Good morning to everyone!
I am accelerating a fortran code with some OpenACC directives.
I am trying to do some profiling with nvvp. However, when I get to the compute analysis, I get the following error:
The source-assembly viewer could not be shown because source-file mappings are missing from the kernel. You can enable source-file mappings by using the -lineinfo flag when compiling the kernels.
When I try to compile the code with the -lineinfo flag, though, I get the error:
cannot find -lineinfo
I found some topics in the forums that suggest to use the -arch=sm_52 flag. Needless, to say, I get:
nvfortran-Error-Unknown switch: -arch=sm_52
Acutally, I have a Quadro P1000 graphic card, cuda 11.7 installed, and a compute capabily 6.1.
Can you please help me to generate these source-file mappings for my code?
Thank you
Related
I'm cross compiling on Windows 7 and I use this compiler:
gcc-linaro-arm-linux-gnueabihf-4.7-2013.03\bin\arm-linux-gnueabihf-g++.exe
I want to use libcurl for some HTTP requests but when I link using -libcurl I get the infamous message that the linker couldn't find that library. I tried searching on the internet for the correct library for this architecture (ARMHF) but I can't find anything. Or maybe I did find it but I'm doing things wrong? My linking call is long but here it is:
This is the one that works, if I add -libcurl or anything related it breaks...
C:\vde\toolchains\windows\vos2\gcc-linaro-arm-linux-gnueabihf-4.7-2013.03\bin\arm-linux-gnueabihf-g++.exe -o out\RefApp -Wl,-rpath=C:\vde\SDKs\vos2\vos2-sdk-winx86-release-31010800\vos2\usr\lib -Wl,-r
path=C:\vde\SDKs\vos2\vos2-sdk-winx86-release-31010800\vos2\usr\local\lib -Wl,-rpath=C:\vde\SDKs\vos2\vos2-sdk-winx86-release-31010800\vos2\lib -Wl,-rpath=C:\vde\SDKs\vos2\vos2-sdk-winx86-release-3101
0800\vos2\usr\local\lib\svcmgr C:\desarrollo\corvet\out\corvet.a -LC:\desarrollo\RefApp\lib -LC:\vde\SDKs\vos2\vos2-sdk-winx86-release-31010800\vos2\usr\local\lib -LC:\vde\SDKs\vos2\vos2-sdk-winx86-re
lease-31010800\vos2\lib -LC:\vde\SDKs\vos2\vos2-sdk-winx86-release-31010800\vos2\usr\local\lib\svcmgr -LC:\vde\ADKs\adk-full-4.6.5-610\vos2\lib -L -LC:\desarrollo\corvet\out -lrt -llog -lvfiguiprt -l
vfiipc -lTLV_Util -lpthread -lvfiplatforminfo -lsvc_utility -lvfibuzzer -lsvc_powermngt -lsvc_led -lvfisysinfo -lvfisysbar -lsqlite -lexpat -linf -lcom -levt -ltecclient -lmsrclient -lEMV_CT_Link -lEM
V_CT_Client -lEMV_CTLS_Link -lEMV_CTLS_Client -lNFC_Framework -lNFC_Client -lsvc_tms -lUtils -lseccmd-static -lsvc_sound -lccp -lcrypto -lsec -lseccmd-static -lsecins -lvfisvc -lvfisyspm -lvfimac -lvf
isec -lAdkCapX
I also tried downloading some libcurl libraries I found on debian page and that were supposedly for ARMHF, put them in some folder and used -L(routetofolder) but that didn't work either.
Hope you guys can help me out.
Thanks!
Well, libcurl is there in what you get from the vendor (if you look at /usr/lib you will find it there). It's a problem with your linking.
You might want to try looking at how dynamic libraries dependencies are resolved at linking and later at runtime on linux systems. It will also help you understand what may be happening on vos2 platform when you will try using your own dynamic libraries that may have conflicting names and linker can resolve them to something existing on the same device but at a location that you might not have permissions to read, because it belongs to different usrXX or sysXX.
Reading /etc/ld.so.conf from the device will help you understand the locations and search order done by dynamic linker. You can get the file from the device or simply find it in OS.rfsbundle (look in skeletonfs) and set your linking options right. I think it's not normally distributed with sysroot by the vendor but you can manually add it.
