I am trying to compile source of milcv7.7.8/ks_spectrum on a cluster with MPICC using version 11.1 20090511 when make the source in milcv7.7.8/ks_spectrum by 'make ks_spectrum_hisq' command then at last the error comes out as
com_mpi.o: In function initialize_machine':
../generic/com_mpi.c:(.text+0xb447): undefined reference to_mm_idivrem_epi32'
I know now that this function _mm_idivrem_epi32 is a part of ia32intrin.h file in intel compiler.
When I use the latest intel mpiicc on new cluster with 14.0.0 20130728 version of intel compiler then code compiles successfully.
So is there any way to tell linker to include function _mm_idivrem_epi32 location...
_mm_idivrem_epi32() is not a function but rather a compiler intrinsic. When properly handled, it is replaced with a call to __svml_idivrem4() from the Intel's Short Vector Math Library libsvml.
You are most likely being hit by a bug in ICC's auto-vectoriser. Try compiling the same source file with -no-vec and see it this has any effect. Or better use the newest ICC version that you have at your disposal.
Related
I have a setup like this:
GDB from "GNU Arm Embedded Toolchain" 10.3-2021.10
GDB server from "Segger JLink" 7.54d
JLink Ultra+ connected to my PC and my embedded device
Arm Compiler 6.15
I'm having problems stepping into a certain function from a C module (let's call it "F1"). When trying, I get the error message
Single stepping until exit from function "F1", which has no line number information.
If I use Segger Ozone, with the same .elf file, stepping into "F1" works fine.
I've tried to narrow down the problem and have the following observations:
A single line of code from the C module holding "F1" makes the difference. If I remove this line, it works. This line is a simple incrementation (++) of a static uint32_t variable and it is in a separate function (i.e. not "F1").
If I don't link with "--inline" option, it stops working - even with the "fix" in (1)
All source files (a mix of C and C++ files) are compiled with -g option.
I may try to reproduce it in a much smaller context which I could share here but until then, I'm hoping for some hints.
Anything is appreciated.
[Update 2021-11-10] Tried with older/newer versions of "GNU Arm Embedded Toolchain" as well as "Segger JLink". Same problem.
[Update 2021-11-10] Compiler/linker command used:
armclang -g --target=arm-arm-none-eabi -mcpu=cortex-m33 -mfloat-abi=soft -MMD -Werror -D__STDC_LIMIT_MACROS -I<my_include_paths>
armlink --inline --info=sizes --info=veneers --info=unused --info=totals --map --symbols --scatter=<my_scatter_file> --list=list.txt
I have a problem when building chromium for ARM platform. Here are some details about my host server:
Linux version 4.2.0-42-generic (buildd#lgw01-55) (gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.3) )
And I use Chromium version 53.0.2785.143. I tried to use gn to build the chromium, and here is my arguments in args.gn file:
target_cpu = "arm"
arm_tune = "generic-armv7-a"
arm_float_abi = "softfp"
Basically, I used these specific arguments above because of my ARM platform. And the gn command ran without errors. However, when building project with ninja, the following errors popped out:
ninja: Entering directory `out/Default_arm64'
[1/1] Regenerating ninja files
[296/46119] LINK ./minidump-2-core
FAILED: minidump-2-core
../../third_party/llvm-build/Release+Asserts/bin/clang++ -Wl,--fatal-warnings -fPIC -Wl,-z,noexecstack -Wl,-z,now -Wl,-z,relro -Wl,-z,defs -fuse-ld=gold -B../../third_party/binutils/Linux_x64/Release/bin -Wl,--icf=all -pthread --target=arm-linux-gnueabihf --sysroot=../../build/linux/debian_wheezy_arm-sysroot -L/home/miaozixiong/workspace/chromium/src/build/linux/debian_wheezy_arm-sysroot/lib/arm-linux-gnueabihf
-Wl,-rpath-link=/home/miaozixiong/workspace/chromium/src/build/linux/debian_wheezy_arm-sysroot/lib/arm-linux-gnueabihf
-L/home/miaozixiong/workspace/chromium/src/build/linux/debian_wheezy_arm-sysroot/usr/lib/arm-linux-gnueabihf
-Wl,-rpath-link=/home/miaozixiong/workspace/chromium/src/build/linux/debian_wheezy_arm-sysroot/usr/lib/arm-linux-gnueabihf
-Wl,-rpath-link=../Default_arm64 -Wl,--disable-new-dtags -o "./minidump-2-core" -Wl,--start-group #"./minidump-2-core.rsp"
-Wl,--end-group -ldl -lrt ld.gold: error: obj/breakpad/minidump-2-core/minidump-2-core.o uses VFP register
arguments, output does not
...
I am new to chromium and have no clue about what do those errors mean. So anybody knows how to work around? You are appreciated.
