Attempting to solve linear programming problems using GLPLK's GLPSOL we've come upon a snag, namely that in very specific cases, the results between glpsol executables created with different compilers are different.
The situation is that we have a problem with several valid solutions. To put it simply, we have a table where each row (X) can be assigned only one column (Y), and viceversa. As such, all combinations that assign unique column/row pairs are valid.
Example, for a 2x2 table, these are valid:
{(X0,Y0),(X1,Y1)} {(X0,Y1),(X1,Y0)}
Now, the original glpsol binary we used under windows, returned the results in order, something like this:
{(X0,Y0),(X1,Y1)...(Xn,Yn)}
We noticed an issue with the Linux binary, in that it returned the solution in a different order, something like this:
{(X0,Y0),(Xn,Y1),(X1,Y2) ....}
Note that the order is not random, every execution follows the same pattern.
After much investigation I discovered that the issue lies in which compiler was used to create each binary. In our example above, the Windows binary was compiled using Visual C++, while the Linux binary used GCC.
I've verified this by recompiling the Windows binary using GCC, resulting in the same pattern. Compiling with Borland results in a different pattern.
So the question is, mainly, why is this happening?
I'm guessing it might be the result of how each compiler optimizes the binary, but I'm not sure, and my objective is to obtain the same results we had with the original executable (the one compiled with Visual C++) both for Windows and Linux. And I am suspecting cross-compiling with the Visual C++ toolchain won't be an option.
Note: I managed to determine the compiler used by each binary by opening them as text and locating text strings within the executable referencing Visual C++ and GNU GCC respectively.
Thanks!
Versions of the solver built with different compilers can take different paths during the optimization process which can result in the behavior you observe. Things that can affect this are: differences in floating-point semantics (possibly caused by -ffast-math), different implementations of sort (qsort is normally not a stable sort) - this is mentioned by Ben Voigt, different implementations of random number generators in standard libraries.
If both solutions are optimal, I wouldn't be too much concerned about this.
Related
i have the following problem (and cannot really produce a minimal test)--
i am porting a package from C++ via Rcpp to R.
the tests (i am testing if the output matrix is exactly what i
would get if calling c++ directly) under linux and osx are absolutely equal, no difference.
but when testing either via build_win() or via a win 8.1 virtual machine i get different results (but the results between both are consistent, so i have linux/osx vs win results)
i already replaced the one rand() call with the corresponding Rcpp sugar, so this should be no problem (i hope at least).
as calling the tests via "R -d valgrind" also produce no error, i am a bit puzzled how to proceed.
all tests are done with R 3.2.0 (local machines) and latest unstable (via build_win())
so my questions are:
are there any known Rcpp differences when compiling (e.g. the compiler provided by Rtools on windows is too old and therefore numeric computations (using STL, no other library like boost/eigen etc) are expected to be slightly different?
is there a good way to debug the problem? i would need to trace basically the C++ code line by line, i am even not sure how to do that except for heavy std::couts.
thanks.
the truth about the 32bit/64bit problem is indeed written up here
different behaviour or sqrt when compiled with 64 or 32 bits
adding the -ffloat-store option did fix my problem.
never expected that, thought the problem is in the source code.
I am trying to figure out a bug (a serious performance downgrade). Unfortunately, I wasn't able to figure out why by going back many different versions of my code.
I am suspecting it could be some modifications to libraries that I've updated, not to mention in the meanwhile I've updated to GHC 7.6 from 7.4 (and if anybody knows if some laziness behavior has changed I would greatly appreciate it!).
I have an older executable of this code that does not have this bug and thus I wonder if there are any tools to tell me the library versions I was linking to from before? Like if it can figure out the symbols, etc.
GHC creates executables, which are notoriously hard to understand... On my Linux box I can view the assembly code by typing in
objdump -d <executable filename>
but I get back over 100K lines of code from just a simple "Hello, World!" program written in Haskell.
If you happen to have the GHC .hi files, you can get some information about the executable by typing in
ghc --show-iface <hi filename>
This won't give you the assembly code, but you can get some extra information that may prove useful.
As I mentioned in the comment above, on Linux you can use "ldd" to see what C-system libraries you used in the compile, but that is also probably less than useful.
You can try to use a disassembler, but those are generally written to disassemble to C, not anything higher level and certainly not Haskell. That being said, GHC compiles to C as an intermediary (at least it used to; has that changed?), so you might be able to learn something.
Personally I often find view system calls in action much more interesting than viewing pure assembly. On my Linux box, I can view all system calls by running using strace (use Wireshark for the network traffic equivalent):
strace <program executable>
This also will generate a lot of data, so it might only be useful if you know of some specific place where direct real world communication (i.e., changes to a file on the hard disk drive) goes wrong.
In all honesty, you are probably better off just debugging the problem from source, although, depending on the actual problem, some of these techniques may help you pinpoint something.
