Firebird, Swarthmore’s HPC cluster, saw unprecedented use over the summer, largely in support of faculty research and the Summer Scholars Program. The cluster accelerated research in many disciplines, including Physics & Astronomy, Economics, Engineering, Biology, Mathematics, and more. As the following image shows, during the month of July, around 80% of Firebird’s CPU cores were consistently in use:

Most of Firebird’s 15 GPUs were likewise in heavy use:

In addition, many users took advantage of Firebird’s revamped and expanded web interface, offering virtual desktops, a file manager, terminal access, Jupyter notebooks, and more:

We are excited to say that we recently added two additional nodes to Firebird: a GPU node offering two NVIDIA H200 double-precision GPUs, and an additional high-memory node offering 2 TB of memory to support larger memory-intensive tasks. During fall, we will also be expanding further with another GPU node, offering two Blackwell RTX6000 Server Edition GPUs, which are purpose-built for AI tasks, and which will increase our total number of GPUs to 19.
We encourage anyone interested in using Firebird, either for research or classroom use, to reach out to Jason Simms (jsimms1@swarthmore.edu).