FireSim 1.21.0 Released

[Jim Fang]

Hi All,

We are back to regular FireSim releases! We aim to publish releases every 3-4 months.

FireSim release 1.21.0 was just published. For a list of changes, you can refer to the release notes here: https://github.com/firesim/firesim/releases/tag/1.21.0.

Major changes in this release include:

- AWS F2 support for FireSim
- Prebuilt bitstreams for cores using the Saturn vector engine
- Updates FPGA/local deployment infrastructure

If you have any questions or concerns, feel free to ask here on the mailing list or via GitHub issues, and we will try our best to answer them!

Thanks!

[Connor Sullivan]

Hi Jim,

Thanks for the work on all of this! We have been using on prem FireSim for a while now, but still have a need for AWS support for larger parallel sim runs at times so we are thrilled to see F2 support.

I was running through a fresh AWS FireSim setup for the new F2 support and ran into a snag. I suppose it is more Chipyard related, but Chipyard's build-setup.sh conda dependency lock step seems to be getting stuck indenfinetely for me. I let it sit for a few hours and all I could see is one of the CPUs being pinned at 100%, but no progress was made. I haven't had this issue before. Any ideas? Did I likely screw something up or is this something else? I was following the 1.21.0 FireSim docs and checked out the 1.14.0 Chipyard tag before doing the setup.

Thanks again!

Connor

[Jim Fang]

Hi Connor,

We weren't able to reproduce this issue on a fresh c5.4xlarge running AMI 1.19.2-prod-rhng4b6alkhdq. We ran `./build-setup.sh riscv-tools` which took ~30 mins to complete. There is a chance that our machine launch script may be buggy. Here's a couple things I'd try:

1) Start a fresh instance with machine launch script included, let that finish (or if it hangs, pkill it), then setup Chipyard according to the instructions at https://ucb.bar/cystart (including installing/reinstalling Conda manually via the wget) -- in theory that machine launch script should've installed Conda, but if it has bugs it might lead to a broken install, hence the recommendation to manually install/reinstall.

2) Try setting up Chipyard on this latest release on a non-AWS machine to see if something in this latest release broke something in our Conda flow.

3) Setup Conda on AWS and install something not from the Chipyard conda-reqs yamls to see if something weird is happening with Conda on AWS.

Let me know if you're still having trouble!

Jim