Most of these tools have Mac and Windows equivalents.
Since much has changed in the last 9 years, and apparently this is still the first result a search engine gives on this question (like for me, again), an updated answer is in order:
First of all, yes, while Haskell does not specify a bytecode format, bytecode is also just a kind of machine code, for a virtual machine. So for the rest of the answer I will treat them as the same thing. The “Core“ as well as the LLVM intermediate language, or even WASM could be considered equivalent too.
Secondly, if your old binary is statically linked, then of course, no matter the format your program is in, no symbols will be available to check out. Because that is what linking does. Even with bytecode, and even with just classic static #include in simple languages. So your old binary will be no good, no matter what. And given the optimisations compilers do, a classic decompiler will very likely never be able to figure out what optimised bits used to be partially what libraries. Especially with stream fusion and such “magic”.
Third, you can do the things you asked with a modern Haskell program. But you need to have your binaries compiled with -dynamic and -rdynamic, So not only the C-calling-convention libraries (e.g. .so), and the Haskell libraries, but also the runtime itself is dynamically loaded. That way you end up with a very small binary, consisting of only your actual code, dynamic linking instructions, and the exact data about what libraries and runtime were used to build it. And since the runtime is compiler-dependent, you will know the compiler too. So it would give you everything you need, but only if you compiled it right. (I recommend using such dynamic linking by default in any case as it saves memory.)
The last factor that one might forget, is that even the exact same compiler version might behave vastly differently, depending on what IT was compiled with. (E.g. if somebody put a backdoor in the very first version of GHC, and all GHCs after that were compiled with that first GHC, and nobody ever checked, then that backdoor could still be in the code today, with no traces in any source or libraries whatsoever. … Or for a less extreme case, that version of GHC your old binary was built with might have been compiled with different architecture options, leading to it putting more optimised instructions into the binaries it compiles for unless told to cross-compile.)
Finally, of course, you can profile even compiled binaries, by profiling their system calls. This will give you clues about which part of the code acted differently and how. (E.g. if you notice that your new binary floods the system with some slow system calls where the old one just used a single fast one. A classic OpenGL example would be using fast display lists versus slow direct calls to draw triangles. Or using a different sorting algorithm, or having switched to a different kind of data structure that fits your work load badly and thrashes a lot of memory.)
I'm re-writing a build that produces a number of things (shared/static libraries, jars, executables, etc). The question came up whether there's a way to verify that the results are functionally equivalent without doing a full top-to-bottom test of the resulting software.
However, that is proving to be more difficult to do than I anticipated.
As an example, I expected that the md5 of two objects produced from the same source (sun studio C++ compiler) and command-line parameters would have the same md5 hash, but that isn't the case. I can build the file, rename it, build again, and they have different hashes.
With that said ... is there a way do a quick check to verify that two files produced from separate build architectures of the same source tree (eg, two shared objects) are functionally equivalent?
edit I am sorry, I neglected to mention this is for a debug build ... when debugging flags aren't used the binaries are identical, but they've been using debugging flags by default for so many years their stuff breaks when you remove the debugging flags (part of the reason I'm re-writing the build is to take that particular 'feature' out of the build so we can get some proper testing going)
Windows DLLs have a link timestamp (TimeDateStamp) as part of PE image.
Looking at linker options, I don't see an option to suppress that. So re-linking a DLL (or an EXE) will always produce a different binary.
You could write a tool to zero out these timestamps (always at a fixed offset from file start), and compare MD5s afterwards. But you'll likely discover lots of other differences as well. In particular, any program that uses __DATE__ or __TIME__ builtins will give you trouble.
We've had to work quite hard to achieve bit-identical rebuilds (using GNU toolchain). It's possible (at least for open-source tools, on Linux), but not easy (as you've discovered).
I forgot about this question; I'm revisiting so I can give the answer I came up with.
objcopy can be used to produce a new binary file in different formats. It's been a few years since I worked on this, so the specifics escape me, but here's what I recall:
objcopy can strip various things out (debug info, symbol information, etc), but even after stripping stuff out I was still seeing different hashes between objects.
In the end I found I could convert it from ELF to other formats. I ended up dumping it to another format (I think I chose SREC) that consistently provided the same MD5 for objects built at different times with identical source/flags.
I'm betting I could have done this a better way with objcopy (or perhaps another binutils tool), but it was good enough to satisfy our concerns.
When the same piece of c++ code is compiled with the same version of visual c++ compiler but at different times and possibly in different computers, does the code reordering performed by the compiler remains same or it may differ. i.e. does the logic behind code optimization by code reordering depend only on the code or it depends on various other parameters?
The context of the question is that I want to create a tool which finds out whether the two dlls are same or different based on their functionalities.
