If you watched or read about NVIDIA CEO Jensen Huang’s SIGGRAPH keynote address last week, you’re aware that NVIDIA has upgraded its workstation GPU line with new RTX Ada generation cards that offer tremendous performance gains over previous generations. These GPUs include the NVIDIA RTX 5000 Ada which delivers a 3x increase in performance over the RTX A5500, the RTX 4500 which delivers a 2.7x increase over the RTX A4500, and the RTX 4000 with a 1.7x increase over the RTX A4000.
However, for the purposes of this blog, I’ll direct your attention to the NVIDIA RTX 6000 Ada generation GPU which delivers state-of the-art performance for artificial intelligence workloads. It’s also the GPU you’ll find inside our line of BOXX AI workstations which bring AI to your desk side.
You may envision AI systems only as data center platforms like the top-of-the-line, elite-class BOXX RAXX W3, a rack mounted workstation powered by an Intel Xeon W9 processor (up to 56 cores) and available with up to four NVIDIA RTX 6000 Ada GPUs. But as AI technology expands beyond compute and is integrated into more software applications, a desk side workstation purpose-built to power AI-augmented workflows is becoming increasingly essential.
Speaking of essential, BOXX AI desk side workstations begin with the essential-class APEXX W3. This liquid-cooled, subcompact system features an Intel Xeon W5 processor and a single NVIDIA RTX 6000 Ada GPU. Next up is the more advanced-class APEXX W4, a liquid-cooled tower featuring an Intel Xeon W7 and available with up to two NVIDIA RTX 6000 Ada GPUs. Both workstations provide ample memory and storage.
Even though current AI news is focused on the NVIDIA GPUs, these Intel Xeon W CPUs are groundbreaking in their own right, providing up to 64 PCIe lanes in the APEXX W3 and 112 in the APEXX W4 to support high bandwidth devices like the new GPUs, high speed NICs, and NVMe storage.
So whether it’s AI training and development, AI inference, data analytics, any combination thereof, or another category entirely, AI on your desk in general, and BOXX AI workstations in particular, provide distinct advantages. Let’s look at a few.
Training models requires accuracy which means multiple iterations, and the faster these are completed, the faster your project can move to discovery and deployment. A BOXX AI workstation enables and accelerates smaller iterative model training, accommodating more iterations and experimentation before you spend time and effort (as well as money) on large models.
The scalability of BOXX AI workstations is another key advantage. With multiple configuration options from GPUs and CPUs, to system memory, you can tailor your workstation based on the requirements of your models. With the APEXX W4, for example, you could always begin with a single GPU and add a second later on as your AI workload increases.
Data security is obviously critical, so BOXX AI workstations enable users to meet requirements for sensitive and proprietary enterprise datasets like intellectual property content or customer info.
Total cost of ownership (TCO) may be another advantage when opting for an AI desk side workstation. Of course, you have the initial upfront cost, but that’s likely a wiser financial decision than metered (pay-to-use) services requiring long term contracts and complex licensing agreements.
As AI drives change across every industry and integrates into the applications we rely on, the future is moving ever closer to the present. To configure your AI desk side system, visit BOXX AI workstations, download the recommended system pdfs, and consult with a BOXX performance specialist via chat or by calling 877.877.BOXX.