e32)
WinXP: -0.39929634612021897 LinuxX86: -0.49093671143542561
Thanks for the info. Scary. It is amazing how we can forget this ugly things in the comfort of our images. I remember a talk of dave thomas (VA) and he said that they burned tons of smart people on such issues.
In short, the results *dramatically* vary depending on processor family, processor version, C compiler libraries, and so on. None of the libraries produce cross-platform bit-identical results. fdlibm does.
i see. And it looks like this library used by java vm. It is very strange that Oracle having no interest in maintaining it anymore:
http://mailman.oakapple.net/pipermail/numeric-interest/2010-September/002054...
Strange. May be not for free.
Stef
2010/12/26 Andreas Raab andreas.raab@gmx.de:
On 12/26/2010 10:24 PM, Igor Stasenko wrote:
btw, how old that fdlibm portions there? maybe updating it will be a simple solution to that issue?
I've never found any newer release than here:
Oh, and please, can you shed some light on the history, why for floating-point math in Squeak, VM using non-standard math lib, which comes with any C compiler, but instead, a portions of separate 3rd party library?
You would be surprised to find out just how much the results of the C compiler libraries differ when it comes to edge cases. I've run the experiment in the past:
Using Python 2.4:
import math math.cos(1.0e32)
WinXP: -0.39929634612021897 LinuxX86: -0.49093671143542561
Yes, I like this one.
Considering that 1.0e32 is exactly 100000000000000005366162204393472, my own approximation would be: (1.0e32 asArbitraryPrecisionFloatNumBits: 53) cos asFloat storeString -> '0.8108648261576407'
which is in agreement with fdlibm because enough decimals of pi are considered for computing the modulo 1.0e32 cos storeString -> '0.8108648261576407'
But considering that 1.0e32 ulp -> 1.8014398509482e16, the least rounding error leads to an uncertainty interval [-1,1] on this result. So this kind of computation has no meaning at all, bit identical or not, and should better be forbidden. The IEEE 754 conceivers could as well have stated that the right answer was NaN (meaning undefined here).
Nicolas
In short, the results *dramatically* vary depending on processor family, processor version, C compiler libraries, and so on. None of the libraries produce cross-platform bit-identical results. fdlibm does.
Cheers, - Andreas
On 12/26/2010 10:19 PM, Igor Stasenko wrote:
low level C code which pretending to know how to do it 'right' + C compiler , who also knowing how to do it 'right' ===>> mess in result :)
Well, I'm sorry but the C code does bit manipulation according to IEEE 754 so it's hard to blame that code for overly aggressive gcc optimizations. FWIW, the FloatMathPlugin must also be compiled with -mno-fused-madd as this would produce another source of differences depending on whether a compiler chooses to use fmadd or not.
Cheers, - Andreas
On 26 December 2010 23:32, Andreas Raab andreas.raab@gmx.de wrote:
On 12/26/2010 10:19 PM, Igor Stasenko wrote:
low level C code which pretending to know how to do it 'right' + C compiler , who also knowing how to do it 'right' ===>> mess in result :)
Well, I'm sorry but the C code does bit manipulation according to IEEE 754 so it's hard to blame that code for overly aggressive gcc optimizations. FWIW, the FloatMathPlugin must also be compiled with -mno-fused-madd as this would produce another source of differences depending on whether a compiler chooses to use fmadd or not.
ok. added :)
Btw, i found another library: http://www.jhauser.us/arithmetic/SoftFloat.html
looks like this one don't have problems with gcc -O2, because at "SoftFloat speed" it says that it is compiled with this option.
Cheers, - Andreas
On 27 December 2010 00:17, Igor Stasenko siguctua@gmail.com wrote:
On 26 December 2010 23:32, Andreas Raab andreas.raab@gmx.de wrote:
On 12/26/2010 10:19 PM, Igor Stasenko wrote:
low level C code which pretending to know how to do it 'right' + C compiler , who also knowing how to do it 'right' ===>> mess in result :)
Well, I'm sorry but the C code does bit manipulation according to IEEE 754 so it's hard to blame that code for overly aggressive gcc optimizations. FWIW, the FloatMathPlugin must also be compiled with -mno-fused-madd as this would produce another source of differences depending on whether a compiler chooses to use fmadd or not.
ok. added :)
Btw, i found another library: http://www.jhauser.us/arithmetic/SoftFloat.html
looks like this one don't have problems with gcc -O2, because at "SoftFloat speed" it says that it is compiled with this option.
oh , forget it.. it only contains basic operations like add , mul, pow etc but not sin / cos / tan
Cheers, - Andreas
-- Best regards, Igor Stasenko AKA sig.
vm-dev@lists.squeakfoundation.org