www.imag.com.cn 站内新闻 TotalView for HPC  
TotalView for HPC[ 2013-6-17 ]


TotalView® 8.12产品更新内容



Dear Customer,

We are proud to announce the release of TotalView® 8.12, which features official support for Intel® Xeon Phi™ coprocessors, full support for the Cray® XC supercomputer series, and support for Apple® Lion and Mountain Lion platforms. 

TotalView's support for Intel Xeon Phi coprocessors enables you to debug scenarios where you leverage Intel Xeon Phi coprocessors through Intel Language Extensions for Offloading (LEO) to incrementally speed up specific parts of programs that already run on the host processor. You can also use TotalView to debug scenarios in which you recompile your application to run natively on the Intel Xeon Phi coprocessor. This is often done in conjunction with MPI and takes advantage of TotalView’s mpiexec integration. If you run your Intel Xeon Phi coprocessor native code manually or want to debug an already running code, then this scenario is supported using the remote debugging functionality and treating the Intel Xeon Phi coprocessor as a remote node. 

New features of this release include:

  • Intel Xeon Phi coprocessor support
  • Cray XC support
  • Apple OS X Lion and Mountain Lion support
  • The new Sessions Manager, a framework that saves time from one debugging session to the next
  • Initial support for AVX instructions
  • Support for the Cray Abnormal Termination Processing (ATP)
  • New capabilities for sorting and filtering breakpoint addresses
  • Extended STLView support for the STL set, multi-set, and multi-map templates
  • Improved start up, stepping, variable viewing, and breakpoint setting performance on C++ programs that make extensive use of templates
  • Improved performance at scale

We also have several platforms updates, including:

  • RH EL 6u2, Fedora 17, Fedora 18
  • gcc/gfortran 4.7.2, pgi 12.8, Intel 2013 compiler suite
  • OpenMPI 1.6.3, Intel MPI 4.1

Download the product update.

For additional information, please visit:




























Developing parallel, data-intensive applications is difficult. We make it easier.


本文内容来自Rogue Wave新闻邮件。