RTX 4090 vs RTX 3090
Manufacturer specifications of the RTX 4090 and the RTX 3090, side by side, per GPU. A value the manufacturer does not publish for the exact model is marked as such, never estimated.
NVIDIA GeForce RTX 4090 24GB
Ada Lovelace
- GPU memory
- 24 GB
- Memory bandwidth
- 1,008 GB/s
Live price on EverXCloud · from $0.33 per GPU-hour
Suited workloads
NVIDIA GeForce RTX 3090 24GB
Ampere
- GPU memory
- 24 GB
- Memory bandwidth
- 936 GB/s
Live price on EverXCloud · from $0.22 per GPU-hour
Suited workloads
Key differences
- Both have 24 GB of GPU memory.
- The RTX 4090 has 1.1× the memory bandwidth of the RTX 3090 (1,008 vs 936 GB/s), which usually sets the pace of LLM token generation.
Specifications
| Specification | RTX 4090 | RTX 3090 |
|---|---|---|
| GPU memory | 24 GB | 24 GB |
| Memory type | GDDR6X | GDDR6X |
| Memory bandwidth | 1,008 GB/s (higher) | 936 GB/s |
| Max power | 450 W | 350 W |
| Architecture | Ada Lovelace | Ampere |
| Interconnect | PCIe 4.0 | NVLink bridge · PCIe 4.0 |
The higher value of each measured row is shown in bold. Manufacturer specifications, per GPU. Tensor throughput is shown without sparsity. Values the manufacturer does not publish for this exact model are left out.
What fits in memory, on one GPU
| Workload | RTX 4090 | RTX 3090 |
|---|---|---|
| LLM inference, 16-bit | ≈ 9B parameters | ≈ 9B parameters |
| LLM inference, 4-bit | ≈ 30B parameters | ≈ 30B parameters |
| Full training, mixed precision | ≈ 1B parameters | ≈ 1B parameters |
Cautious estimates counting 80% of memory for model weights; context length, batch size and framework change the real limit. Several GPUs together hold larger models.
RTX 4090 vs RTX 3090: frequently asked questions
Which has more memory, the RTX 4090 or the RTX 3090?
Both have 24 GB of GPU memory.
Which is faster for LLM inference, the RTX 4090 or the RTX 3090?
Generating tokens is usually bound by memory bandwidth, and the RTX 4090 has 1.1× the bandwidth of the RTX 3090. Memory size then decides which models fit. Measure your own model before a large run.
Can I rent the RTX 4090 and the RTX 3090 by the hour?
Yes, when sellers list them on the EverXCloud spot exchange: you buy GPU hours at the live price from a prepaid balance in euros, US dollars or pounds. Prices are shown excluding tax and EverXCloud is the merchant of record for every rental.
Be ready on launch day
Create your account now: you can top up your balance and rent the GPU that fits your workload as soon as the market opens.
Have GPU servers?
Businesses can join the EverXCloud Network and sell hours of their GPU servers on the spot exchange.