NVIDIA's GeForce RTX 5080 GPU is rumored to get a 24 GB upgrade after the launch of the 16 GB model as reported at Chiphell Forums.
NVIDIA GeForce RTX 5080 GPU Rumored To Get 16 GB and 24 GB Variants, Possibly A 3 GB GDDR7 Module Upgrade
NVIDIA's GeForce RTX 5080 and RTX 5090 specifications were revealed yesterday and the first thing that everyone was able to point out was the halving of core and memory specs for the RTX 5080 vs its bigger brother. The RTX 5080 is said to feature a similar VRAM configuration as the existing RTX 4080 GPUs with 16 GB capacities and a 256-bit bus interface. However, the memory itself will utilize the newer GDDR7 standard which should bring a nice uplift in bandwidth and we can also expect a L2 cache boost.
Now obviously, 32 GB VRAM on the NVIDIA GeForce RTX 5090 sounds much more enticing, and 24 GB for the RTX 5080 would've been a better choice but it looks like the initial memory configuration might be limited due to GDDR7 yields. GDDR7 is going to be a fairly new memory architecture and the process needs to be mature to allow for higher speeds and capacities. Early reports suggest 28 Gbps die speeds and 2 GB capacities per module but that might change.
According to prominent Chiphell Forum member, Polymorph, who has previously leaked the first GPU pictures of the GA102 "Ampere" GPU core, revealed the configurations of RTX 40 series cards and also detailed RX 6000 "RDNA 2" GPU specs, it looks like NVIDIA may offer a 24 GB variant of the RTX 5080 in the future.
Now there are two ways that a 24 GB NVIDIA GeForce RTX 5080 is going to happen. The first and least likely way is that the green team uses the GB202 GPU core, allowing for a 384-bit memory bus and a 24 GB configuration. But there's another path and the more likely one which is to use 3 GB VRAM modules which will be available with GDDR7.
The RTX 5090 and RTX 5080 will feature 16 & 8 GDDR7 modules, respectively. That gives us 32 GB and 16 GB memory using 2 GB modules but with 3 GB, you can get up to 24 GB for the RTX 5080 while keeping the same PCB design, core specs, and also the same 256-bit memory configuration. It all comes down to when the 3 GB GDDR7 modules will be ready for mass production for next-gen GPUs but it is well-known that GDDR7 scales from 16Gb to 64Gb so a memory upgrade is definitely in the equation.
GDDR Graphics Memory Evolution:
| GRAPHICS MEMORY | GDDR7 | GDDR6X | GDDR6 | GDDR5X |
|---|---|---|---|---|
| Workload | Gaming / AI | Gaming / AI | Gaming / AI | Gaming |
| Platform (Example) | GeForce RTX 5090? | GeForce RTX 4090 | GeForce RTX 2080 Ti | GeForce GTX 1080 Ti |
| Die Capacity (Gb) | 16-64 | 8-32 | 8-32 | 8-16 |
| Number of Placements | 12? | 12 | 12 | 12 |
| Gb/s/pin | 32-42.5 | 19-24 | 14-16 | 11.4 |
| GB/s/placement | 128-144 | 76-96 | 56-64 | 45 |
| GB/s/system | 1536-1728 | 912-1152 | 672-768 | 547 |
| Configuration (Example) | 384 IO (12pcs x 32 IO package)? | 384 IO (12pcs x 32 IO package) | 384 IO (12pcs x 32 IO package) | 384 IO (12pcs x 32 IO package) |
| Frame Buffer of Typical System | 24 GB (16 Gb) 36 GB (24 Gb) | 24 GB | 12GB | 12GB |
| Module Package | 266 (BGA) | 180 (BGA) | 180 (BGA) | 190 (BGA) |
| Average Device Power (pJ/bit) | TBD | 7.25 | 7.5 | 8.0 |
| Typical IO Channel | PCB (P2P SM) | PCB (P2P SM) | PCB (P2P SM) | PCB (P2P SM) |
In addition to the NVIDIA GeForce RTX 5080 24 GB GPU rumor, there is finally a sight of the first RTX 5080 & other RTX 50 GPU shipments being sent out to the testing labs. Harukaze5719 has spotted several RTX 50 GPU configs, mainly pointing toward the 16 GB RTX 5080 variation which you can see below (via NBD.ltd):
- 699-1G144-0030-TS1 (RTX 5090)
- 699-1G144-0050-TS1
- 699-1G144-0045-TS1 (RTX 5080)
- 691-1G145-2030-TS1 (RTX 5090)
- 699-1G147-0070-TS1
- 699-1G147-0050-TS1
- 699-1G147-0045-EB1 (RTX 5080)
699-1G144-0030-TS1
699-1G144-0050-TS1
699-1G144-0045-TS1
691-1G145-2030-TS1
699-1G147-0070-TS1
699-1G147-0050-TS1
699-1G147-0045-EB1https://t.co/VfBUmrurmp https://t.co/pKUBcyw818 pic.twitter.com/X3LNFYif2z— 포시포시 (@harukaze5719) September 27, 2024
We also recently covered a whole bunch of details and what to expect from the upcoming NVIDIA GeForce RTX 50 family including the RTX 5080 GPUs so give it a read here.
NVIDIA GeForce RTX 5090 & RTX 5080 GPU Specs (Rumor):
| Graphics Card Name | NVIDIA GeForce RTX 5090 | NVIDIA GeForce RTX 5080 | NVIDIA GeForce RTX 4090 | NVIDIA GeForce RTX 4080 |
|---|---|---|---|---|
| GPU Name | Blackwell GB202-300 | Blackwell GB203-400 | Ada Lovelace AD102-300 | Ada Lovelace AD103-300 |
| GPU SMs | 170 (192 Full) | 84 (84 Full) | 128 (144 Full) | 76 *80 Full) |
| GPU Cores | 21760 (+33%) | 10752 (+11%) | 16384 | 9728 |
| Clock Speeds | TBD | TBD | 2520 MHz | 2505 MHz |
| L2 Cache | TBD | TBD | 72 MB | 64 MB |
| Memory Capacity | 32 GB GDDR7 (+33%) | 16 GB GDDR7 (0%) | 24 GB GDDR6X | 16 GB GDDR6X |
| Memory Bus | 512-bit (+33%) | 256-bit (0%) | 384-bit | 256-bit |
| Memory Speed | 28-32 Gbps | 32 Gbps | 21.0 Gbps | 23.0 Gbps |
| Bandwidth | 1792-2048 GB/s | 1024 GB/s | 1008 GB/s | 736 GB/s |
| TBP | 600W (+33%) | 400W (+25%) | 450W | 320W |
| Power Interface | 1 12V-2x6 (16-Pin) | 1 12V-2x6 (16-Pin) | 1 12VHPWR (16-Pin) | 1 12VHPWR (16-Pin) |
