# Superluminal vs iRender

> Superluminal reached local disk sooner in all four scenes. iRender required boot, setup, rendering, and file sync. Only Classroom's upload activation predates its recording.

- Canonical HTML: https://superlumin.al/compare/superluminal-vs-irender
- Markdown endpoint: https://superlumin.al/compare/superluminal-vs-irender.md
- Structured page data: https://superlumin.al/compare/superluminal-vs-irender.json
- Published: 2026-08-04
- Last updated: 2026-08-07
- Author: Superluminal Engineering

## Quick answer

The test used one rented iRender workstation with four RTX 4090 GPUs. Its cost covers the complete paid machine window.

## Data access

- [This page as structured JSON](https://superlumin.al/compare/superluminal-vs-irender.json)

- Every chart below includes a full-resolution PNG and its underlying values as a Markdown table.

- Every screenshot below is embedded with its alt text, caption, and full-resolution URL.

## Related documentation

- [Inspect a render job](https://superlumin.al/docs/web-ui/job-details): Use the resizable frame grid, preview, details, scene settings, charts, machines, and logs on a job page.

## Complete report

Superluminal recorded the shorter render pipeline in all four scenes. Its 10-node run on the Fox scene led iRender by 14.1 seconds.

![Superluminal and iRender selected on the 2026 render farm score scatterplot](https://superlumin.al/seo/charts/superluminal-vs-irender-ranking-2026.png "Superluminal and iRender in the benchmark ranking")

*[Open the full-resolution PNG](https://superlumin.al/seo/charts/superluminal-vs-irender-ranking-2026.png).*

#### Underlying data: Superluminal and iRender in the benchmark ranking

Scores are frozen against the 2026 benchmark. 100 is the reference on both axes.

| Initial state | Rank | Render farm | Time Score | Price Score | Combined score | Coverage |
| --- | --- | --- | --- | --- | --- | --- |
| selected | 1 | Superluminal | 179.1 | 187.8 | 182.0 | all 4 scenes |
| muted | 2 | GarageFarm | 138.2 | 102.2 | 126.2 | all 4 scenes |
| muted | 3 | RenderJuice | 93.7 | 176.2 | 121.2 | all 4 scenes |
| muted | 4 | RebusFarm | 102.0 | 59.9 | 88.0 | all 4 scenes |
| muted | 5 | RenderStreet | 68.8 | 121.7 | 86.5 | all 4 scenes |
| muted | 6 | Fox Renderfarm | 67.5 | 115.3 | 83.5 | all 4 scenes |
| selected | 7 | iRender | 51.4 | 66.4 | 56.4 | all 4 scenes |
| muted | 8 | Render Network | 49.7 | 37.7 | 45.7 | 1 of 4 scenes with both metrics; no coverage penalty |
| muted | 9 | Blendergrid | 22.9 | 80.6 | 42.1 | 3 of 4 scenes with both metrics; no coverage penalty |

[Open the complete benchmark dataset](https://superlumin.al/seo/evidence/render-farm-benchmarks-2026-08.json)

iRender ranked seventh with an Overall Score of 56.4. Its Time Score was 51.4, and its Price Score was 66.4. Three paid windows are exact. Classroom is an upper bound because the transfer client was hidden immediately before the first full completion proof.

## A rented workstation, not a managed job

We rented one GPU Server 5S Windows workstation with four RTX 4090 GPUs. We booted the machine, opened Blender, started each render, and synchronized the output.

Renting is work. The machine took 733.7 seconds from start to the Connect button becoming ready, boot time iRender did not bill. From there you operate Windows over a remote session: open Blender, press render, watch it, and copy the output to iRender Drive yourself. During retrieval the client showed misleading transient states, at one point reporting 39 of 50 files at 0.0 Mbps, and needed a manual refresh before the counts moved.

Superluminal is the managed version of the same hardware wish. You [submit from Blender](/interface), the farm runs the machines, and finished frames arrive on your disk while later frames render. Nothing boots on your clock, and nothing synchronizes by hand.

iRender charged 30 iR per machine-hour on its pay-as-you-go rate. The account required a $50 deposit.

iRender's own meter starts when the Connect button becomes ready, not when the machine activates. The Classroom window starts there. We removed the 29.9 seconds spent extracting the portable Blender archive from the scored time, and the cost keeps those seconds because billing continued. The later scenes reused the running workstation.

