Troubleshooting `install-quartus.sh` Missing `fiftyfivenm_atoms.vhd` In GitHub Actions

Hey everyone,

I'm currently facing an interesting challenge while trying to set up a GitHub Action workflow. The goal is to integrate NVC (which, by the way, is awesome!) with the Quartus simulation libraries. Basically, I want to get my CI (Continuous Integration) environment to play nicely with my hardware simulation setup.

Setting the Stage: Cloning and Copying Libraries

So, to kick things off, I decided to recreate the Quartus library environment inside my CI container. How? Well, I cloned the asicguy/gplgpu repository. This repo conveniently includes a copy of the Quartus simulation libraries tucked away under the hdl/sim_lib/ directory. Think of it as a treasure chest of simulation goodies! I then expose this path using the QUARTUS_ROOTDIR environment variable, essentially telling NVC where to find these libraries.

The Plot Thickens: The nvc --install quartus Command

Now comes the crucial part. I try to install the Quartus libraries using the following command:

nvc --install quartus

This is where things get a bit dicey. The script, in its current form, seems to expect a specific file named fiftyfivenm_atoms.vhd. Unfortunately, this file is playing hard to get and refuses to show up in my copy of the libraries. Bummer!

** Fatal: opening /opt/intelFPGA/20.1/quartus/eda/sim_lib/fiftyfivenm_atoms.vhd: No such file or directory

It's like going to a party and realizing you're missing the key ingredient for the perfect cocktail. Locally, I've noticed a similar trend. I don't have the fiftyfivenm variant either. Instead, I've got libraries for some of the newer process nodes, like 20nm, maxv, and others. It's like the hardware world is evolving faster than my script can keep up!

Diving into the Workflow Snippet

To give you a clearer picture, here's a snippet of my GitHub Actions workflow:

- uses: nickg/setup-nvc@v1
  with:
    version: latest
- run: |
    git clone https://github.com/nselvara/gplgpu
    cd gplgpu
    mkdir -p /opt/intelFPGA/20.1/quartus/eda/
    cp -r ./hdl/sim_lib/ /opt/intelFPGA/20.1/quartus/eda/

    ls /opt/intelFPGA/20.1/quartus/eda/sim_lib/

    nvc --version
    nvc --install xpm_vhdl
    nvc --install quartus
  env:
    QUARTUS_ROOTDIR: "/opt/intelFPGA/20.1/quartus"

Let's break this down a bit:

1. I'm using the nickg/setup-nvc@v1 action to set up NVC. Pretty standard stuff.
2. I clone the gplgpu repository, which, as we discussed, contains the Quartus simulation libraries.
3. I then do a bit of file juggling, creating directories and copying the library files to the expected location (/opt/intelFPGA/20.1/quartus/eda/).
4. I even throw in an ls command to verify that the files are indeed where I expect them to be. Always good to double-check!
5. I run nvc --version to make sure NVC is up and running. Gotta love those version checks!
6. I install xpm_vhdl using nvc --install xpm_vhdl. This seems to work just fine.
7. And then comes the moment of truth: nvc --install quartus. This, as we know, is where the script throws a tantrum about the missing fiftyfivenm_atoms.vhd.
8. Finally, I set the QUARTUS_ROOTDIR environment variable, pointing NVC to the Quartus installation directory.

The Million-Dollar Questions

This whole adventure has left me with a few burning questions, and I'm hoping the NVC wizards out there can shed some light on them:

Is there a supported way to override or skip missing device families during --install quartus?

This is probably my biggest question. Is there a secret incantation, a magic flag, or a hidden setting that I can use to tell the script, "Hey, it's cool if you can't find fiftyfivenm. Just skip it and move on to the libraries that are actually available"? It would be incredibly helpful to have a way to customize the installation process to match the specific libraries I have on hand.

Could the install script be updated to detect which libraries are available instead of assuming a specific one like fiftyfivenm?

This feels like a more robust, long-term solution. Instead of hardcoding a dependency on fiftyfivenm, the script could potentially scan the directory for available libraries and install them accordingly. This would make the script much more adaptable to different Quartus versions and configurations.

Would you recommend an alternative strategy for CI integration without relying on unavailable libraries?

I'm always open to new ideas and approaches. If there's a completely different way to tackle this problem, I'm all ears. Maybe there's a better way to set up the Quartus environment in CI, or perhaps there's a different way to integrate NVC with the simulation libraries.

A Huge Thank You in Advance!

