> For the complete documentation index, see [llms.txt](https://docs.viesus.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.viesus.com/operations/benchmarks.md).

# Benchmarks

VIESUS is designed to scale across hardware tiers and production volumes. The figures below are baselines under controlled conditions — your real-world results vary with hardware, image content, and the quality vs. speed settings you choose.

{% hint style="info" %}
These benchmarks are baselines, not guarantees. Always run a representative sample of your own production images on your target hardware before sizing infrastructure.
{% endhint %}

***

## Throughput by preset

All configurations are benchmarked on the **same reference system**. The **Standard** test set represents average customer images — a realistic spread of sizes and content. With this content, roughly **10% of images trigger 2× AI upscaling, 20% trigger 4×, and 20% trigger Artifact Removal**, and **vScene runs on every image**. **Background Removal** is measured on a **separate** set where 100% of images have their background removed.

Each configuration maps to a ready-made preset — see the [Presets Gallery](/configuration/presets-gallery.md).

{% hint style="warning" %}
The figures below are placeholders (**TBD**) pending a full test run.
{% endhint %}

| Configuration              | Image set      | Avg time / image | Throughput (images / hour) |
| -------------------------- | -------------- | ---------------: | -------------------------: |
| Default (base enhancement) | Standard       |              TBD |                        TBD |
| vScene on                  | Standard       |              TBD |                        TBD |
| 2× AI upscaling            | Standard       |              TBD |                        TBD |
| 4× AI upscaling            | Standard       |              TBD |                        TBD |
| Artifact Removal           | Standard       |              TBD |                        TBD |
| Background Removal         | Background set |              TBD |                        TBD |

Reference system: **TBD**.

***

## What affects performance

| Category          | Factors                                                                                                                                                                                      |
| ----------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Image**         | Source resolution (cost grows with pixel count), content complexity (many faces, dense textures), compression level (heavily compressed JPEGs may trigger AI Artifact Removal as a pre-step) |
| **Configuration** | Upscaling factor, quality mode (Fast / Standard / High Quality), AI upscaling intensity, Facial Reconstruction settings                                                                      |
| **Hardware**      | GPU model and VRAM (insufficient VRAM causes paging and slowdowns), CPU, storage (SSD I/O for high-batch throughput)                                                                         |
| **Workflow**      | Number of parallel instances (scales until the GPU saturates), batch size (amortises model load), surrounding pipeline steps (download, archive, upload)                                     |

For how to tune these levers, see [Performance Tuning](/operations/performance-tuning.md).

***

## Sizing your deployment

A practical sizing approach:

{% stepper %}
{% step %}
**Pick a representative sample** of your actual production images (1,000–5,000 is plenty).
{% endstep %}

{% step %}
**Run them on the candidate hardware** with your target configuration (`viesusini.json`).
{% endstep %}

{% step %}
**Measure end-to-end time** including I/O, not just enhancement time.
{% endstep %}

{% step %}
**Multiply by your daily volume** to estimate capacity and add 30–50% headroom.
{% endstep %}
{% endstepper %}

For hardware recommendations and the full set of tuning levers, see [Performance Tuning](/operations/performance-tuning.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.viesus.com/operations/benchmarks.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.