Benchmarks vs. System Metrics - When and Why to Use Them

Optimize your simulation performance and costs using two complementary data sources: benchmarks for upfront machine selection and system metrics for post-simulation analysis and refinement.

Benchmarks help you choose the right machine before you run a full simulation. They’re a quick test that shows how your code performs across different machine types — so you can pick the best fit for speed and cost efficiency without wasting credits on trial and error. Always run a benchmark when you’re setting up a new project, trying new software, or testing different workloads.

System Metrics help you understand how your chosen machine was actually used after the task completes. By looking at CPU, memory, and disk stats, you can check if you over- or under-provisioned resources. This helps you tweak future runs — maybe you can save money with a smaller machine, or speed things up with more cores or RAM.

In short:
✅ Use benchmarks upfront to pick smartly.
✅ Use system metrics afterward to refine and optimize.

Both features give you clear, actionable data about how your workloads use compute resources — so you can choose the right machine, avoid waste, cut unnecessary costs, and get faster results without guesswork. In short: they help you optimize your spend and performance, every time you run a simulation.