Question 1

How does Go compare to .NET Core in the Skynet benchmark?

Accepted Answer

On a Macbook Core M, Go runs in about 979 ms while .NET Core takes 650 ms, but on an i7 with TPL, .NET can be as fast as 118 ms. Results depend heavily on hardware and specific implementations like async vs. promises.

Question 2

How to run the Skynet benchmark for Rust?

Accepted Answer

Navigate to the rust-coroutine directory, use 'cargo build --release' and 'cargo run --release'. This requires Rust and the coroutine-rs library, as noted in the README for the coroutine-based implementation.

Question 3

Why are Haskell results so fast on Ubuntu but slower on other systems?

Accepted Answer

The README shows Haskell at 41-44 ms on Ubuntu with an i7, but 6181 ms on a Macbook Core M. This discrepancy highlights GHC optimizations, runtime settings like +RTS -N, and hardware differences, emphasizing the benchmark's sensitivity.

Question 4

Can Skynet benchmark Java virtual threads?

Accepted Answer

Current implementations include standard Java and Quasar fibers (marked TODO), but not Java virtual threads (Project Loom). You would need to adapt the code or find other benchmarks for that specific concurrency model.

Question 5

What hardware was used for the Skynet benchmark results?

Accepted Answer

The README specifies tests on a Macbook 12" '2015 with Core M, an i7-4770 on Win8.1, and an i7-4771 on Ubuntu 15.10. These varied setups mean results are not standardized, requiring caution in cross-language comparisons.

Question 6

How accurate is Skynet for real-world concurrency scenarios?

Accepted Answer

Skynet is a synthetic microbenchmark useful for measuring raw concurrency overhead, but it doesn't simulate real-world factors like I/O bottlenecks or complex state management, so it should be supplemented with application-specific tests.

skynet

What is skynet?

Overview

Use Cases

Best For

Related Projects

Found a gem we're missing?

Not Ideal For

Pros & Cons

Pros

Cons

Frequently Asked Questions