ANALYSIS CLUSTER ADD-ON
The Analysis Cluster add-on facilitates parallel analysis in HALO. The add-on was designed to facilitate faster processing speeds for laboratories analyzing large number of images or requiring fast turn-around for time-sensitive analysis jobs.
By default, HALO will use all processing cores or threads available to analyze multiple image tiles in parallel within a single image. Most image analysis systems are limited to one tile at a time. This Multi-core processing feature is one of the main reasons why HALO analysis is so fast! The analysis cluster add-on takes multi-core processing to the next level by facilitating parallel analysis of multiple images. Depending on the hardware configuration, this dual parallelization (multiple cores across multiple images) cuts the processing time down several fold and allows customers to make the most of their server’s processing power. Each analysis cluster can facilitate the analysis of four jobs in parallel. Depending on your hardware configuration, you may have the option to add more than one analysis cluster for even faster speeds.
In the presentation playing above, you can see a comparison of analysis speeds using tile-by-tile (single core) processing, multicore processing with HALO alone and multi-core processing with analysis clusters. In the Ki67 study, 50 image images are analyzed with a standard cytonuclear IHC module. While it took almost a day (21.7 hours) to analyze this batch of images using tile-by-tile analysis, HALO with two analysis clusters was able to analyze in less than an hour. Similarly impressive results were obtained when we analyzed 50 multi-channel fluorescent slides (Insulin/Brdu study) with our cytonuclear FL module. HALO with analysis clusters was able to complete analysis across the 50 images in less than 2 hours while it took more than a day to complete (31.8 hrs) with tile-by-tile analysis.
Want to know more? Contact us at firstname.lastname@example.org to discuss your options with an application scientist.