I am trying to compile some source codes about UNIX scokets programs, on Linux I have no problems but on macOS I get stuck in front of types definition problems. I don't know how many details I can put here, but I'll try.
The source codes to be compiled are:
errlib.c
errlib.h
server_test.c
sockwrap.c
sockwrap.h
where the main is located in server_test.c.
To compile I use:
gcc -Wall -DTRACE -o server_test server_test.c errlib.c sockwrap.c
There are no problems running this on Linux, but on macOS I get more than 20 error and all of these are about a (perhaphs) missing definition of bool_t. I suppose something not working in /usr/include/rpc/* files located in macOS internal libraries.
So I looked for if <sys/types.h> is included in /usr/include/rpc/xdr.h and it seems to be not missing.
Some of the first lines of the gcc output are:
In file included from server_test.c:16:
/usr/include/rpc/xdr.h:126:3: error: type name requires a specifier or qualifier
bool_t (*x_getlong)(struct __rpc_xdr *, int *);
^
/usr/include/rpc/xdr.h:126:10: error: function cannot return function type 'int (struct __rpc_xdr *, int *)'
bool_t (*x_getlong)(struct __rpc_xdr *, int *);
^
/usr/include/rpc/xdr.h:128:3: error: type name requires a specifier or qualifier
bool_t (*x_putlong)(struct __rpc_xdr *, const int *);
^
/usr/include/rpc/xdr.h:128:10: error: function cannot return function type 'int (struct __rpc_xdr *, const int *)'
bool_t (*x_putlong)(struct __rpc_xdr *, const int *);
^
/usr/include/rpc/xdr.h:128:3: error: duplicate member 'bool_t'
bool_t (*x_putlong)(struct __rpc_xdr *, const int *);
^
/usr/include/rpc/xdr.h:126:3: note: previous declaration is here
bool_t (*x_getlong)(struct __rpc_xdr *, int *);
^
/usr/include/rpc/xdr.h:136:3: error: type name requires a specifier or qualifier
bool_t (*x_getbytes)(struct __rpc_xdr *, char *, unsigned int);
...
and many others lines that are very similar.
For those who want to deepen, the sources are available here.
What could be the problem?
The bad answer
Issue solved by simply including <rpc/types.h> in the server_test.c source as follows:
#include <rpc/types.h>
The better one
This is a solution above is valid only for this server_test.c implementation.
If you want fix this issue "globally", you can follow follown steps:
disable SIP (how to? here)
get your macOS's xdr.h (from here /usr/include/rpc/xdr.h)
copy it elsewhere two times: one copy is for backup and modify the other one simply adding the required include (#include <rpc/types.h> and see the picture below)
overwrite your modified xdr.h onto the original one in /usr/include/rpc/
re-enable SIP
Here is how I did it:
Im stuck on compiling the tsr.asm code provided in the book rootkit arsenal.
I installed open watcom on a XP maschine and the first asm listing was compiled well.
When compiling, it throws the error: "multiple starting address found" (nothing found on google). Can anyone confirm that this code is compilable, and how?
Im thankful for any suggestions.
When you're writing some code, there is a particular address where the execution is to begin (the main function in C for example), but in your code there are more than one starting address, and it crashes when compiling. But without seeing the code I can't tell you more.
Sorry for the late answer, but I was searching for an answer to this and just figured it out - hopefully it'll help someone else Googling around for an answer.
Since you're using OpenWatcom (I'm using version 1.9), I'll assume that you have tsr.asm in its own OpenWatcom 16-bit DOS COM project. In the IDE, go to Targets -> Target Options -> Linker Switches. In the window that appears, select "2. Import, Export and Library Switches" from the drop-down at the top and remove the cstart_t entry under "Library files(,): [libf]".
Recompile, and your TSR COM file should be generated.
I am running a Eclipse for C/C++ using cygwin GCC. I need to get the library path of "libws2_32.a", which is in C:\cygwin\lib\w32api for me in windows. However I do not understand how this is translated into a cygwin-path.
I've tried stuff like:
/cygdrive/c/cygwin/lib/w32api
/lib/w32api
/usr/lib/w32api
Any ideas?
