Extreme Editions: New Releases for All Intel® Software Tools

I’m pleased to share that today, Intel released Intel® Parallel Studio XE 2011 and Intel® Cluster Studio 2011 for Linux* and Windows*.

These bring together, the latest versions of Intel’s industry-leading C/C++ and Fortran compilers, performance and parallelism libraries, correctness analyzers and performance profilers to help improve application performance, code quality, and reliability. Above and beyond the features of Intel Parallel Studio XE 2011, Intel Cluster Studio includes capabilities for distributed memory programming using MPI.

In terms of prior products, these products are:

  • Significant upgrades still focused on top performance for your applications.
  • v12.0 – Intel’s latest C/C++ and Fortran compilers
  • Updated MKL and IPP libraries (Intel® Math Kernel Library and Intel® Integrated Performance Primitives respectively)
  • Update Intel® MPI Library
  • Intel® VTune™ Amplifier XE is the new product name for the best VTune Intel has ever made! The performance tuning tool of choice for many, sports new features and completely reworked user interface based on the very popular WhatIf prototype known affectionately as “PTU.” The coolest new feature: “frame analysis”… we need a whole blog just on this feature – it is amazing!
  • Debut of Linux versions of key innovations from the Intel Parallel Studio 2011, including: memory checking, and parallel programming models.
  • Intel® Inspector XE 2011 is the new product name for combining the groundbreaking Intel®Thread Checker with memory checking capabilities.
  • Intel® Composer XE 2011, new product name for Intel® Compiler Suite Professional Edition with the latest compiler and libraries from Intel
  • Intel® Trace Analyzer and Collector with enhancements. The new Ideal Interconnect Simulator helps figure find imbalances by simulating application behavior in the "ideal communication environment."
  • Fortran includes the Microsoft Visual Studio shell (no need to buy Visual Studio again!), and the Intel® Composer XE and Intel® Amplifier XE work within that shell.

Many new features, highlights include:

  • Intel® Parallel Building Blocks (PBB)
  • solutions for parallel programming in C and C++
  • composable, interoperable solutions for task and data parallelism
  • explicit support for vectorization to utilize SIMD instructions including SSE and AVX
  • elegant new array notation, including elemental functions, for C and C++ programmers
  • solutions for multicore today and ready for many core programming tomorrow
  • includes version 3 of Intel® Threading Building Blocks
  • includes Intel® Cilk™ Plus support
  • Support for AVX, tuned and ready for use. Being ready for AVX can be a big performance boost for applications!
  • Frame analysis in VTune
  • Threading error detection extended to cover .NET codes. So possible dead-locks and race conditions can now be detected even in .NET code.
  • Co-array Fortran support – both node-level and distributed (via MPI).
  • Standards tracking for C++ (0x) and Fortran (nearly all of Fortran 2003, key features of Fortran 2008 including coa-array)
  • Static Security Analysis tool for tracking down more problems such as uninitialized data and buffer overflows.
  • Support for the latest environments: Microsoft Visual Studio 2010 (while still supporting 2005 and 2008), Eclipse, Windows 7, and the latest Linux distributions.
  • And, of course, updates to support the latest processors and processor features.  As always, striving to give the best performance of any compiler library – one Intel and compatibles.

Intel works to win business by striving to offer the best performance of any compiler or library. Please send feedback if you find any issues. Lawyers won’t allow me to guarantee Intel will fix it and always win, but the results speak for themselves. I’m confident you’ll find Intel can help the performance of your application. You can refer to the optimization notice for more information regarding performance and optimization choices in Intel® software products.

More information on these products, including how to get trial editions, can be found at intel.com/software/products.

For more complete information about compiler optimizations, see our Optimization Notice.


anonymous's picture

Do you have a data sheet? Do you support windows?

anonymous's picture

Thanks sir really cool post Intel chipsets are truly fantastic

Michal Kvasnicka's picture

The new fortran standard and co-arrays are of course very important part of recent fortran development, but on the other hand, after more than one decade of development there are still no comprehensive tool for, at least, semi-automatic parallelization of standard serial fortran codes.

For example in nuclear energy (ractor physics) industry are more than 75% of simulation codes still strictly and only serial Fortran77 codes. These codes are extremely realiable and robust, but manual parallelization is more or less imposible or extremely complicated task, because in some cases the authors are not able to help with reprogramming.

So finally, at least from my point of view, is something like parallel Advisor for Fortran most demanded tool.

James R.'s picture

Our Fortran energy for this release went into a great implementation of the new Fortran standard (2008) and its parallel programming support... co-arrays. We support co-arrays on nodes (shared memory) and clusters (distributed memory). Our tools support co-array very aggressively.

I'd be interested in your experiences with "Intel(r) Parallel Advisor" so far. I'm glad it is so useful that you want it for Fortran too!
Yes, the initial version (released only 9 weeks ago) is focused exclusively on Windows and C/C++ code... we want to learn from that and see if it can be applied more places. Your experiences can help.

The co-array approach for parallelism is very promising for Fortran - we'll be eager to hear from people on their experiences with co-arrays.

Michal Kvasnicka's picture

OK ... and where is equivalent of the "Intel® Parallel Advisor 2011"??? Something like Parallel Advisor for Fortran is most demanded tool, from my point of view.

This is a really disapointed, because my job is to parallelize old legacy and large (about 200k lines per package) Fortran77 codes where is very difficult to parallelize via standard programming, due to the fact, that practicaly nobody knows detailed algoritmic implementation of the Fortran code. So, only automatic and/or semi-automatic parallelization on source level is possible way how to renew these codes for future usage on modern multicore computers.


Add a Comment

Have a technical question? Visit our forums. Have site or software product issues? Contact support.