Educational Guide
Learn How Proof Workloads Work
Understand proof generation, compute intensity, operator tiers, and workload planning so you can deploy on Fermah with confidence.
Core Concepts
Learn the terminology behind proving workloads and infrastructure planning.
What is Proof Generation?
Zero-knowledge proof generation is the process of computing cryptographic proofs that verify a statement without revealing the underlying data. Fermah helps developers estimate how demanding this proving workload will be before deployment.
What is Batch Size?
Batch size represents how many proofs or operations are processed together. Larger batches improve throughput but increase proving time and compute requirements.
What is Complexity?
Complexity measures how difficult a proof is to generate on a scale from 1 to 10. Simple arithmetic circuits may be low complexity, while ZKML inference or advanced cryptographic workloads are much more intensive.
What is Compute Intensity?
Compute intensity estimates how much proving capacity a workload consumes. Higher percentages indicate larger resource requirements and stronger infrastructure needs.
What is Operator Tier?
Operator tiers suggest the infrastructure level required to handle the workload. Starter is suitable for testing, Pro for production systems, and Enterprise for highly demanding workloads.
Estimated Cost
Estimated cost provides directional budgeting guidance for proof generation workloads, helping teams plan spending before committing to production deployments.
Real-World Examples
Example workloads and their expected infrastructure recommendations.
Small Rollup Test
Chain: zkSync
Proof Type: Rollup
Batch Size: 1,000
Complexity: 3/10
Starter Tier • ~$4 • Low Compute
Production Rollup
Chain: Scroll
Proof Type: Rollup
Batch Size: 5,000
Complexity: 7/10
Pro Tier • ~$25 • Medium Compute
Large ZKML Inference
Chain: Scroll
Proof Type: ZKML
Batch Size: 10,000
Complexity: 10/10
Enterprise Tier • ~$120 • Extreme Compute
Best Practices
• Start with smaller batch sizes when testing.
• Use comparison mode to evaluate different chains and proof types.
• Monitor complexity carefully for ZKML and advanced workloads.
• Share results with your team before committing infrastructure.
• Use cost estimates for budgeting and deployment planning.
• Upgrade operator tiers before workloads become bottlenecks.
Ready to Estimate Your Workload?
Use Fermah Estimate to calculate proving time, cost, compute intensity, and recommended infrastructure before deployment.