## The paid window by scene

| Scene | Active render | Render pipeline | Paid workstation window | Rate-derived cost | File synchronization |
|---|---:|---:|---:|---:|---:|
| ![](/seo/benchmarks/thumb-classroom.jpg) Classroom | 418.5 s | 434.8 s | ≤1,225.7 s | ≤10.46 iR (about $10.46) | ≤503.9 s |
| ![](/seo/benchmarks/thumb-junkshop.jpg) Junk Shop | 154.2 s | 159.7 s | 1,033.0 s | 8.61 iR (about $8.61) | 686.9 s |
| ![](/seo/benchmarks/thumb-monster.jpg) Monster Under the Bed | 154.7 s | 159.3 s | 617.1 s | 5.14 iR (about $5.14) | 387.6 s |
| ![](/seo/benchmarks/thumb-fox.jpg) Fox (EEVEE) | 74.5 s | 80.6 s | 736.7 s | 6.14 iR (about $6.14) | 556.5 s |

The render-pipeline column covers provider setup, loading, rendering, and finalization. It excludes upload, download, and operator time. iRender is the machine-rental exception in the Time Score: its paid workstation window includes billed operator work and file synchronization. Classroom starts at Connect-ready and removes only the one-time Blender extraction.

The cost estimates apply the verified 30 iR hourly rate to each paid window, and each main-balance iR had a $1 face value. The windows overlap while files synchronize, so their costs do not add up to one four-scene total. No final invoice was captured.

The billing model is the deeper difference. iRender's meter runs on the machine, so your own clicking, waiting, and file copying all bill at the machine rate. Superluminal's meter runs on frames: each finished frame bills its measured render seconds, and nothing you do at the keyboard costs anything.

The recordings verify the local frame count. They do not prove that Blender used all four GPUs.

![Superluminal submission terminal confirming the queued job after packaging and upload](https://superlumin.al/seo/interfaces/superluminal-submission-terminal.png)

*Submission ends in the add-on's terminal: the job is packaged, uploaded, and queued before any browser opens.*

![Superluminal Classroom job starting active rendering on 30 nodes](https://superlumin.al/seo/interfaces/superluminal-live-progress-classroom.webp)

*Active rendering started after the upload and queue phases in the same recording.*

![Superluminal Classroom job after all 145 frames completed](https://superlumin.al/seo/interfaces/superluminal-completed-classroom.webp)

*Provider completion showed all 145 Classroom frames finished.*

![Superluminal Classroom job after all 145 frames reached local disk](https://superlumin.al/seo/interfaces/superluminal-downloaded-classroom.webp)

*The final boundary confirms the complete local frame set 106.7 seconds after upload began.*

![iRender remote workstation hardware configuration](https://superlumin.al/seo/interfaces/irender-machine-configuration.webp)

*The operator chooses a remote workstation instead of a queued job.*

![iRender remote machine start dialog](https://superlumin.al/seo/interfaces/irender-machine-start-dialog.webp)

*The rented Windows machine must boot before Blender opens.*

![Blender rendering Classroom through iRender remote desktop](https://superlumin.al/seo/interfaces/irender-remote-desktop-render-start.webp)

*The operator starts Blender and the render through remote desktop.*

![iRender Drive synchronizing the 145-file Classroom output folder](https://superlumin.al/seo/interfaces/irender-drive-classroom-synchronizing.webp)

*After remote rendering, iRender Drive shows the Classroom output synchronization in progress.*

iRender Drive required a manual synchronization step, and in every tested scene that step took longer than the render itself. The [Classroom](/seo/evidence/runs/irender-gpu-server-5s-classroom-2026-08-02.json) and [Fox (EEVEE)](/seo/evidence/runs/irender-gpu-server-5s-fox-eevee-2026-08-02.json) records show each synchronization step and completion proof.

## Sources

| # | Source | Type | Checked | URL |
| --- | --- | --- | --- | --- |
| 1 | Render farm benchmark dataset, July to August 2026 | First-party benchmark evidence | 2026-08-04 | https://superlumin.al/seo/evidence/render-farm-benchmarks-2026-08.json |