Nvidia has quietly began promoting its RTX 6000 Ada Technology graphics card. Primarily based on the AD102 GPU with 18,176 CUDA cores enabled, that is the ‘fattest’ AD102 configuration out there to this point — 142 of the 144 Streaming Multiprocessors are enabled. Rated for 300W of energy consumption, the board is designed for laptop aided design, digital content material creation, digital desktop infrastructure, and different skilled purposes.
Pricing varies fairly extensively, starting from $6,800 to $8,600 relying on the retailer. Nvidia has the cardboard listed at $6,800 (opens in new tab), with a restrict of 5 per buyer — simply in case you are attempting to place collectively a bunch of workstations. Different shops like CDW (opens in new tab) are marking it up an extra 25% or so, with availability set for 4–6 weeks out.
Nvidia’s RTX 6000 Ada primarily based on the AD102 GPU has 18,176 CUDA cores enabled, and whereas Nvidia does not specify clocks anyplace that we are able to discover, it is companions like PNY (opens in new tab) and Leadtek (opens in new tab) quote to 91.1 FP32 TFLOPS of compute efficiency. That equates to a GPU enhance clock of 2505 MHz and is about 10% greater than the GeForce RTX 4090, which options 16,384 CUDA cores. Nevertheless, there are clear variations between the RTX 6000 Ada and the consumer-focused RTX 4090.
The RTX 6000 Ada is designed for skilled purposes, and so it carries 48GB of GDDR6 reminiscence with ECC enabled, that includes a peak reminiscence bandwidth of 960 MB/s — barely decrease than 1,008 MB/s provided by the GeForce RTX 4090. That means Nvidia is utilizing 20 Gbps GDDR6 reminiscence moderately than the marginally quicker 21 Gbps GDDR6X, however energy use may additionally be barely decrease on the GDDR6 chips.
The RTX 6000 comes geared up with 4 DisplayPort 1.4a connectors that may drive 4 4K or 5K shows (4K at 240 Hz is supported by way of DSC), or two 8K shows (120 Hz with DSC).
One of many intriguing issues concerning the RTX 6000 Ada is its energy consumption. Regardless of its greater compute efficiency and 48GB of reminiscence onboard, the RTX 6000 Ada is rated for simply 300W of energy, down from the 450W ranking on the gamer-oriented GeForce RTX 4090. Each GPUs have related official enhance clocks of round 2.5 GHz, although the minimal assured clocks on the RTX 6000 are decrease and we suspect it will not clock fairly as excessive in heavy computational workloads.
Due to its comparatively restricted energy consumption, the RTX 6000 Ada Technology graphics card comes with a dual-wide cooling system and a blower fan to make sure that it matches into workstations and servers. It makes use of one 12VHPWR (16-pin CEM 5.0 PCIe) energy connector for energy supply, so it is going to require an applicable cable adapter to suit into present machines that do not need a local 16-pin connector.
One other attention-grabbing factor about Nvidia’s RTX 6000 Ada skilled graphics card is its value. Nvidia sells the board for $6,800, barely decrease than the $6,999 launch value of its RTX A6000 predecessor. However worth added reseller CDW lists the product for $8,615, whereas a Japanese retailer discovered by @momomo_us lists the unit for $8,524 with out tax. We’re not certain why resellers are charging a lot greater than Nvidia itself, however for now it appears prefer it makes essentially the most sense to purchase the RTX 6000 Ada board immediately from the Nvidia.