Nvidia is updating its entire mobile lineup of workstation graphics cards to Ada Lovelace and is also launching an RTX 4000 SFF card.
At the end of January, Nvidia announced the first professional graphics card to use the new Ada Lovelace architecture: RTX 6000 Ada Generation. This powerhouse came with 48 GB of GDDR6 ECC memory, 18,176 CUDA cores, 568 Tensor cores and 142 RT cores. Since then, things have remained quiet around the rest of the line-up, until now.
Nvidia is not yet announcing any new desktop models, but is launching the full mobile line-up. Only at the bottom of the range does the offer remain unchanged. The RTX A1000 and RTX A500 are still closing ranks one year after their launch. These two GPUs run on the older Ampere architecture.
Ada Lovelace Mobile GPUs
The new chips all use the Ada Lovelace architecture. According to Nvidia, the GPUs double the performance compared to Amps. The new architecture is also a lot more energy efficient, which is a handy bonus in a workstation.
| Reeks | CUDA cores | RT cores | The Tensor Core | GDDR6 memory | Memory interface | Memory bandwidth | Peak consumption | |
| Nvidia RTX 5000 | There’s Lovelace | 9.728 | 76 | 304 | 16 GB ECC | 256 bits | 576 GB/s | 175 watt |
| Nvidia RTX 4000 | There’s Lovelace | 7.424 | 58 | 232 | 12 GB ECC | 192 bits | 432 GB/s | 175 watt |
| Nvidia RTX 3500 | There’s Lovelace | 5.120 | 40 | 160 | 12 GB ECC | 192 bits | 432 GB/s | 140 watt |
| Nvidia RTX 3000 | There’s Lovelace | 4.608 | 36 | 144 | 8 GB ECC | 128 bits | 256 GB/s | 140 watt |
| Nvidia RTX 2000 | There’s Lovelace | 3.072 | 24 | 96 | 8 GB | 128 bits | 256 GB/s | 140 watt |
| Nvidia RTX A1000 | Ampere | 2.560 | 20 | 80 | 6 GB | 96 bits | 168 GB/s | 95 watt |
| Nvidia RTX A500 | Ampere | 2.048 | 16 | 64 | 4 GB | 64 bits | 112 GB/s | 60 watt |
The new Ada Lovelace graphics cards have more CUDA cores than their predecessors, more RT cores (3rd generation) and more Tensor cores (4th generation). For the first time, the 16 GB memory (GDDR6 ECC) cape is also passed with the mobile RTX 5000 Ada Generation.
RTX cards enjoy additional optimizations that traditional Nvidia GeForce GPUs don’t have through the Studio drivers. For example, Nvidia RTX GPUs work closely with more than 110 creative apps to deliver better stability and clean performance through dedicated Enterprise drivers.
RTX 4000 SFF Ada Generation GPU
In addition to the complete renewal of the mobile line-up, Nvidia also makes one desktop GPU available under the name RTX 4000 SFF Ada Generation. SFF here stands for Small Form Factor, which means that the card is suitable for compact desktop systems. It is only the second RTX Ada Lovelace desktop card after the RTX 6000.
The RTX 4000 SFF has 6,144 CUDA cores, 192 tensor cores and 48 RT cores on board. The card is equipped with 20 GB GDDR6 with ECC. The maximum consumption is 70 watts to realistically fill the compact spaces within SFF PCs. The card has a suggested retail price of $1,250 and is available through specialty retailers worldwide.
Confusing names
By the way, don’t call the Nvidia RTX cards a Quadro card, because Nvidia ditched that sub-brand a while ago. Unfortunately, a consistent new naming strategy remains missing. For the sake of clarity, you’re looking for ‘Nvidia RTX Ada Generation’, not to be confused with the Nvidia RTX A-series, because that is the predecessor, and neither for Quadro RTX, because it is even older.
