Compared to Other Google Cloud Virtual Machines (VMs) the N2 VMs Featuring 2nd Gen Intel® Xeon® Scalable Processors Reduced the Time and Cost to Execute Genomics Analytics Toolkit Tasks by As Much As 25%
Genomics workloads demand a great deal of compute resources. If you’re shopping for public cloud VMs, it’s important to investigate the performance of different options. The right choice can dramatically reduce the amount spent running these workloads. To illustrate the ways VMs can vary in terms of performance and cost, Intel conducted benchmark testing on clusters of three Google Cloud VM series: N2 Standard and C2 Standard, both featuring 2nd Gen Intel® Xeon® Scalable processors, and N1 Standard with 1st Gen Intel® Xeon® Scalable processors.
The Cromwell on Google Cloud benchmark measured how quickly these VMs executed 24 tasks in the Genomics Analytics Toolkit (GATK) application. Test results showed that the N2 Standard cluster performed the tasks in less time than both other clusters, which resulted in a lower cost. For efficiency and cost-effectiveness, run your genomics workloads on N2 Standard VMs featuring 2nd Gen Intel® Xeon® Scalable processors.
N2 Standard VM Clusters Outperformed N1 Standard VM Clusters and Cost Less
As Figure 1 shows, in our testing the N2 VM cluster featuring 2nd Gen Intel® Xeon® Scalable processors completed genomic tasks in the GATK application in 75% of the time the N1 Standard VM cluster needed. When your VM cluster completes a set of tasks in less time, you pay for less VM uptime, which helps your bottom line. This efficiency translates to savings: the cost of executing the tasks on the N2 VM cluster would be 21% lower than on the N1 VM cluster.
N2 Standard VM Clusters Outperformed C2 Standard VM Clusters and Cost Less
As Figure 2 shows, in our testing the N2 VM cluster featuring 2nd Gen Intel® Xeon® Scalable processors completed genomic tasks in the GATK application in 78% of the time the C2 Standard VM cluster needed. This efficiency translates to savings: the cost of executing the tasks on the N2 VM cluster would be 22% lower than on the N1 VM cluster.
Because genomics analysis applications are so compute intensive, selecting a cloud VM with robust performance is vital. In our testing, we showed Google Cloud N2 Standard VMs featuring 2nd Gen Intel® Xeon® Scalable processors reduced both the time and cost to complete genomics tasks, compared to N1 Standard and C2 Standard VMs.
To begin running your genomics workloads on Google Cloud N2 virtual machines with 2nd Gen Intel® Xeon® Scalable processors, visit https://cloud.google.com/compute/docs/general-purpose-machines#n2_machines.
All tests by Intel on GCP/us-central1 on 9/10/2020-10/9/2020. All tests: CentOS 7, Input Data Set 30X Coverage Human Whole Genome Sequence (WGS); NA12878, Workload GATK Best Practices Pipeline for Germline Variant Calling with pre-processing, GATK 220.127.116.11, Genomics Kernel Library (GKL) 0.8.6, Cromwell 52, Picard 2.20, BWA 0.7.15-r1140, Samtools 1.3.1. Tools in https://hub.docker.com/r/broadinstitute/genomes-in-the-cloud/: us.gcr.io/broad-gotc-prod/genomes-in-the-cloud:2.4.3-1564508330, us.gcr.io/broad-gatk/gatk:18.104.22.168. Workflow defined: https://github.com/gatk-workflows/gatk4-genome-processing-pipeline/blob/master/WholeGenomeGermlineSingleSample.wdl. Run Iterations:3. VM details: N1 machine series: Intel Xeon 1st Generation Scalable Processor (Skylake): N1-standard-2: 2 vCPUs, 7.5GB RAM; N1-standard-16: 16 vCPUs, 60GB RAM. N2 machine series: Intel Xeon 2nd Generation Scalable Processor (Cascade Lake): N2-standard-2: 2 vCPUs, 8GB RAM; N2-standard-16: 16 vCPUs, 64GB RAM. C2 machine series: Intel Xeon 2nd Generation Scalable Processor (Cascade Lake): C2-standard-4: 4 vCPUs, 16GB RAM. C2-standard-16: 16 vCPUs, 64GB RAM.