NVIDIA GeForce RTX 5080 Might Come With 24GB variant and GDDR7 Config
New reports have revealed that NVIDIA is aiming for a 24GB variant of the GeForce RTX 5080 with GDDR7 config.
NVIDIA’s CEO Jensen Huang’s refusal to merge with AMD in the 2000s has been justified many times over the past decade. The latest manifestation of this decision is the NVIDIA GeForce RTX 5080 GPU, the leaks of which are fresh all over the internet. As reported at Chiphell Forums, along with the expected 16 GB variant of the graphics card, we would also see a grander 24 GB version.
Speculations about NVIDIA’s GeForce RTX 5080 move from 16GB to 24GB
Yesterday, NVIDIA’s GeForce RTX 5080 specifications were splattered across the internet, and people were surprised. Among the prominent changes was the halving of core and memory specs for the RTX 5080 compared to its successor.
The GeForce RTX 5080 will sport a similar VRAM configuration as the existing RTX 4080 GPUs: 16GB capacity and 256-bit bus interface. However, the memory will be upgraded to include the newer GDDR7 standard. This will bring a significant increase in the bandwidth. We can also expect an L2 cache increment.
Following the trend, we naturally expect the successor to the GeForce RTX 5080, that is, RTX 5090, to raise the VRAM’s limit to 32 GB. While that would be a great improvement on NVIDIA’s part, things don’t point this way. It seems the memory configuration would be inhibited by the GDDR7 yields.
We must remember that the GDDR7 memory architecture is still new. Each technology shift needs a certain amount of time to mature and for the supporting hardware to develop. Higher speeds and memory bandwidth will come along in good time. Right now, the reports indicate a 28 Gbps die speed and 2GB capacities per module.
We have a lot to thank Polymorph for, a Chiphell Forum member. They were responsible for leaking the first images of the GA102 “Ampere” GPU core, the RTX 20 series cards specifications, and the RX 6000 “RDNA 2” GPU configurations. Polymorph says that NVIDIA might manufacture a 24GB variant of RTX 5080 sometime in the future.
The tech guys at NVIDIA could go about creating a 24GB GeForce RTX 5080 card in a couple of different ways. An unlikely and wasteful method is to use the GB202 core with a 384-bit memory bus and a 24GB configuration. A second, innovative, and resourceful method is to use 3GB VRAM modules, which is a GDDR7 prerogative.
With the old 2GB modules, the RTX 5080 and RTX 5090 would sport 16 & 8 GDDR7 units, amounting to 32GB and 16GB of memory respectively. However, with a 3GB module, the RTX 5080 can shoot up to 24GB while sticking to the same PCB design, core specifications, and 256-bit memory configuration.
Everything comes down to how well and how quickly the 3GB modules can be mass-produced. The GDDR7 can accommodate anything between 16GB to 64GB of memory, so the jump to the next-gen GPUs is only a matter of economic and temporal feasibility.
That was everything about the technical aspects of the NVIDIA GeForce RTX 5080. Let’s now talk about the logistics. The rumors have it that the first RTX 5080 and RTX 50 GPU shipments are on the move and on their way to the testing labs.
Harukaze5719 took notice of the various configurations of the RTX 50 GPU. Sadly, most of these point towards the 16GB version of the GPU.
That’s all the internet has on NVIDIA’s GeForce GTX 5080. If you would like to know the nitty-gritty of the configurations, go and have a look at the Chiphell Forums where the leaks originated.
