Renato Golin

At Linaro, we have an HPC lab where we can setup multiple clusters and machines, install OpenHPC, run the test-suite, benchmarks using Mr-Provisioner, Jenkins and Ansible automation. This allows us to test new kernels/initrd, new versions of OpenHPC, compilers, applications, etc. on different architecture at the same time, from scratch. The work we did is open source, on GitHub, and is intended to be used by other people. We started with hard-coded rules but are slowly moving towards generic configuration-based setup from the ground up, so that anyone else can clone, change the configuration files and run the Ansible playbooks. From scratch.

These are the repositories that we contribute to / use:

Mr-Provisioner, the bare metal provisioner that installs OSs on all our servers (Arm or Intel): https://github.com/Linaro/mr-provisioner

Our infrastructure scripts, setting up Jenkins and running the jobs logic: https://github.com/Linaro/hpclabjenkins

Our benchmark harness, which runs benchmarks in a repeatable way: https://github.com/Linaro/benchmark_harness

Our OpenHPC Ansible scripts, which install OpenHPC on bare-metal/VM, clusters (Arm or Intel): https://github.com/Linaro/ansible-playbook-for-ohpc

Auto-Vectorization has come a long way since the early vector-processing CPUs, and compilers generally take a long time to implement it, prioritizing other more generic features instead. But with all recent high-end chips containing some form of SIMD operations, auto-vectorization became a necessary feature on any modern compiler. LLVM was perhaps the last of the big compilers to have a decent vectorization engine, but it has grown considerably for the last year, and the investment on SIMD code generation will not diminish. This presentation outlines the past implementations, what we currently have available and peeks into the engineering pipeline to see what else we are working on.