NVIDIA has announced the launch of the RTX A4500, a new dedicated graphics for professional work machines that we already anticipated, and another entry-level RTX 2000 model.
The RTX A4500 it comes to fill the gap in the 4000 and 5000 series. It uses the basic GA102 core, which means that it will have the same base as the RTX A5000, although it reduces its number of CuDA cores to 7168; the Tensor cores at 224 and their working frequency at 1653 MHz.
It also lowers the memory configuration with 20 Gbytes dedicated GDDR6 on a 320-bit bus, while its TDP consumption also drops to 200 watts. Its estimated FP32 performance is 23.7 TFLOPS, while that of RT rises to 46.2 TFLOPS, more than enough for a graphics dedicated to mid-level workstations. Its price will also be cheaper than that of the RTX 5000 and, incidentally, it offers a greater number of alternatives to the user.
NVIDIA has also announced the RTX A2000, the lowest performance ‘Ampere’ architecture workstation card, but also the most economical. It is based on the GA106 kernel and has 3328 stream processors; 104 Tensor cores and 1200 MHz frequency. It includes 12 Gbytes of dedicated memory and its TDP is only 70 watts. It is estimated at a power of 8 TFLOPS.
Both cards are for desktop workstations and offer compatibility with DisplayPort 1.4 4-connector base, full-size on the RTX A4500 and mini-size on the RTX A2000 as the RTX A2000 employs a low-profile design. Of course, they include all the manufacturer’s technologies and certification for common professional applications in CAD, architecture, manufacturing, engineering, etc.
They are expected models and with a known base, but NVIDIA expands with them its catalog of professional graphics that lead -by far- the sales of its segment and offers greater alternatives for desktop work machines.