SolidRun announced what they’re calling an “AI Inference Server,” the Janux GS31. The Janux GS31 has a 1U form factor, up to 128 Gyrfalcon Technology Inc. SPR2803 AI acceleration chips and targets the edge computing market. SolidRun was founded in 2010 and is known for their tiny ARM architecture CuBox computers, which are cubes about 2 inches in all dimensions. They also provide other edge and IoT products and services. Gyrfalcon Technology was founded more recently, early in 2017, and primarily develops low power AI processors.
SolidRun announced what they’re calling an “AI Inference Server,” the Janux GS31. The Janux GS31 has a 1U form factor, up to 128 Gyrfalcon Technology Inc. SPR2803 AI acceleration chips and targets the edge computing market. SolidRun was founded in 2010 and is known for their tiny ARM architecture CuBox computers, which are cubes about 2 inches in all dimensions. They also provide other edge and IoT products and services. Gyrfalcon Technology was founded more recently, early in 2017, and primarily develops low power AI processors.
The Janux GS31 is, like other many other SolidRun products, an ARM-based appliance. It can support decoding and video analytics of up to a staggering 128 channels of 1080p/60Hz video. That many channels are honestly overkill for monitoring and managing all but the largest surveillance needs, but it comes in a package small enough to be used virtually anywhere. Gyrfalcon’s Lightspeeur 2803S Neural Accelerator chips make it surprisingly energy efficient, delivering up to 24 TOPS per Watt. The entire appliance has a maximum power consumption of 900W (single phase) and needs a power supply of 100V~240V via an IEC60320 connector.
Engage with StorageReview
Newsletter | YouTube | Podcast iTunes/Spotify | Instagram | Twitter | Facebook | RSS Feed
