View Full Version : Announcing the Release of TotalView 8.9.1

06-09-2011, 09:41 AM
I'm pleased to announce the availability of TotalView 8.9.1. The focus of our development for this release cycle has been to include support for CUDA 3.1 and 3.2. This release includes the following new features when TotalView is used with the (separately priced) CUDA option:

* Support for apps built with the CUDA 3.0, 3.1, and 3.2 SDK
* Compatibility with CUDA 3.0, 3.1 and 3.2 drivers
* Support for CUDA function calls on the stack (as opposed to inlined)
* Handles exceptions in CUDA code
* Display of variables in GPU hardware registers
* Support for host pinned memory regions
* Support for CUDA contexts
* New CLI commands for CUDA functionality

In addition to our CUDA work we made improvements to our test suite, fixed quite a number of bugs and added the following new features.

* CLI commands for array statistics
* Improved array type inheritance in the multi-dimensional array display
* Improved handling of recursion in the parallel backtrace

The following operating systems are now supported

* RHEL 6
* Fedora 14
* AIX 6.1.5

The following compilers have been added to the supported compiler list:

* GCC/GFortran 4.5.2
* Intel Composer XE 2011 for Linux and Mac
* PGI 11.2

The following new MPI libraries are supported:

* Intel MPI 4.0.1
* POE 5.2

Current users with maintenance are encouraged to obtain copies of the updated software for 30 days at the following URL


after that time, contact Rogue Wave to get access to the latest and greatest in parallel debugging.

Chris Gottbrath