How to do ultrafast Deep Neural Network inference on FPGAs

Connecting to the cloud

All exercises will be on the Amazon Cloud

• Morning: CPU (t2.2xlarge) Afternoon:  FPGA machines (f1.2xlarge)

Go to this Google Docs spreadsheet

• put your name down next to one of the CPU/FPGA instances.

Download the key FPGA4HEP.pem from the email you have received and store it somewhere on your computer (set the correct permissions: chmod 700)

You can log in with

• CPU exercise:                    ssh -i "FPGA4HEP.pem" centos@CPUpublicIP**
  Also mount the storage: sudo mount /dev/xvdf fpga4hep/
• FPGA exercise:                 ssh -i "FPGA4HEP.pem" centos@FPGApublicIP**