RTX 4090 vs A100 PCIe
Manufacturer specifications of the RTX 4090 and the A100 PCIe, 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 A100 PCIe 80GB
Ampere
- GPU memory
- 80 GB
- Memory bandwidth
- 1,935 GB/s
Live price on EverXCloud · from $1.79 per GPU-hour
Suited workloads
Key differences
- The A100 PCIe has 3.3× the GPU memory of the RTX 4090 (80 GB vs 24 GB).
- The A100 PCIe has 1.9× the memory bandwidth of the RTX 4090 (1,935 vs 1,008 GB/s), which usually sets the pace of LLM token generation.
- The RTX 4090 has ray-tracing cores for 3D rendering; the A100 PCIe does not.
Specifications
| Specification | RTX 4090 | A100 PCIe |
|---|---|---|
| GPU memory | 24 GB | 80 GB (higher) |
| Memory type | GDDR6X | HBM2e |
| Memory bandwidth | 1,008 GB/s | 1,935 GB/s (higher) |
| FP16 tensor | Not published | 312 TFLOPS |
| BF16 tensor | Not published | 312 TFLOPS |
| Max power | 450 W | 300 W |
| Architecture | Ada Lovelace | Ampere |
| Interconnect | PCIe 4.0 | NVLink bridge (600 GB/s) · 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 | A100 PCIe |
|---|---|---|
| LLM inference, 16-bit | ≈ 9B parameters | ≈ 30B parameters |
| LLM inference, 4-bit | ≈ 30B parameters | ≈ 100B parameters |
| Full training, mixed precision | ≈ 1B parameters | ≈ 4B 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 A100 PCIe: frequently asked questions
Which has more memory, the RTX 4090 or the A100 PCIe?
The A100 PCIe has 80 GB of GPU memory, against 24 GB for the RTX 4090.
Which is faster for LLM inference, the RTX 4090 or the A100 PCIe?
Generating tokens is usually bound by memory bandwidth, and the A100 PCIe has 1.9× the bandwidth of the RTX 4090. Memory size then decides which models fit. Measure your own model before a large run.
Can I rent the RTX 4090 and the A100 PCIe 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.