Correct me if I'm wrong in assuming that since you want to compare dlls based on their functionality, you don't care about implementation details. Based on this assumption, it is clear that your tool could only look at the function signatures and classes, structs, etc definitions exposed by the dlls which would always be the same regardless of compiler for the same dll.
I responded to another question about developing for the iPhone in non-Objective-C languages, and I made the assertion that using, say, C# to write for the iPhone would strike an Apple reviewer wrong. I was speaking largely about UI elements differing between the ObjC and C# libraries in question, but a commenter made an interesting point, leading me to this question:
Is it possible to determine the language a program is written in, solely from its binary? If there are such methods, what are they?
Let's assume for the purposes of the question:
That from an interaction standpoint (console behavior, any GUI appearance, etc.) the two are identical.
That performance isn't a reliable indicator of language (no comparing, say, Java to C).
That you don't have an interpreter or something between you and the language - just raw executable binary.
Bonus points if you're language-agnostic as possible.
Short answer: YES
Long answer:
If you look at a binary, you can find the names of the libraries that have been linked in. Opening cmd.exe in TextPad easily finds the following at hex offset 0x270: msvcrt.dll, KERNEL32.dll, NTDLL.DLL, USER32.dll, etc. msvcrt is the Microsoft 'C' runtime support functions. KERNEL32, NTDLL, and USER32.dll are OS specific libraries which tell you either the target platform, or the platform on which it was built, depending on how well the cross-platform development environment segregates the two.
Setting aside those clues, most any c/c++ compiler will have to insert the names of the functions into the binary, there is a list of all functions (or entrypoints) stored in a table. C++ 'mangles' the function names to encode the arguments and their types to support overloaded methods. It is possible to obfuscate the function names but they would still exist. The functions signatures would include the number and types of the arguments which can be used to trace into the system or internal calls used in the program. At offset 0x4190 is "SetThreadUILanguage" which can be searched for to find out a lot about the development environment. I found the entry-point table at offset 0x1ED8A. I could easily see names like printf, exit, and scanf; along with __p__fmode, __p__commode, and __initenv
Any executable for the x86 processor will have a data segment which will contain any static text that was included in the program. Back to cmd.exe (offset 0x42C8) is the text "S.o.f.t.w.a.r.e..P.o.l.i.c.i.e.s..M.i.c.r.o.s.o.f.t..W.i.n.d.o.w.s..S.y.s.t.e.m.". The string takes twice as many characters as is normally necessary because it was stored using double-wide characters, probably for internationalization. Error codes or messages are a prime source here.
At offset B1B0 is "p.u.s.h.d" followed by mkdir, rmdir, chdir, md, rd, and cd; I left out the unprintable characters for readability. Those are all command arguments to cmd.exe.
For other programs, I've sometimes been able to find the path from which a program was compiled.
So, yes, it is possible to determine the source language from the binary.
I'm not a compiler hacker (someday, I hope), but I figure that you may be able to find telltale signs in a binary file that would indicate what compiler generated it and some of the compiler options used, such as the level of optimization specified.
Strictly speaking, however, what you're asking is impossible. It could be that somebody sat down with a pen and paper and worked out the binary codes corresponding to the program that they wanted to write, and then typed that stuff out in a hex editor. Basically, they'd be programming in assembly without the assembler tool. Similarly, you may never be able to tell with certainty whether a native binary was written in straight assembler or in C with inline assembly.
As for virtual machine environments such as JVM and .NET, you should be able to identify the VM by the byte codes in the binary executable, I would expect. However you may not be able to tell what the source language was, such as C# versus Visual Basic, unless there are particular compiler quirks that tip you off.
what about these tools:
PE Detective
PEiD
both are PE Identifiers. ok, they're both for windows but that's what it was when i landed here
I expect you could, if you disassemble the source, or at least you may know the compiler, as not all compilers will use the same code for printf for example, so Objective-C and gnu C should differ here.
You have excluded all byte-code languages so this issue is going to be less common than expected.
First, run what on some binaries and look at the output. CVS (and SVN) identifiers are scattered throughout the binary image. And most of those are from libraries.
Also, there's often a "map" to the various library functions. That's a big hint, also.
When the libraries are linked into the executable, there is often a map that's included in the binary file with names and offsets. It's part of creating "position independent code". You can't simply "hard-link" the various object files together. You need a map and you have to do some lookups when loading the binary into memory.
Finally, the start-up module for C, C++ (and I imagine C#) is unique to that compiler's defaiult set of libraries.
Well, C is initially converted the ASM, so you could write all C code in ASM.
No, the bytecode is language agnostic. Different compilers could even take the same code source and generate different binaries. That's why you don't see general purpose decompilers that will work on binaries.
The command 'strings' could be used to get some hints as to what language was used (for instance, I just ran it on the stripped binary for a C application I wrote and the first entries it finds are the libraries linked by the executable).