If in source CUDA file main header (cuda_runtime.h or cuda.h) is not presented, corresponding HIP main header (hip_runtime.h) should be explicitly included in output hipified file.
[Algorithm]
1. If #pragma once is presented, HIP main header should be placed just after it;
2. Otherwise if any other (not CUDA main) header is presented, HIP main header should be placed just before it;
3. Otherwise HIP main header should be placed in the beginning of output file.
P.S.
There might be one more situation when #ifndef #define ... #endif guard for the entire file is presented (make sense for *.h, *.hpp, *.cuh files). In this case HIP main include should be placed just after such #ifdef, or after #pragma once, if it is also presented. This situation will be handled in a separate change.
// hipified to #include<hip/hip_runtime.h>
#include<cuda.h> // 1st cuda main include (Driver API)
// to eliminate
#include<cuda_runtime.h> // 2nd cuda main include (Runtime API)
HIP has one header hip_runtime.h for both CUDA APIs, thus second cuda main include directive is eliminated entirely.
Function require_program erroneously doesn't report "Can't find program" on any missing program except the first one due to the cached FOUND_PROGRAM value.
Additionally:
+ Do not throw FATAL_ERROR on missing program in order to obtain the whole list of missing programs (if any).
+ Report also found program location.
Much pretty. Very wow
This gives users all the usual power when it comes to manipulating
clang diagnostics. People can pass -Werror can have hipify fail if
it doesn't completely translate a file, for example. Much nicer
than reinventing the wheel.
Most of what hipify does is really just replacing CUDA idenitifers
with HIP ones. CUDA function calls, preprocessor macro calls,
enum references, types, etc.
This is problematic: calls/types/enum-refs require name resolution
for the AST matcher to work. This fails in the presence of code
deleted by the preprocessor, and in two-pass template compilation.
Instead, we can simply hook the lexer and have it rewrite the
identifiers for us.
This approach means identifier transformations will work correctly
regardless of where they appear (and we get to delete lots of code)
- Fixes#260
- Helps a bit with #207 - it will still fail to translate kernel
calls in preprocessor-ignored code, but everything except kerel
launches should translate correctly now, even in
preprocessor-deleted code.
The original implementation had the statistics system woken very
tightly into things like PPCallbacks, with counters duplicated
in two places, and all the output code duplicated. This made it
very difficult to alter the structure of the program without
breaking the statistics system.
Since the planned approach for solving the remaining preprocessor
bugs needs the introduction of a custom FrontendAction, and such
a restructure was incompatible with the way the statistics system
was set up, this rewrite was required.
'tis rather simpler now, mind you :D
This commit also fixes an issue where some stats were counted
twice, and allows `-print-stats` to operate independently of
`-stat-output`, allowing you to print stats to a file without
printing them to a terminal (or vice-versa).
operator[] default-constructs the map value if no value exists
for that key. Default-construction of int yields a zero. So all
the manual faffing around is just unnecessary.
See [the documentation](https://cmake.org/cmake/help/v3.0/command/find_package.html)
for exactly how the search procedure works. If you want to use an
LLVM from a specific location, use CMAKE_PREFIX_PATH as normal.
No longer do we have a nonstandard HIPIFY_CLANG_LLVM_DIR variable
for people to learn about.
Since there's now an option for toggling hipify-clang, omitting the
path is no longer something we need to check for. We'll still
abort if LLVM isn't found, due to `REQUIRED`.
Instead of deciding whether to build hipify-clang based on
the presence of an LLVM path on the command line, have an
explicit option.
Do we want this default-on or default-off? I've defaulted it to
on for now, but maybe we want the opposite?
All but the most contrived use of macros is now properly handled -
have a look at the new testcases this commit adds. You can have
macros in kernel calls, macros spanning chunks of your arguments,
the call, call parameters, or callee can all be macros or
partially macros.
It seems like there was a lot of machinery here that is no longer
needed now we have hipLaunchKernelGGL (which doesn't require us
to insert an extra argument into kernel functions). We no longer
need to waste cycles scanning the AST for callees.
We can literally just do "Take the callee expression, and dump
it into the first argument of hipLaunchKernelGGL()".
There's three functions here that all do the same thing...
There was also logic that looks for numeric literals and works
backwards to find the macro name from which they are expanded.
I previously introduced code that rewrites macro references at
expand-time in the `MacroExpands` callback, so that code is no
longer doing anything useful.
Calls to macros that were themselves CUDA API calls were often
being missed - this applies the identifier transform to macro
names at the callsites, too.
Evgeny Mankov
Evgeny Mankov
Chris Kitching