The 3rd Generation Intel Xeon Scalable processors and life sciences are a bit like apple pie and ice cream: good separately, even better together.
To optimize life sciences codes, Intel has engaged with the open-source community for more than 20 years. Now Intel has applied those learnings to its Xeon family of data center CPUs.
One thing Intel has learned over those 20 years is that high performance computing (HPC) is increasingly important to life sciences. Researchers in this field work with some truly immense datasets. They also use AI techniques that require boatloads of processing power.
To help them, Intel now offers developers in life sciences an open and unified programming model called Intel oneAPI. Developers can also use Intel toolkits for HPC.
Intel is also helping life sciences with the performance improvements in its 3rd Gen Intel Xeon Scalable processors. These gains are impressive. For one example, Intel says its optimized version of NAMD — software for nanoscale molecular dynamics — performs 54% better than a non-optimized version.
3rd gen Intel Xeon Scalable processor
Many of these performance gains are the direct result of the processor line’s feature enhancements. These include increased core counts (now up to 40 per socket); more memory bandwidth (raised to 8 DDR4 channels); and support for both Intel Optane Persistent Memory 200 Series and PCIe 4.0.
Life-sciences users can also enjoy additional HPC and AI gains by using Intel AVX-512, which offers 512-bit vector operations capabilities, and Intel Deep Learning Boost, a set of embedded acceleration instructions.
Life-sciences users can also benefit from a new offering in Intel Select Solutions. That’s Intel’s offering of benchmark-tested and verified solutions from third-party providers.
To create this new offering, Intel collaborated with Broad Institute, a research organization run by Harvard University and MIT. Together, they’ve created a suite of optimized software called Intel Select Solutions for Genomic Analytics.
This Genomic Analytics solution includes optimized hardware. It also features a suite of optimized software that includes reference architectures for turnkey configuration, setup and deployment.
The solution is qualified for genomics-related resources. These include GATK (genome analysis toolkit) pipelines; Cromwell, a workflow-management system; and GenomicsDB, an open-source library and tools.
Intel says this software suite has demonstrated a 5x performance improvement when running GATK 4.0, compared with previous versions of the genomics software. The solution also reduces setup time.
In addition, the Genomics Analytics solution is highly scalable. It can run on anything from small computers to very large clustered supercomputers.
The original version of the Intel Select Solutions for Genomic Analytics was developed using the 2nd gen Intel Xeon Scalable processor. Based on its success, a version based on the newer 3rd gen Xeon Scalable CPU is now being developed, and it’s expected to ship in this year’s fourth quarter.
To discover how HPC performance is made flexible with 3rd Gen Intel Xeon Scalable processors, watch this short Intel video: