GigaIO Engineered Solution for life sciences, Powered by AMD
Speed Genomic Analysis by Boosting Accelerator Utilization and Lowering Costs
Challenges
- Sequencers are getting faster and faster, swamping genomics pipelines
- Secondary Analysis processing times are now the bottleneck in genomics
- Personalized Medicine is throttled by the lack of computing resources
Solution
The GigaIO Engineered Solution for Life Sciences, Powered by AMD
- Accelerate analysis with FPGA and GPU resources tailored to the growing needs of expanding genomics workloads
- Works with your existing HPC storage
Benefits
- Speed Secondary Analysis
- Deploy FPGA and GPU resources independently to optimize acceleration
- Lower processing times with GigaIO’s low latency PCIe fabric
- Boost FPGA and GPU utilization by more than 100%
- Easily expand accelerator resources to keep up with future sequencers
- Significantly reduce CAPEX and OPEX
- Eliminate HPC network complexity
- Immediate availability with no supply chain delay
Composable Disaggregated Infrastructure for Genomic Workloads
A new composable architecture is emerging that promises to dramatically improve Genomic Secondary Analysis computational resource utilization. Composable Disaggregated Infrastructure (CDI) disaggregates FPGA and GPU accelerator resources previously locked in physical servers, enabling the provisioning of genomics analysis workloads with the optimal type and amount of accelerator resources to speed analysis, increase resource utilization, and lower costs.
Speed Analysis With Targeted FPGA and GPU Acceleration
Breakthrough Performance and Compelling Economics
Because it disaggregates FPGAs and GPUs from the servers into pools of resources available throughout the genomics analysis cluster, FabreX improves accelerator utilization by more than 100%. In one side-by-side test shown below, a rack built with FabreX ran the same customer weekly workload with 16 GPUs instead of the 32 initially deployed on InfiniBand, lowering costs significantly.
| Fabrex Infrastructure | Infiniband Infrastructure |
 |
|
The GigaIO Engineered Solution for Life Sciences, Powered By AMD
Under the Hood: What Makes GigaIO’s Composable Infrastructure So Fast?
GigaIO’s software-first composable platform increases resource utilization and lowers the cost of ownership by allowing genomics researchers to compose individual resources as needed and on the fly to adapt to changing workflows.
As shown above, the GigaIO Engineered Solution for Life Sciences delivers the industry’s lowest latency and the highest effective bandwidth by integrating compute, FPGA, and GPU acceleration I/O into a single-system cluster fabric, using industry-standard PCI Express (PCIe) technology. GigaIO enables true server-to-server communication across PCIe and makes cluster-scale computing possible via direct memory access by an individual server to system memories of all other servers in the cluster fabric, for the industry’s first in-memory network.
Patented GigaIO technology strips away unnecessary conversion, software layers, and overheads that add latency to legacy interconnects. HPC storage is easily integrated via the HPC Cluster servers as they normally would in conventional infrastructures.
FabreX Building Blocks Compose Ultimate Power and Flexibility
A complete Engineered Solution for Life Sciences can be built out of a few simple FabreX Building Blocks. Lower CapEx and OpEx through less hardware, higher utilization of resources, quicker adoption of new technology, lower power consumption, and less cooling. Avoid over-provisioning and add just the components you need. Maximize utilization of your datacenter footprint and contribute to your bottom line.
The FabreX Fabric Manager (FM) is the central building block and the software engine that drives the performance and dynamic composability of GigaIO’s software-defined hardware. This Linux-based, resource-efficient software layers onto FabreX hardware for easy-to-use composing of computing clusters on-the-fly. A classic Top-Of-Rack network switch, the FabreX Switch communicates with FabreX server drivers to identify and coordinate resources required by the servers.
The FabreX Network Adapter card is the high-performance, cabled interface to cluster subsystems across FabreX. The card includes both server and target (for PCIe I/O) modes and is FPGA-powered for configuration flexibility. Applications can access remote PCIe devices as if they were attached to the local system. The GigaIO Accelerator Pooling Appliance is the industry’s highest-performing PCIe accelerator device, with up to 1Tb/sec bandwidth into and out of the box. It delivers advanced provisioning and monitoring of accelerators, supporting up to 8 double-width PCIe Gen 4.0 x16 accelerator cards and 2 PCIe Gen 4.0 x16 low-profile slots. Industry-standard HPC cluster servers and GPU servers are also integrated as shown.
Outstanding Simplicity and Easy Management
Accessing the total power of GigaIO FabreX is as simple as using the same cluster management and workload scheduling managers as you always have. Moreover, there are no problematic NIC or OFED driver bugs, incompatibilities, etc., and there are no more supply chain delays. And down the road, you can add compute, storage, and application accelerators at the component level that plug-n-play with your environment. Every primary subsystem can now operate on its own upgrade cycle. The total cost of the system is optimized over its lifecycle, as FabreX drives much higher utilization of all resources.
Resources:
From Engineering.com: There is an Easy and Affordable Way to Improve HPC Performance
