This image shows the FDM SearchFaster(tm) software performing three(3) searches in quick succession on about 900,000 spectra, most of which are mixture spectra, which is a much heavier task than searching single spectra. The search operation is similar to common image processing operations. This was run on a Windows 11 computer with a top end CPU and high bus speeds, but the take-away is the (formerly) immense computational job is now multithreaded, or multicore, and smoothly scales to all available logical cores and each core is used with great efficiency by maximizing task and data parallelism.

Do you publish CPU-bound scientific, engineering or advanced imaging software written in C++? 

​

Do you struggle with multithreading?

​

We can help you think parallel and improve performance ​with analysis, profiling, modernization, and prototyping.  

​

Please contact us with your requirements.

​

Keywords: parallelism, TBB, MKL, IPP, DAL, DNN, VPL, CCL, image processing, kernel, voxel, scalable, cluster, GPU, FPGA. Numerical Recipes. Scientists and Engineers Guide to Digital Signal Processing. Corp-to-corp. Spark, Yarn, Hadoop.

​