NVIDIA debuts its RTX 5000 ADA graphics card
The new workstation GPU is now available in China.
NVIDIA RTX 4500 ADA
In August, NVIDIA made an announcement regarding updates to its RTX ADA series, introducing the new RTX 5000 and RTX 4500/4000 models. It’s important to note that these models should not be confused with the previous generation of RTX Ampere series. The easiest way to distinguish them is by looking for product names containing “ADA.” This family of workstation GPUs has grown sustainably to include 5 models from the initial 2.
To provide some details, the RTX 5000 ADA model boasts the AD102 GPU with a newly configured 12,800 CUDA cores. This represents a significant reduction of 5,376 CUDA cores from the 18K config in the RTX 6000 ADA, which is NVIDIA’s flagship workstation desktop model. The 5000 ADA comes equipped with 32GB of GDDR6 memory with ECC support, marking the first AD102 model to feature a 256-bit memory bus. In contrast, the RTX 4080 relies on the AD103 GPU with a native 256-bit memory bus, powering the 16GB GDDR6X configuration.
The memory bus, along with a lower memory speed of 18 Gbps, will result in notably reduced bandwidth when compared to the 6000 ADA, dropping from 960 to 578 GB/s. Additionally, the single-precision compute performance will decrease from 91 to 65 TFLOPs.
Workstation with four RTX 6000 ADA GPUs
The card has been officially listed by Gigabyte, a certified OEM, on the Chinese JD platform, with a price tag of 35,999 RMB (equivalent to 4,932 USD). Notably, this price represents a 50% reduction compared to the RTX 6000 ADA, which is also available from the same store. It’s important to mention that this price includes the local sales tax, and it’s worth noting that the official MSRP for the RTX 5000 ADA is $4,000. The card is currently in stock and available for immediate shipping.
NVIDIA RTX ADA Workstation GPUs | |||||
---|---|---|---|---|---|
VideoCardz.com | RTX 6000 ADA | RTX 5000 ADA | RTX 4500 ADA | RTX 4000 ADA | RTX 4000 SFF ADA |
Picture | ![]() | ![]() | ![]() | ![]() | ![]() |
GPU | AD102 | AD102 | AD104 | AD104 | AD104 |
CUDA Cores | |||||
Tensor Cores | |||||
Boost Clock | |||||
Memory | |||||
Memory Bus | |||||
Memory Clock | |||||
Max Bandwidth | |||||
FP32 Compute | |||||
TDP | |||||
MSRP |
Source: JD