Note: I need my arm_float_abi attribute to be "softfp" according to my ARM platform. So please note I cannot change it to "hard". Also, when set float abi = "hard", there is no building errors.
ld.gold: error: obj/breakpad/minidump-2-core/minidump-2-core.o uses VFP register arguments, output does not
This a linking error to indicate that minidump-2-core cannot be linked, due to a mismatch in the floating point ABI: the object minidump-2-core.o is compiled for hard floats (the generated code takes advantage of the ARM VFP unit - "uses VFP register arguments"), but the target executable is requested to use soft floats (in which floating point support is emulated, rather than using specialized FP hardware instructions).
According to this bug report, Chromium should build fine with soft float.
My best guess is, try replacing softfp by just soft: arm_float_abi = "soft".
According to gcc documentation, softfp maintains the soft ABI but still 'allows the generation of code using hardware floating-point instructions', which could lead to the seen error.
If that won't work, you might want to check this tutorial on cross building Chromium for ARM:
https://unix.stackexchange.com/questions/176794/how-do-i-cross-compile-chromium-for-arm
I posted this question and finally solved it. I used my local tool chain on ARM platform and compiled it successfully with g++.
I created a C program that uses libmcrypt. I compiled it for many times in a CentOS machine, but now, trying to do the same thing in Ubuntu, I get this:
/tmp/ccM2ugbq.o: In function `encrypt':
util.c:(.text+0xa1): undefined reference to `mcrypt_module_open'
util.c:(.text+0xc6): undefined reference to `mcrypt_enc_get_iv_size'
util.c:(.text+0x134): undefined reference to `mcrypt_generic_init'
util.c:(.text+0x1dc): undefined reference to `mcrypt_generic'
util.c:(.text+0x1fa): undefined reference to `mcrypt_module_close'
/tmp/ccM2ugbq.o: In function `decrypt':
util.c:(.text+0x288): undefined reference to `mcrypt_module_open'
util.c:(.text+0x296): undefined reference to `mcrypt_enc_get_iv_size'
util.c:(.text+0x304): undefined reference to `mcrypt_generic_init'
util.c:(.text+0x375): undefined reference to `mdecrypt_generic'
util.c:(.text+0x380): undefined reference to `mcrypt_module_close'
, although I have installed libmcrypt and have spacified the flags properly when compile (-I/usr/include -L/usr/lib -lmcrypt)
Can anyone guide me on how to get the compilation work properly with libmcrypt?
EDIT: man pages says: Compile as "cc prog.c -lmcrypt", or "cc prog.c -lmcrypt -lltdl" depending on your installation.
I also installed libltdl-dev, but it does not work.
EDIT 1:
System information:
version (running uname -r):
3.5.0-25-generic
release (running lsb_release -a):
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 12.10
Release: 12.10
Codename: quantal
EDIT 2:
Because #teppic requested me to show the compile command, I solved my issue, but I need some explanations. Bellow I'll detail the topic:
the command that I used and failed:
cc -Wall -I../path/to/include -lmcrypt ./a_file.c ./some/other/file.c ./some/other/files.c ./main/file.c -I/some/include/path -lother_lib -o ./bin/out
After that, I only put the -lmcrypt at the end of command and deleted it from the first place, and it worked.
In my initial question, I told you that I did compile it before in CentOS and everything worked fine. That's true, even if in CentOS the compile command were the former one (probably some files differ in order or some libraries).
Can anybody explain me how the flags order influences the compilation success or failure?
In general you should put the library dependency after the source file that references it in the command line. In older versions of gcc, this wasn't always the case.
1) Try installing libmrypt-dev
2) Locate the path of libmcrypt using "locate" command
3) If not in standard library path, try adding the path to LD_LIBRARY_PATH environment variable
4) If nothing helps try installing again using "yum install mcrypt-devel"
I was using Fortran g77 and experienced this problem:
c this program calculates runoff and sediment
1 2
Unrecognized statement name at (1) and invalid form for assignment or statement-function definition at (2)
Also, the compiler can recognized only .for file extension, not .f.
Does anyone know, where is the problem? I downloaded it from http://www.cse.yorku.ca/~roumani/fortran/ftn.htm.
The compiler is not recognizing that statement as a comment. As a comment it should ignore the line but it is trying parse it. Are you sure that the "C" is in the first column?
Why are you using g77? It hasn't been supported for years. gfortran is the current GNU Fortran compiler. It can compile FORTRAN 77, Fortran 90, 95 and portions of 2003 and 2008.
EDIT: Perhaps its wants an upper-case "C".
The page you have linked to states that the f2exe wrapper passes -ffree-form to the compiler:
Compilation Command
The above f2exe command is just a batch file that invokes g77, the "real" compilation command. The command:
g77 -ffree-form prog.for -oprog.exe
directs the compiler to compile the file prog.for and stores the output in the file prog.exe. The -ffree-form switch indicates free-form style (remove it if you are using the old style).