EDIT: update
Here's the make Eclipse log used for building, if it helps:
make all
Building target: Filesharing_core.dll
Invoking: Cygwin C Linker
gcc -L/cygdrive/c/cygwin/lib/w32api -shared -o"Filesharing_core.dll" ./src/test.o -llibws2_32.a
/usr/lib/gcc/i686-pc-cygwin/3.4.4/../../../../i686-pc-cygwin/bin/ld: cannot find -llibws2_32.a
collect2: ld returned 1 exit status
make: *** [Filesharing_core.dll] Error 1
It's not a problem with the path; it's a problem with the name of the library (the -l option).
Where you have entered libws2_32.a, instead enter ws2_32.
See the GCC manual for how the -l option works.
Generally, when asking for help with programming, don't describe, but show. Don't describe what you're doing; don't describe what you think the error is. Instead, show exactly what you're doing, show the exact code (or a cut down version that exhibits the same problem), and show the exact error message.
This way others don't have to guess as to what's wrong, and you are more likely to receive an answer right away that solves your problem.
You can use the find utility to find the file, although it may take a while.
find / -name libws2_32.a -print
Edit: Based on your edit to the question, simply move -llibws2_32.a to before ./src/test.o.
The folder C:\Opt\Cygwin\lib\w32api or whatever it is in your system is simply /lib/w32api from within Cygwin. The library you're looking for is /lib/w32api/libws2_32.a.
But I'm not sure what you're trying to achieve. I guess your Eclipse in running in Windows, and not from a Cygwin'ized version of Java? Why are you using Cygwin GCC? Do you want to compile for Cygwin? If not, MinGW or TDM-GCC would be the tool to pick.
If you do want to compile for the Cygwin environment using Eclipse (a combination I haven't tried), then user experience like the following might be helpful:
http://www.benjaminarai.com/benjamin_arai/index.php?display=/eclipsecygwingcc.php
(I simply googled: eclipse cygwin gcc)
I am working on Linux environment. I have two 'C' source packages train and test_train.
train package when compiled generates libtrain.so
test_train links to libtrain.so and generates executable train-test
Now I want to generate a call graph using gprof which shows calling sequence of functions in main program as well as those inside libtrain.so
I am compiling and linking both packages with -pg option and debugging level is o0.
After I do ./train-test , gmon.out is generated. Then I do:
$ gprof -q ./train-test gmon.out
Here, output shows call graph of functions in train-test but not in libtrain.so
What could be the problem ?
gprof won't work, you need to use sprof instead. I found these links helpful:
How to use sprof?
http://greg-n-blog.blogspot.com/2010/01/profiling-shared-library-on-linux-using.html
Summary from the 2nd link:
Compile your shared library (libmylib.so) in debug (-g) mode. No -pg.
export LD_PROFILE_OUTPUT=`pwd`
export LD_PROFILE=libmylib.so
rm -f $LD_PROFILE.profile
execute your program that loads libmylib.so
sprof PATH-TO-LIB/$LD_PROFILE $LD_PROFILE.profile -p >log
See the log.
I found that in step 2, it needs to be an existing directory -- otherwise you get a helpful warning. And in step 3, you might need to specify the library as libmylib.so.X (maybe even .X.Y, not sure) -- otherwise you get no warning whatsoever.
I'm loading my library from Python and didn't have any luck with sprof. Instead, I used oprofile, which was in the Fedora repositories, at least:
operf --callgraph /path/to/mybinary
Wait for your application to finish or do Ctl-c to stop profiling. Now let's generate a profile summary:
opreport --callgraph --symbols
See the documentation to interpret it. It's kind of a mess. In the generated report, each symbol is listed in a block of its own. The block's main symbol is the one that's not indented. The items above it are functions that call that function, and the ones below it are the things that get called by it. The percentages in the below section are the relative amount of time it spent in those callees.
If you're not on Linux (like me on Solaris) you simply out of luck as there is no sprof there.
If you have the sources of your library you can solve your problem by linking a static library and making your profiling binary with that one instead.
Another way I manage to trace calls to shared libraries, is by using truss. With the option -u [!]lib,...:[:][!]func, ... one can get a good picture of the call history of a run. It's not completely the same as profiling but can be very usefull in some scenarios.