Graphics testing, benchmarking & performance analysis support.

After updating from patch 5.00b, I noticed that HairWorks hair looked worse indoors, with more visible shimmering. I first saw it in 5.00c, and it was still there in 5.01. I wanted to understand what had changed and whether I could recover the earlier appearance.

Comparing the builds helped me narrow the change down to the hair-rendering pipeline. I then tested a hybrid configuration that combined the earlier rendering components with the newer game data. In the scenes I checked, it brought back the appearance I had seen in 5.00b.

Getting a closely matched pair of screenshots felt almost impossible at first. To make these comparisons easier, I developed a mod that lets me run multiple instances of the game, hold Geralt still and save both his position and the camera setup. I can then load that setup in the other instance, matching Geralt's position and the camera angle across builds.

I started GameStutterLab because of the stutters I kept running into while playing PC games. I wanted to understand what was causing them, and to spend less time sorting through data before finding something worth investigating.

The tool lines up frame times with available CPU, GPU, memory and disk activity, so I can see what else was happening around a hitch. The idea is to make that first pass through a capture quicker and help narrow down what to test next.

It's still in development, and I'm checking its suggestions against real gameplay captures.

I wanted to see how much GPU power consumption I could reduce while keeping similar gaming performance. I developed Silicon Tuner and tested it on my RTX 5090 and RTX 5070, comparing stock settings with tuned profiles.

For the RTX 5070 comparison shown here, I used Cyberpunk 2077's built-in benchmark with the same game settings for each profile, recording frame rate, GPU power draw and temperature. In my runs, the Balanced profile averaged 160.3 W against 230.7 W at stock, 30.5% less power, while averaging 50 FPS compared with 49 FPS at stock.

That's one card in one game's built-in benchmark. Results will vary with the GPU, workload and settings, so I treat it as a measured example rather than a general rule.

While playing Resident Evil Requiem, I wanted a closer look at character expressions and details the normal camera left out of view, so I built RE9 FreeCam.

The hardest part was keeping control during cutscenes, when the game kept pulling the camera back to its scripted position. I also added field-of-view controls and options to disable depth of field and lens flare for closer inspection.

I've been adapting RTX Remix to work with The Sims 2, comparing path-traced lighting, shadows, reflections and materials with the original rasterized rendering. An important part of that work is telling expected differences apart from things that are rendering incorrectly.

Along the way, I've investigated distorted characters, missing textures and incorrectly rendered glass: reproducing each issue, comparing it with the original game, then testing possible fixes and checking whether they introduce other problems.