I want to express my sincere gratitude to the NVC community for their hard work and dedication. NVC is an amazing tool, and I'm really excited about its potential. Any insights or suggestions you can offer would be greatly appreciated. Let's conquer this fiftyfivenm mystery together!

Keywords and SEO Optimization

This article focuses on a specific issue encountered while using the install-quartus.sh script with NVC in a GitHub Actions environment. The core problem revolves around the missing fiftyfivenm_atoms.vhd file and the need for a more flexible way to install Quartus libraries. The keywords we're targeting include:

• NVC
• Quartus
• GitHub Actions
• install-quartus.sh
• fiftyfivenm_atoms.vhd
• Simulation Libraries
• CI Integration
• VHDL

By incorporating these keywords naturally throughout the article, we aim to improve its visibility in search engine results and help other users facing similar challenges find a solution.

Understanding the Core Issue: The Missing fiftyfivenm_atoms.vhd

At the heart of this issue is the missing fiftyfivenm_atoms.vhd file. This file is part of the Quartus simulation libraries and is expected by the install-quartus.sh script. However, in many modern Quartus installations, this specific library might not be present, especially if users are working with newer process nodes. This discrepancy between the script's expectations and the actual available libraries leads to the installation failure.

The Importance of Flexibility in Installation

This situation highlights the importance of flexibility in the installation process. A script that rigidly expects a specific file or library can easily break when faced with different environments or Quartus versions. A more robust approach would involve the script dynamically detecting the available libraries and adapting its installation process accordingly. This would ensure that the script works seamlessly across a wider range of setups.

CI Integration Challenges and Solutions

CI (Continuous Integration) environments present unique challenges when it comes to setting up hardware simulation tools. The environment needs to be reproducible and consistent across different runs. This often involves setting up the necessary libraries and tools in an automated fashion. The issue with install-quartus.sh underscores the need for robust and adaptable installation scripts in CI environments. Potential solutions include:

• Overriding missing device families: A mechanism to explicitly tell the script to ignore missing libraries.
• Dynamic library detection: The script automatically identifies and installs available libraries.
• Alternative CI integration strategies: Exploring different approaches to setting up the Quartus environment in CI.

Exploring Potential Solutions and Workarounds

Let's dive deeper into the potential solutions and workarounds for this issue. Understanding the nuances of each approach will help us choose the best strategy for our CI integration.

Overriding Missing Device Families: A Short-Term Fix

One immediate solution is to find a way to override or skip the missing device families. This would involve modifying the install-quartus.sh script or providing a configuration option that allows us to specify which libraries to install. While this might be a quick fix, it's not the most elegant solution in the long run. It requires manual intervention and might not be sustainable as Quartus versions evolve.

Dynamic Library Detection: A More Robust Approach

A more robust approach is to update the install-quartus.sh script to dynamically detect which libraries are available. This would involve the script scanning the Quartus installation directory and identifying the present libraries. The script could then adjust its installation process accordingly, installing only the available libraries and skipping the missing ones. This approach would make the script much more adaptable to different Quartus versions and configurations.

Alternative CI Integration Strategies: Thinking Outside the Box

Sometimes, the best solution is to think outside the box and explore alternative CI integration strategies. This might involve using pre-built Docker images that contain the necessary Quartus libraries, or setting up a dedicated CI environment with a specific Quartus version. These approaches can offer more control over the CI environment but might also require more setup and maintenance.

Conclusion: Towards a More Flexible Installation Process

The issue with the missing fiftyfivenm_atoms.vhd file highlights the need for a more flexible and adaptable installation process for Quartus simulation libraries. The current install-quartus.sh script, while useful, can be brittle when faced with different environments and Quartus versions. By implementing solutions like overriding missing device families, dynamic library detection, or exploring alternative CI integration strategies, we can create a more robust and reliable CI environment for hardware simulation. This will ultimately lead to faster development cycles and fewer integration headaches. So, let's work together to make the Quartus installation process smoother and more user-friendly for everyone!

Key Takeaways

• The install-quartus.sh script may fail due to missing libraries like fiftyfivenm_atoms.vhd.
• A more flexible installation process is needed to handle different Quartus versions and configurations.
• Potential solutions include overriding missing device families, dynamic library detection, and alternative CI integration strategies.
• A robust CI environment is crucial for efficient hardware development.
• Community collaboration is key to solving these challenges and improving the NVC ecosystem.

I hope this article helps you understand the issue and explore potential solutions. Feel free to share your thoughts and experiences in the comments below. Let's learn and grow together!