Frequently Asked Questions

RTX A6000 (48 GB), NVIDIA L40S (48 GB), and NVIDIA DGX A100 slices or full pods (80 GB per GPU, NVLink). Live availability is shown in the deployment panel before you launch.

Yes. All instances run in sovereign data centers in Rabat, ensuring low latency for local users and compliance with national data residency requirements.

No. Every instance ships with the correct NVIDIA drivers, CUDA 11.8/12.1/12.2, and the Docker/NVIDIA Container Toolkit preinstalled. We recommend pulling containers from NVIDIA NGC to keep dependencies consistent.

Hourly, charged at the end of each hour an instance runs. A stopped instance only accrues storage charges at 0.05 MAD/GB per month.

Running instances are suspended automatically, and storage volumes are kept for 7 days so you can top up before anything is deleted.

Personal accounts get instant access to RTX A6000/L40S with standard quotas. Organization accounts require ICE/RC verification but unlock higher quotas, invoicing, and DGX A100 access.

Yes. Bank transfers (minimum 2,000 MAD) are credited within 24–48 business hours, and verified organization accounts can request purchase-order billing via sales@mydgx.ma.

Go to Formations & Certifications on the homepage, pick a course, and submit the registration form. An invoice follows approval, and payment confirms your seat.

Most instances move from "Provisioning" to "Running" within a few minutes. DGX A100 pods can occasionally take longer during high-demand periods since they are shared cluster resources.

Yes. You can start with a personal account and request organization verification at any time from Account Settings — your existing instances and data are preserved during the upgrade.

MyDGX offers Ubuntu-based images preconfigured with NVIDIA drivers, CUDA, and Docker. Custom images can be requested for organization accounts by contacting support.

Storage volumes cannot be resized in place. To increase capacity, create a new larger volume, mount both, and copy your data over before detaching the old one.

This is usually because the SSH host key changed after a rebuild, or a persistent volume was not re-mounted via /etc/fstab. Check the SSH troubleshooting and storage articles in the knowledge base.

Yes, as long as combined VRAM usage stays within your GPU's capacity. Use nvidia-smi to monitor shared memory usage across containers running on the same instance.

Yes. You can launch JupyterLab inside any NGC container (e.g., the PyTorch image) and access it via SSH port forwarding, or use a pre-configured Jupyter image if available in the deployment panel.

Verified organization accounts can request custom reserved-capacity pricing for long-running workloads by contacting sales@mydgx.ma with expected duration and GPU type.

Yes. Invoices for all payments, including card top-ups, are available for download from the Billing tab in your dashboard.

No, prepaid credits do not expire and remain on your account balance until used.

Public universities and verified academic institutions can request discounted DLI course pricing through dli-support@mydgx.ma when submitting a purchase order.

Yes. Each completed DLI course concludes with an official NVIDIA certificate of competency, issued directly by NVIDIA and listed under your profile.