Nvidia’s Tiny RTX 4000 SFF 20GB Offers RTX 3070 Performance at 70W
There are many compact modern workstations with very high performance CPUs, but at the same time lack the space to accommodate standard high performance workstation grade graphics cards. This typically limits performance to mediocre entry-level GPUs. If you want a compact his SFF workstation with more graphics, Nvidia has introduced a new his ProViz-oriented workstation. RTX 4000 SFF Ada Generation (opens in new tab) graphics card. This is one of his most interesting products in recent years, packing a high-end GPU into a thin form factor with a power consumption of just 70W.
The Nvidia RTX 4000 SFF Ada board uses the company’s AD104 graphics processing unit, with 6144 CUDA cores (out of a total of 7680) enabled. This is the same GPU as the RTX 4070 Ti, but with fewer active cores and a boost frequency capped at around 1560 MHz, lowering the total power of the board. The graphics card, on the other hand, comes with 20 GB of GDDR6 memory with ECC that connects to the GPU using a 160-bit interface, so there’s plenty of memory for workstation use.
The GPU has two NVENC encoders and two NVDEC decoders active, but Nvidia has not mentioned the exact functionality of these units. It should be similar to NVDEC, and video encoding performance and quality can be found in a recent roundup of GPUs.
The GA104 chip in this configuration delivers 19.2 TFLOPS of peak single precision performance, theoretically matching the GeForce RTX 3070. It has a peak RT performance of 44.3 TFLOPS and a peak FP8/INT8 tensor performance of 306.8 TFLOPS/TOPS.
Nearly 20 FP32 TFLOPS may be dwarfed by the blazing performance of Nvidia’s RTX 6000 Ada Generation or GeForce RTX 4090, but the RTX 4000 SFF is a thin dual-slot graphics card that can be used in almost any desktop computer. Fits No spare auxiliary PCIe power connector. Interestingly, the RTX 4000 Ada’s 153/306.8 INT8 TFLOPS (non-sparse and sparsity respectively) performance is very close to that of Nvidia’s GeForce RTX 3090 Ti, which is more expensive and consumes much more power. It’s getting bigger.
|row 0 – cell 0||GPUs||FP32 CUDA core||FP32 TFLOPS||INT8 TFLOPS||memory configuration||TBP||Manufacturer’s suggested retail price|
|GeForce RTX 4070 Ti||AD104||7680||40 TFLOPS||160/320 TFLOPS||12GB 192bit 21GT/s GDDR6X||285W||$799|
|GeForce RTX 4070||AD104||5888 (?)||?||?||12GB 192bit 21GT/s GDDR6X||250W (?)||?|
|RTX 4000 Ada Generation||AD104||6144||19.2 TFLOPS||153/307 TFLOPS||20GB 160bit 16GT/s GDDR6 ECC||70W||$1,250|
|GeForce RTX 3090 Ti||GA102||10,752||40 TFLOPS||160/320 TFLOPS||24GB 384bit 20GT/s GDDR6X||450W||$1,999|
|GeForce RTX 3070||GA104||5888||20.31 TFLOPS||81/160 TFLOPS||8GB 256bit 14GT/s GDDR6||220W||$499|
Since this is a workstation grade add-in board, it comes with 4 DisplayPort 1.4a connectors and has a 3 pin mini DIN connector for stereoscopic 3D output (like Nvidia 3D Vision) for multi-application viewing.
When it comes to multi-display applications, one of the advantages of Nvidia’s RTX 4000 Ada Generation graphics cards’ compact dimensions, low power consumption, and wide compatibility is the ability to install a large number of such boards in a relatively compact system. It is possible. It can drive multi-display and video wall applications without the need for his high wattage PSU. Such systems are widely used in various industries such as aerospace, healthcare, military, professional A/V, digital signage and security.
Starting in April, Nvidia’s newly released RTX 4000 SFF graphics card for professional visualization applications will be available from the company’s reseller partners including Leadtek, PNY and Ryoyo Electro for a suggested price of $1,250. Additionally, workstation manufacturers plan to offer this product later this year.