performance

CppCon 2014 Writing Data Parallel Algorithms on GPUs--Ade Miller

While we wait for CppCon 2015 in September, we’re featuring videos of some of the 100+ talks from CppCon 2014. Here is today’s feature:

Writing Data Parallel Algorithms on GPUs

by Ade Miller

(watch on YouTube) (watch on Channel 9)

Summary of the talk:

Today most PCs, tablets and phones support multi-core processors and most programmers have some familiarity with writing (task) parallel code. Many of those same devices also have GPUs but writing code to run on a GPU is harder. Or is it?

Getting to grips with GPU programming is really about understanding things in a data parallel way. This talk will look at some of the common patterns for implementing algorithms on today's GPUs using examples from the C++ AMP Algorithms Library. Along the way it will cover some of the unique aspects of writing code for GPUs and contrast them with a more conventional code running on a CPU.

What Every Programmer Should Know About Compiler Optimizations--Hadi Brais

Everything is in the title:

What Every Programmer Should Know About Compiler Optimizations

by Hadi Brais

From the article:

High-level programming languages offer many abstract programming constructs such as functions, conditional statements and loops that make us amazingly productive. However, one disadvantage of writing code in a high-level programming language is the potentially significant decrease in performance. Ideally, you should write understandable, maintainable code—without compromising performance. For this reason, compilers attempt to automatically optimize the code to improve its performance, and they’ve become quite sophisticated in doing so nowadays. They can transform loops, conditional statements, and recursive functions; eliminate whole blocks of code; and take advantage of the target instruction set architecture (ISA) to make the code fast and compact. It’s much better to focus on writing understandable code, than making manual optimizations that result in cryptic, hard-to-maintain code. In fact, manually optimizing the code might prevent the compiler from performing additional or more efficient optimizations...

Yes, we must replace it++ with ++it--Andrey Karpov

I decided to find out if there is practical sense in writing ++iterator instead of iterator++ when handling iterators.

Is it reasonable to use the prefix increment operator ++it instead of postfix operator it++ for iterators?

by Andrey Karpov

From the article:

I will always write ++it. I did so before but I did it "just in case". Now I can see how useful it is because I regularly launch debug versions. In general, of course, ++it has a very slight influence on the running time. But if I don't make such small optimizations in different places of the code, it will be too late and the profiler won't help me. Bottlenecks will be spread throughout the code.

 

Expression Templates Revisited

A new video from Meeting C++ 2014!

Expression Templates Revisited

by Klaus Iglberger

From the talk description:

Since their invention in 1995, Expression Templates (ETs) have proven to be a valuable tool for many C++ template libraries. Especially numerics libraries quickly embraced them as salvation for the performance deficiencies of standard C++. This reputation as performance optimization...

C++ SIMD parallelism with Intel Cilk Plus and OpenMP 4.0

A new video from Meeting C++ 2014

C++ SIMD parallelism with Intel Cilk Plus and OpenMP 4.0

by Georg Zitzlsberger

From the talk description:

Performance is one of the most important aspects that comes to mind if deciding for a programming language. Utilizing performance of modern processors is not as straight forward as it has been decades ago. Modern processors only rarely improve serial execution of applications by increasing their frequency or adding more execution units.

Generic Parallel Programming

A new video from Meeting C++ 2014:

Generic parallel programming for scientific and technical applications

by Guntram Berti

From the talk description:

Technical and scientific applications dealing with a high computational load today face the challenge to match the increasingly parallel nature of current and future hardware. The talk shows how the increased complexity of software can be controlled by using generic programming technologies. The process and its advantages are introduced using many concrete examples...

C++11 library components to make a programming scientist happy

A new video from Meeting C++ 2014:

C++11 library components to make a programming scientist happy

by Peter Gottschling

From the talk description:

This talk presents a survey over components of the C++11
standard library which facilitate the life of a programming scientist.
It will provide a gentle introduction to the random number generator
and evolve this to a complete application: the stochastic simulation
of share price development.

The C++ Memory Model - Valentin Ziegler @ Meeting C++ 2014

A new video from Meeting C++ 2014:

The C++ Memory Model

by Valentin Ziegler

From the talk description:

The C++ memory model defines how multiple threads interact with memory and shared data, enabling developers to reason about concurrent code in a platform independent way. The talk will explain multi-threaded executions and data races in C++...

Monads in Chains - Ivan Cukic @meetingcpp 2014

Ivan Cukic gave a very good talk on monads at Meeting C++ 2014:

Monads in Chains

by Ivan Cukic

From the talk description:

Monads are scary, and monads are evil. But they are still useful.

In the recent years, the abuse of multi-threading has become apparent and more and more systems started being developed in the reactive, or event-processing style. It allows lowering the number of threads to match the system cores instead of items that are being processed by using non-blocking asynchronous APIs.