Model fine-tuning
Budget 80 GB across weights, training state and working buffers.
A100 SXM 80 GB rental
Ampere in the SXM form factor for training, inference and scientific computing.
$1.432/ hour
Workloads for this GPU configuration
Inference with memory headroom
Measure the context length and concurrency your runtime can support.
CUDA and HPC
Run reproducible compute experiments on Ampere with HBM2e.
Software compatibility
- The GPU has 80 GB of HBM2e. Include runtime memory and working buffers in the total.
- This is the SXM version of A100. Validate results instead of assuming PCIe measurements apply.
- For distributed runs, consider the separate four-A100 configuration.
Rent A100 SXM 80 GB
Provide your model, software environment and rental period. We will confirm the rate, availability and connection method before the rental starts.
$1.432/ hour
About this configuration
How much memory does A100 SXM 80 GB have?
The GPU has 80 GB of HBM2e. Include runtime memory and working buffers in the total.
One A100 80 GB or four?
Start with one if the workload fits and its runtime is acceptable. Four GPUs support distributed or independent jobs, but the application must explicitly use multiple devices.
How do I request this configuration?
This configuration is open for rental requests. Share your workload, preferred dates and duration. Hardware availability, pricing and access dates are confirmed individually; a request does not create a confirmed booking.
