Getting consistent gaming PC benchmark results has long been a problem I’ve grappled with when looking to upgrade or test graphics cards. Many synthetic benchmarks can be unreliable or unrelatable, while benchmarking real games can be tricky and inconsistent, and both can lack granularity. Thankfully, though, there’s a new benchmarking tool that fixes all these problems, and it’s called Evolve.
Evolve is a new GPU benchmarking app that’s packed with tests for the latest ray tracing and path tracing techniques, as well as traditional rasterization (that’s game rendering without ray tracing) and compute testing too. It will consistently test the performance of your system and deliver relatable, easy-to-read charts that cover a wide range of performance metrics, including scores for rasterization, ray tracing, and power usage. Dig deeper into its results, and you can even analyze frame times too.
To test just how useful this new benchmarking tool can be, I tested three GPUs that might represent a typical upgrade path for a gamer. Starting with an Nvidia GeForce RTX 4070, I ran the software, then tested an AMD Radeon RX 9070 and an Nvidia GeForce RTX 5070, to see how these two possible upgrades compared to the original card.
Setting up the software is easy, as it can be downloaded from Epic Games Store, Google Play, Microsoft Store, and Steam (via this link). With a download size of just 2.84GB, it doesn’t take long to install, and once you’re up and running, there’s blessedly little setup or tinkering to do, other than setting up an account. Just jump into the “Run Benchmark” mode and fire off a benchmark run – you can set up resolution and other settings, but the default settings are fine for most runs.

The app offers five different benchmark runs, each of which focuses on a slightly different type of performance, with the main difference being between running a ray-traced test or a path-traced test. I concentrated on the first ray tracing test, as path tracing is still quite uncommon even in very modern games.
Hitting the benchmark button, you’re greeted to a fantastic-looking dinosaur-filled game world, through which the benchmark works its way, following its unnamed human character. After a couple of minutes, the benchmark run is done, and you’re greeted with a simple score-based results page.

On this page, we can see that the software has correctly noted that I’m running an RTX 4070 and an AMD Ryzen 5 7500X3D CPU, with the full test results – accessible by clicking the “View Your Result” button – also showing the motherboard, memory, and Windows version I was using.

The results page provides scores for Rasterization, Raytracing, Compute, as well as power consumption (Energy) and Driver and Acceleration Structure Build scores. You can also then view graphs for frame time and power over time, using the navigation just above the scores.

With a suite of test runs recorded, I then swapped over to the AMD Radeon RX 9070 and the RTX 5070 to see how these two cards performed. The results are shown below.
As you can see, in just one graph, I’m able to compare the overall rasterization, raytracing, and compute performance, and power consumption, of three GPUs. No longer do I need to run multiple games, turning ray tracing on or off across multiple runs.

Crucially, from this comparison, it’s really easy to see that the RTX 5070 offers a solid upgrade over the RTX 4070, but that the RX 9070 is far faster in rasterization and compute but slower in ray tracing, while the RTX 5070 offers a smaller rasterization upgrade but a larger ray tracing upgrade. So, if you’re into games that don’t use ray tracing – and that’s still most of them – the RX 9070 would be the ideal upgrade, while the RTX 5070 is better suited to cutting-edge ray-traced titles.
To take advantage of this incredibly easy GPU benchmark – whether you’re a gamer, reviewer, hardware vendor, or more, you can download it via this link to the company’s Steam page.

