In this sponsored post on behalf of Intel, we see that in today’s high-performance computing applications, many different pieces of hardware can perform data-centric functions. With diverse accelerators entering the market, programming for multiple architectures has created significant development barriers for software developers.
Paradigm Change: Reinventing HPC Architectures with In-Package Optical I/O
In this white paper, our friends over at Ayar Labs discuss an important paradigm change: reinventing HPC architectures with in-package optical I/O. The introduction of in-package optical I/O technology helps HPC centers accelerate the slope of compute progress needed to tackle ever-growing scientific problem sizes and HPC/AI convergence. Ayar Labs expects to not only see its technology extend the traditional type of architecture to put the HPC industry back on track, but also result in an inflection point that fundamentally changes the slope of the compute performance efficiency curve. The key will be enabling converged HPC/AI centers to build systems with disaggregated CPUs, GPUs, FPGAs and custom ASICs interconnected on equal footing.
2nd Generation Intel® Xeon® Scalable Processors Demonstrate Amazing HPC Performance
In this guest article, our friends at Intel discuss how benchmarks show key workloads average 31% better on Intel Xeon Platinum 9282 than AMD EYPC “Rome” 7742. Intel analysis provides strong evidence that the 2nd Generation Intel Xeon Scalable Processor (Cascade Lake “CLX”) architecture provides dramatic performance for real-world workloads. An impressive array of benchmarks shows 2S systems built with Intel’s 56 core processors (Intel Xeon Platinum 9282 processor) solidly ahead of systems built with AMD’s 64 core processors (AMD EYPC 7742).
Optimizing in a Heterogeneous World is (Algorithms x Devices)
In this guest article, our friends at Intel discuss how CPUs prove better for some important Deep Learning. Here’s why, and keep your GPUs handy! Heterogeneous computing ushers in a world where we must consider permutations of algorithms and devices to find the best platform solution. No single device will win all the time, so we need to constantly assess our choices and assumptions.