GPU guide
GPU for university research in Quebec: rent a GPU by the hour
Updated on
Short answer
A university lab, a CEGEP or a school service centre can rent an NVIDIA GPU in Montreal by the hour, from CA$0.67/h for an RTX A6000 with 48 GB, paid from prepaid credit in Canadian dollars. It complements the national computing clusters, for a deadline, a class or a prototype.
When to rent rather than wait for an allocation
The clusters of the Digital Research Alliance of Canada remain the right choice for large computations covered by an allocation. Renting by the hour fits the other cases: a conference submission next week, a student who has no account yet, a test before applying for an allocation, or a class that needs GPUs at a set time.
On a rented server, you are the administrator: you install the libraries you want, with no modules and no scheduler, and the GPU is shared with nobody.
For a CEGEP or a school service centre, a server with JupyterLab lets a class learn machine learning without buying graphics cards or maintaining them.
A budget that does not overflow
- Credit is prepaid, by credit card, from CA$25.00: spending follows the credit you load, and auto-reload stays optional.
- For an organization, payment by bank transfer is possible on request, to be confirmed with our team before the order.
- Invoices and a monthly usage statement download as PDF in the console, in Canadian dollars.
- The price shown is the price billed in Canadian dollars, taxes extra: no exchange rate to explain in an expense report.
The GPUs for JupyterLab, with their availability
| GPU | GPUs per server | Server GPU memory | On demand, per hour | Availability |
|---|---|---|---|---|
| NVIDIA RTX A6000 | 1 | 48 GB | $0.67 | |
| NVIDIA L40 | 1 | 48 GB | $1.34 | |
| NVIDIA A100 PCIe | 1 | 80 GB | $1.81 | |
| NVIDIA H100 PCIe | 1 | 80 GB | $3.35 |
Prices in Canadian dollars, taxes extra, read from the catalog when the page is built. Availability is read live; a machine out of stock shows the closest equivalent machine in stock.
The JupyterLab one-click template installs JupyterLab with the GPU, served over HTTPS and protected by a token. For a thesis that needs large models, servers with 2 to 8 GPUs are described in the rent an 8-GPU server guide.
Cost examples for a class and a thesis
- A class lab: 25 students × 3 hours on an RTX A6000, that is 75 hours, costs $50.25.
- A thesis project: 40 hours of training on an A100 PCIe with 80 GB cost $72.40.
- Formula: hourly rate × hours, rounded to the cent, taxes extra. Delete or hibernate the servers between sessions.
Research data and ethics
If your protocol states where the data may be hosted, the servers and their disks stay in Canada, in Montreal, in a single region. The details are on the data sovereignty page; for health data, see also the GPUs and health data guide.
Good to know before you launch
- Billing is hourly, at the server’s hourly rate, from prepaid credit. There is no per-minute billing, and the first hour is charged when you deploy.
- A stopped server is still billed at the full hourly rate. To pay less, hibernate it ($0.02 per hour (about $14.60 per month), disk kept) or delete it. A Spot machine cannot be hibernated: delete it when you are done.
- The minimum top-up is CA$25.00, paid by credit card.
- Servers and their disks stay in Canada (region canada-montreal, in Montreal). Account information (identity, billing, e-mails) may be processed outside Quebec, as the privacy policy explains.
Frequently asked questions
How much does a GPU for a research project cost?
From CA$0.67/h for an RTX A6000 with 48 GB, and CA$1.81/h for an A100 PCIe with 80 GB, taxes extra, billed by the hour.
Can a university pay by bank transfer?
Yes, on request for an organization: the bank transfer is confirmed with our team before the order. Payment by credit card is open to everyone.
Can I set up GPUs for a class?
Yes. Each student or team can have its own server with JupyterLab installed in one click; delete the servers after the session to stop billing.
Does research data stay in Canada?
The servers and their disks stay in Canada, in Montreal. Account information may be processed outside Quebec, as the privacy policy explains.
Does it replace the Alliance clusters?
No, it complements them: the national clusters remain the right choice for large computations covered by an allocation. Renting by the hour fits when you need a GPU outside an allocation.
Need GPUs? We’ve got you.
Launch an NVIDIA GPU server by the hour in Canada, paid in Canadian dollars by credit card, or reserve a sold-out GPU: it launches as soon as one frees up.
Other guides
- Your AI says you need a GPU? Run your Python script on a GPU in minutes
- CUDA out of memory: causes, fixes, and when to move to a GPU with more VRAM
- torch.cuda.is_available() returns False: causes and fixes (Mac, PC without NVIDIA, Docker)
- How much does a GPU cost per hour in Canada, in Canadian dollars?
- OpenAI API too expensive? Host your LLM on a GPU by the hour, in Canada
- Rent a GPU by the hour in Canada: prices, steps and billing
- GPU cloud in Montreal, Canada: NVIDIA GPU servers hosted in the country
- Rent an H100 in Canada: PCIe, NVLink or SXM, by the hour
- Which GPU to fine-tune an LLM? Memory, machine and cost
- Cheap GPU cloud: how to pay as little as possible for a GPU
- Ollama on a cloud GPU: run an LLM online on your own server
- GPU for Stable Diffusion and ComfyUI: generate images on a rented server
- Rent an 8-GPU server: 8x H100 or 8x A100 NVLink, by the hour
- RunPod alternative in Canada: GPUs by the hour, in Canadian dollars
- Health data and AI: GPUs hosted in Canada