I'm trying to get started with OpenCL but came across weird behavior of the OpenCL compiler with respect to white-space and can't seem to find any documentation about that.
C-style single-line comments (// foo) immediately cause a meaningless build error: At end of source: error: expected a "}". Multi-line comments (/* bar */) seem to work fine.
Line breaks seem to get stripped without adding whitespace which can cause errors. This example will not compile because of that:
__kernel
void TestKernel() {}
line 1: error: identifier "__kernelvoid" is undefined
This may totally depend on my machine and/or configuration but can somebody confirm that these things should not be this way?
I am using OpenCL via Cloo from .net/C#. The driver is from AMD OpenCL 2.0 AMD-APP (1642.5)
I think I figured it out. I was doing this:
var program = new ComputeProgram(context, File.ReadAllLines(filename));
File.ReadAllLines() returns an array of strings without the line-break characters which is the root of the errors I was getting.
Using File.ReadAllTest() instead fixed all the problems:
var program = new ComputeProgram(context, File.ReadAllText(filename));
But in my opinion some of the blame goes to either Cloo or the OpenCL API for accepting a string array but just concatenating it together..
Related
I have a simple Chapel program to test the NetCDF module:
use NetCDF;
use NetCDF.C_NetCDF;
var f: int = ncopen("ppt2020_08_20.nc", NC_WRITE);
var status: int = nc_close(f);
and when I compile with:
chpl -I/usr/include -L/usr/lib/x86_64-linux-gnu -lnetcdf hello.chpl
it produces a list of errors about SysCTypes:
$CHPL_HOME/modules/packages/NetCDF.chpl:57: error: 'c_int' undeclared (first use this function)
$CHPL_HOME/modules/packages/NetCDF.chpl:77: error: 'c_char' undeclared (first use this function)
...
Would anyone see what my error is? I tried adding use SysCTypes; to my program, but that didn't seem to have an effect.
Sorry for the delayed response and for this bad behavior. This is a bug that's crept into the NetCDF module which seems not to have been caught by Chapel's nightly testing. To work around it, edit $CHPL_HOME/modules/packages/NetCDF.chpl, adding the line:
public use SysCTypes, SysBasic;
within the declaration of the C_NetCDF module (around line 50 in my copy of the sources). If you would consider filing this bug as an issue on the Chapel GitHub issue tracker, that would be great as well, though we'll try to get this fixed in the next release in any case.
With that change, your program almost compiles for me, except that nc_close() takes a c_int argument rather than a Chapel int. You could either lean on Chapel's type inference to cause this to happen:
var f = ncopen("ppt2020_08_20.nc", NC_WRITE);
or explicitly declare f to be of type c_int:
var f: c_int = ncopen("ppt2020_08_20.nc", NC_WRITE);
And then as one final note, I believe you should be able to drop the -lnetcdf from your chpl command-line as using the NetCDF module should cause this requirement to automatically be added.
Thanks for bringing this bug to our attention!
I'm getting this on the console in a QML app:
QFont::setPointSizeF: Point size <= 0 (0.000000), must be greater than 0
The app is not crashing so I can't use the debugger to get a backtrace for the exception. How do I see where the error originates from?
If you know the function the warning occurs in (in this case, QFont::setPointSizeF()), you can put a breakpoint there. Following the stack trace will lead you to the code that calls that function.
If the warning doesn't include the name of the function and you have the source code available, use git grep with part of the warning to get an idea of where it comes from. This approach can be a bit of trial and error, as the code may span more than one line, etc, and so you might have to try different parts of the string.
If the warning doesn't include the name of the function, you don't have the source code available and/or you don't like the previous approach, use the QT_MESSAGE_PATTERN environment variable:
QT_MESSAGE_PATTERN="%{function}: %{message}"
For the full list of variables at your disposal, see the qSetMessagePattern() docs:
%{appname} - QCoreApplication::applicationName()
%{category} - Logging category
%{file} - Path to source file
%{function} - Function
%{line} - Line in source file
%{message} - The actual message
%{pid} - QCoreApplication::applicationPid()
%{threadid} - The system-wide ID of current thread (if it can be obtained)
%{qthreadptr} - A pointer to the current QThread (result of QThread::currentThread())
%{type} - "debug", "warning", "critical" or "fatal"
%{time process} - time of the message, in seconds since the process started (the token "process" is literal)
%{time boot} - the time of the message, in seconds since the system boot if that can be determined (the token "boot" is literal). If the time since boot could not be obtained, the output is indeterminate (see QElapsedTimer::msecsSinceReference()).
%{time [format]} - system time when the message occurred, formatted by passing the format to QDateTime::toString(). If the format is not specified, the format of Qt::ISODate is used.
%{backtrace [depth=N] [separator="..."]} - A backtrace with the number of frames specified by the optional depth parameter (defaults to 5), and separated by the optional separator parameter (defaults to "|"). This expansion is available only on some platforms (currently only platfoms using glibc). Names are only known for exported functions. If you want to see the name of every function in your application, use QMAKE_LFLAGS += -rdynamic. When reading backtraces, take into account that frames might be missing due to inlining or tail call optimization.
