The Visual Studio 11 Developer Preview brings some excellent additions to the experience of debugging parallel computing applications for both multi-core CPUs and GPUs. One of the key new areas is the Parallel Watch Window, allowing you to set up watches on variables across many threads and contexts. When the contexts get as high as thousands, you can use data visualizers in the window to more easily sort, arrange, and view the values, helping you identify issues and outliers very quickly.
This is a pack of 4 visualizers for the parallel data, created by a brilliant team of interns who came up with some very useful and attractive ways to display this information.
Histogram bucketizes the thread values, making it easy to identify and flag outliers in the data distribution.
Scatterplot graphs 1 or 2 dimensions of values, and is another great way to spot outliers.
Heatmap is an attractive visualizer that’s useful for spotting strange distributions of data.
Object Expander helps drill into all of the members and classes of specific objects, find errors, and minimum and maximum values.
Read more: Greg's Cool [Insert Clever Name] of the Day
Read more: Microsoft Parallel Visualization Pack
QR:
Posts
