This article is for providing information about how to reserve a virtual machine with virtual GPUs created from the limited number of H100 GPUs in the lab. If you do not specifically need to use the H100 GPU, you can send a general request for computing servers or workstations through the following post. We have plenty of CPU only computing servers and tens of RTX 3070 GPU workstations. If you are a student who wants to use the H100 virtual GPU, before you can make a reservation, you need to send a request using the following post in order to get an approval from your advisor. Please select "Remote Access to Computing Servers" and specifically request H100 virtual GPU in your Request Details.
Through the support from Office of the Provost and School of Business, we had the chance to test and set up servers with H100 GPUs when it was first released. In many cases in research with financial applications, we see significant speed increases when GPU is used. However, not all projects can fully use the 80 GB of GPU RAM in a H100 GPU and not many research groups will demand running large-scale GPU jobs everyday. To let more people use the expensive GPU, we deployed the GPU virtualization software from NVIDIA (https://www.nvidia.com/en-us/data-center/virtual-solutions/). However, virtual machines still cannot dynamically share the GPU RAM. Therefore, we have to statically assign equal shares of the physical GPU RAM to each virtual machine. For example, we can have two virtual machines with each having 40 GB of GPU RAM, but we cannot have one virtual machine with 15 GB of GPU RAM and another one with 65 GB of GPU RAM. If one user requested a 40 GB virtual GPU for tomorrow, there will be only one 40 GB virtual GPU available for others to reserve for the same time period.
Here are the types of virtual machines that we can provide from the host servers with H100 GPUs:
Type A: 96 CPU cores, 1000 GB CPU RAM, 80 GB GPU RAM
Type B: 48 CPU cores, 500 GB CPU RAM, 40 GB GPU RAM
Type C: 24 CPU cores, 250 GB CPU RAM, 20 GB GPU RAM
Type D: 12 CPU cores, 125 GB CPU RAM, 10 GB GPU RAM
Type Z: 4 CPU cores, 8 GB CPU RAM, No GPU
Type A, B, C and D are types that require reservations to use. When a user is not reserving any of those types, their virtual machine will be converted to Type Z. Since Type Z virtual machines are dynamically sharing CPU time and CPU RAM from the host, we can allow a large number of users to keep their Type Z virtual machines running or idling. This is very important for most users because it takes time to work on coding and setting up of an experiment before the code is ready to be tested on large GPUs. When the virtual machine needs to be converted, one reboot will be enough. There will be no changes or very little changes needed from the user in their code. The user will be assigned to a specific host and migrations of the user and the virtual machine can be requested through emails to firstname.lastname@example.org.
Each host with one H100 GPU can host one Type A, two Type B, four Type C and eight Type D at any given time.
Here are the rules for all virtual GPU users of the Hanlon Labs.
- If this is your first H100 virtual GPU request, you must obtain approval: Students must receive approval from their advisor using the form located at https://fsc.stevens.edu/how-to-request-computing/. Faculty or staff can email us at email@example.com. We will email you for setting up access.
- Before submitting a reservation request using the form below, carefully check the H100 GPU reservation calendar below for available numbers and types of virtual machines.
- Reservations need to be made in advance based on the type: Type A requires 14 days notice, Type B needs 7 days, and Type C and D require just 24 hours notice. If you have an urgent need and there are virtual GPUs available on the calendar, please email us at firstname.lastname@example.org to make a reservation.
- Reservations cannot start more than 21 days from today.
- Reservation duration is limited to 72 hours except for Type Z. You can make reservations for Type Z without time limitations.
- Each user can only reserve one virtual machine for any time period. Please email us for special cases where more than one virtual machine is needed for one user.
- If your chosen period overlaps with an existing reservation (Type Z not included), you must select a same type with the already reserved virtual machine at that overlapping time.
- Check your email for acceptance notifications; only accepted requests are guaranteed.
For any questions, you can contact us at email@example.com or visit Hanlon Financial Systems Lab for assistance.