On an unrelated note, the %{time [format]} placeholder is quite useful to quickly "profile" code by qDebug()ing before and after it.
I think you can use qInstallMessageHandler (Qt5) or qInstallMsgHandler (Qt4) to specify a callback which will intercept all qDebug() / qInfo() / etc. messages (example code is in the link). Then you can just add a breakpoint in this callback function and get a nice callstack.
Aside from the obvious, searching your code for calls to setPointSize[F], you can try the following depending on your environment (which you didn't disclose):
If you have the debugging symbols of the Qt libs installed and are using a decent debugger, you can set a conditional breakpoint on the first line in QFont::setPointSizeF() with the condition set to pointSize <= 0. Even if conditional breakpoints don't work you should still be able to set one and step through every call until you've found the culprit.
On Linux there's the tool ltrace which displays all calls of a binary into shared libs, and I suppose there's something similar in the M$ VS toolbox. You can grep the output for calls to setPointSize directly, but of course this won't work for calls within the lib itself (which I guess could be the case when it handles the QML internally).
I have legacy code for an embedded 8051 core (in a cypress FX2) that used to compile with other versions of SDCC. However, current SDCC doesn't know the _naked qualifier:
delay.c:27: syntax error: token -> '_naked' ; column 21
as triggered by
static void
udelay1 (void) _naked
{
_asm ; lcall that got us here took 4 bus cycles
ret ; 4 bus cycles
_endasm;
}
and other occurrences.
As _naked practically is supposed to tell the C compiler to "nah, ignore the fact that you're a C compiler and understand that you'd need to save frame context", I don't feel like I should just #define it away.
Is there any solution to this? Should I just go ahead and manually inline the assembler wherever a _naked function is used? I feel like I'd betraying the compiler on a CALL there, and that would change the timing.
_naked was replaced by __naked in newer versions of SDCC. Same applies to asm/__asm, at/__at, interrupt,bit,xdata/__….
So, this turned out to be an exercise in regex replacements.
I'm still having linker/ranlib/mostly ar problems, and CMake ignores what I instruct it to use as compilers, but oh well.
I have a bizarre one here with the serial output when trying to write some code for my Arduino Uno.
I have this proto-code:
MyClass myclass;
void setup()
{
Serial.Begin(9600);
Serial.println("Starting...");
}
void loop()
{
int status = myclass.DoWork();
Serial.println("Status: " + status);
}
class MyClass
{
int DoWork()
{
Serial.println("Doing some work...");
return 1;
}
}
Now when this runs I get the following output:
Starting...
Doing some work...
atus: 1
So the strange part is the "Status: 1" missing the first few characters. Is this because I am using serial in an object improperly or something?
I have noticed when I reference another library that also uses serial like MyClass does that I get other strange output behavior... so I assume that I am doing something wrong.
EDIT: In the end this turned out to actually be a memory issue. A library I was including was quite large and it was consuming the available memory. I found this by adding a few more debugging statements and found the corruption shifted around based on the string lengths and positions. By using the F() function I moved the strings to flash memory (e.g. I now run Serial.println(F("Starting...")); and it has corrected the strange output.
You cannot add strings and integers in C++. It would have been better for you if this failed to compile:
Serial.println("Status: " + status);
Instead the compiler guessed at something. It guessed wrong. Use this:
Serial.print("Status :");
Serial.println(status);
or for complete control of outputting numbers and strings learn to use C string formatting, sprintf()
In the end this turned out to actually be a memory issue. A library I was including was quite large and it was consuming the available memory. I found this by adding a few more debugging statements and found the corruption shifted around based on the string lengths and positions. By using the F() function I moved the strings to flash memory (e.g. I now run Serial.println(F("Starting...")); and it has corrected the strange output.
One more possible explanation.
I was running minicom to monitor, and I usually like that it auto-reconnects after resetting my device.
Well I minimized a terminal running minicom last night, and today I started a new instance that somehow also got connected to the same serial port (if that is even possible).
I think the two instances of minicom were each reading ~50% of the serial characters roughly at random, leaving me with quite a mess of text.
I'm using mingw32-make to compile a qt project that uses opengl, it compiles correctly and everything, but it spits countless warning messages of the form:
c:/qt3/include/qcolor.h:67: warning: inline function `int qGray(int, int,
int)' declared as dllimport: attribute ignored
For this particular instance, the function declaration is:
Q_EXPORT inline int qGray( int r, int g, int b )// convert R,G,B to gray 0..255
{ return (r*11+g*16+b*5)/32; }
My question is, why is it spitting all these warning? how can I silence them without silencing other legitimate warnings (i.e. warnings that are related directly to my code and could be potential problems)?
More importantly, why is mingw ignoring the dll import attribute in the first place?
I think Qt ought to only define Q_EXPORT (Q_DECL_EXPORT in Qt 4) to be the dllexport/import attribute if one of the following macros is defined, so make sure your makefiles or code that includes Qt headers (which eventually will include qglobal.h) aren't defining any of them: WIN32, _WIN32, __WIN32__, WIN64, _WIN64, __WIN64__. Or you can just define Q_EXPORT to be nothing in your compile (or preprocessor) flags, then Qt should skip defining it.