Andrew Fagan

Render Graph Problem Analysis

What am I working to fix?

Important: Don’t solve problems I don’t have. Trying to predict future problems and trying to solve them just makes more problems and complicates things.

Context:

Why am I solving this problem?

Identify and Balance Constraints:

How much research do I have to do on this?

What is the performance budget?

What is the latency/overhead requirements?

Will this be handled differently in different builds (Debug vs Release vs Dist)?

Value:

Identify specific, concrete, measurable value.

How much time is this problem worth?

Cost:

What is the cost of solving the problem?

Platform:

How will this handle the platform targets?

How will this interact with shared resources (RAM, SSD)?

Data:

What is read and written over time? What is the access pattern?

Possible Future Features:

  1. Adding support for buffers as a render pass resource
  2. Adding support for texture aliasing