The compiler is to blame for everything—Andrey Karpov

Many programmers are very fond of blaming the compiler for different errors. Let's talk about it.

The compiler is to blame for everything

by Andrey Karpov

From the article:

When a programmer tells you that the compiler causes an error, it is a lie in 99% of cases. When you start investigating the problem, you usually find out the following reasons:

  • an array overrun;
  • an uninitialized variable;
  • a misprint;
  • a synchronization error in a parallel program;
  • a non-volatile variable used;
  • code leading to undefined behavior;
  • etc.

OpenMP* Support in Clang/LLVM: Status Update and Future Directions—Alexey Bataev & Zinovy Nis

A status update for the work done on OpenMP for Clang/LLVM and future directions presented at the 2014 LLVM Developers' Meeting. The slides are also available here.

OpenMP* Support in Clang/LLVM: Status Update and Future Directions

OpenMP is a well-known and widely used API for shared-memory parallelism. Support for OpenMP in Clang/LLVM compiler is currently under development. In this talk, we will present current status of OpenMP support, what is done and what remains to be done, technical details behind OpenMP implementation. Also, we will elaborate on accelerators and pragma-assisted SIMD vectorization, introduced in the latest 4.0 edition of the OpenMP standard.


ccache 3.2 released

Version 3.2 of ccache was recently released. It can help you be more productive as "It speeds up recompilation by caching previous compilations and detecting when the same compilation is being done again" (from the official page). The new version that has been released is a feature version which means lots of goodies and new features.

ccache 3.2 released

From the announcement:

  • Added support for configuring ccache via one or several configuration files instead of via environment variables. Environment variables still have priority but are no longer the recommended way of customizing ccache behavior. See the manual for more information.
  • Added support for compiler error/warning messages with color.
  • Made creation of temporary directories and cache directories smarter to avoid unnecessary stat calls.
  • Improved efficiency of the algorithm that scans for __DATE__ and __TIME__ tokens in the hashed source code.
  • Added support for several binaries (separated by space) in CCACHE_PREFIX.
  • The -c option is no longer passed to the preprocessor. This fixes problems with clang and Solaris’s C++ compiler.
  • ccache no longer passes preprocessor options like -D and -I to the compiler when compiling preprocessed output. This fixes warnings emitted by clang.

And much more ...


New optimizations for X86 in upcoming GCC 5.0—Evgeny Stupachenko

Fresh on the Intel Developer Zone blog:

New optimizations for X86 in upcoming GCC 5.0

by Evgeny Stupachenko

From the article:

Part 1. Vectorization of loads/stores group.

GCC 5.0 significantly improves vector code quality for load groups and store groups. By loads/stores group I mean iterated consecutive sequence of loads/stores. For example:

x = a[i], y = a[i + 1], z = a[i + 2] iterated by “i” is loads group of size 3


The most frequent case where loads/stores groups are applicable is array of structures.
  1. Image conversion (RGB structure to some other) ...
  2. N-dimentional coordinates. (Normalize array of XYZ points) ...
  3. Multiplication of vectors by constant matrix: ...

... GCC 5.0:

  1. Introduces vectorization of load/store groups of size 3
  2. Improves load groups vectorization for all supported sizes
  3. Maximizes load/store groups performance by generating code that is more optimal for particular x86 CPU...