In free-form Fortran the only allowed comment format is that of a line starting with !. As a matter of fact, this is also written on the same page directly under the above text:
Comments
In free-form style, use ! for both full-line and in-line comments. In the old style, use a "C" in column-1.
If you are not using the provided f2exe wrapper, don't pass -ffree-form option when compiling fixed-form FORTRAN 77 code.
I'll assume you want to stick with this compiler.
As noted above, the problems you have come from using the F2EXE batch file, which is not very useful: first it automatically adds ".for" to the file name, so you can't compile ".f" files, and it assumes free-form syntax, which is unusual when programming in Fortran 77 (and if you want Fortran 90, find another compiler, other answers give you links).
Now, suppose you have written a program myprogram.f, and you are in a Windows command line, in the same directory where the program resides (use "cd C:\mydirectory" for example, to change)
You will compile with
g77 myprogram.f
If you use SLATEC, you use
g77 myprogram.f -lslatec
If you want to specify a name for your .exe file (default is a.exe), you write
g77 myprogram.f -o myprogram.exe
There are other useful options
g77 -O2 myprogram.f to optimize (within g77 2.95 limitations)
g77 -Wall myprogram.f to enable all compiler warnings, very useful
to find errors in your code
g77 -c myprogram.f to only compile (you get a .o file), this is
useful to compile functions and subroutines, to
later build a static library (.a file), like
libslatec.a which is given with the compiler
And to build a library, using ar.exe:
ar cru mylib.a myfunc1.o myfnuc2.o ...
Then you can use is with
g77 myprogram.f mylib.a
G77 runs in command line under Windows. You write programs in a text editor.
Notepad++ is fairly good and its free. See http://notepad-plus-plus.org/
If you have problems with compilation, maybe it comes from environment variables, so here are some precisions. You have to tell Windows where to find the G77 compiler (g77.exe).
You can follow instructions on the site where you downloaded it to change Windows' environment variables PATH and LIBRARY_PATH. It needs you install the compiler in the C:\F directory : that is, you will have C:\F\G77\bin, etc.
Slight modification to the instructions on that page :
You should set PATH to C:\F\G77\bin
And LIBRARY_PATH to C:\F\G77\lib;C:\F\SLATEC\lib
This modification to LIBRARY_PATH allows you to compile with SLATEC simply with "-lslatec" as above.
A note about the compiler. It's G77, also know as GNU Fortran 77. An old compiler, integrated with the well known GCC suite until GCC 3.4.6 (we are at GCC 4.7.2 now). And the compiler you downloaded is for version GCC 2.95.
It's a good Fortran 77 compiler, but it's not very well optimized, and of course, you don't get any support for new processor features such as Intel SSE.
Modern Fortran compilers can still understand most if not all of Fortran 77, plus all the newer features of Fortran 90 and newer standards, which are extremely useful.
It may also be interesting to know there is another place to download the same compiler (eccept there is no SLATEC), just in case the page gets destroyed :
http://www.mbr-pwrc.usgs.gov/software/g77.html
so i compiled the GSL from source with Cygwin. Everything went fine, i didnt get any errors. But when i try to compile some simple code, i get the following errors:
..\..\..\Programme\CodeBlocks\MinGW\lib\libgsl.a(error.o) In function `gsl_error':
\usr\include\gsl\err\error.c|43|undefined reference to `__getreent'
\usr\include\gsl\err\error.c|44|undefined reference to `__getreent'
\usr\include\gsl\err\error.c|45|undefined reference to `__getreent'
..\..\..\Programme\CodeBlocks\MinGW\lib\libgsl.a(stream.o) In function `gsl_stream_printf':
\usr\include\gsl\err\stream.c|37|undefined reference to `__getreent'
..\..\..\Programme\CodeBlocks\MinGW\lib\libgsl.a(stream.o) In function `gsl_set_stream':
\usr\include\gsl\err\stream.c|61|undefined reference to `__getreent'
I linked the libgsl.a and libgscblas.a libraries, and there are no complains about any other functions.
For future reference: Dont press cancel instead of Ok when you want to set up a toolchain in code::blocks
The error you're seeing is due to the fact that MinGW doesn't link the Cygwin DLL by default.
MinGW- and Cygwin-compiled libraries don't play that well with each other - while it's possible to get it to work, it's not really a good idea.
Choose one toolchain for your project, which, in this case, means either setting up Code::Blocks so it uses the Cygwin toolchain instead of the bundled MinGW distribution or compiling GSL with MinGW as well.
You may use the following option in Cygwin:
./configure CC=i686-w64-mingw32-gcc
Then you may use GSL in CodeBlocks.