The NVIDIA RTX A4500 was launched with little fanfare. This new GPU is designed for professionals, but perhaps the strangest part is that it was not a card with a featured cameo at NVIDIA’s fall 2021 GTC event. The GPU itself is designed for professional applications making it occupy a segment that we would have formerly called Quadro.
NVIDIA RTX A4500 20GB GPU for Professionals
Let us just skip ahead to the specs since that is going to be the big one many will want to see first:
A few features stand out. This is a 200W TDP GPU with 20GB of GRRD6 memory. That memory is made more interesting since it is ECC memory. This GPU also has four DisplayPort 1.4 outputs. Perhaps the other very interesting point of this actively cooled card is that it supports NVLink.
One can use two of these cards in NVLink. For users that need ECC memory, this could be a very interesting solution. The card itself is sold through OEM partners like PNY, Leadtek, and systems OEMs but the 16GB A4000 retails for around the same price as a RTX 3070 Ti. In a GPU supply-constrained world, many may find an option like this intriguing with the addition of NVLink and ECC memory.
Can we just say it? We miss the Quadro branding. This feels a lot like a workstation GPU but without the workstation branding. Instead, we get the addition of “RTX” making it sound a bit more consumer-ish but with the cooling solution that is more aligned to the server and workstation brand design. Still, having another OEM-focused solution may help on the supply side for some workstation users and that is very welcome. Getting GPUs at reasonable prices is still very difficult so we welcome this move for the STH user base.
Those so called “professionals” are gonna be crypto miners.
Will these be replaced by something like the Mythic Processor: https://www.mythic-ai.com/product/mp10304-quad-amp-pcie-card/ ?
It can do 200 TOPs with 2 cards (double wide) using 1/4 the power.
Rob – these are GPUs with display outputs so those are not comparable.