The peril of data races. Shows how even the most innocent ones can break badly.
The generation of Huffman codes is used in many applications, among them the DEFLATE compression algorithm. The classical way to compute these codes uses a heap data structure. This approach is fairly efficient, but traditional software implementations contain lots of branches that are data-dependent and thus hard for general-purpose CPU hardware to predict. On modern processors with deep...
One Stop for Optimizing Your Data Center From AI to Big Data to HPC: End-to-end Solutions
Learn more about an in-depth analysis of code modernization performance conducted by optimizing original CPU code and re-running tests on the latest GPU/CPU hardware.