Labs
Research, measurements, and evidence from FARPY.
What FARPY tested, what the evidence showed, and what you can reuse. FARPY Labs is the public research shelf. Visitors come here to read findings, methods, and evidence from GPU tests FARPY actually ran. It is not the place to upload a .blend or track a customer render — that is Your renders.
How to read Labs
Research holds studies and notes. Graduated holds capabilities that became public tools. Benchmark evidence lives on the GPU result pages. There is no live public activity feed on this page.
- ResearchResearch archive
Current experiments and measured findings.
- GraduatedGraduated capabilities
Work that has moved into product or stable methodology.
- Verified GPUsVerified GPU evidence
Published GPU evidence that can be inspected independently.
- Submit benchmarkHow evidence is published
Contribute a bounded benchmark result. Public evidence is reviewed on the GPU pages.
Published records
Each record states What we learned, the Evidence, and What you can reuse. Open the page for methods and negative findings.
- Blender 4.1.1 on RTX 3090View study
What qualified, what was blocked, and where the claim stops.
- RTX 3090 Blender evidenceOpen record
One frozen, public RTX 3090 Blender result you can inspect.
- Blender GPU runtime reproducibilityView study
When a frozen Blender workload stays comparable, and when runtime variance means it does not.