Member Avatar for SillyBilly

I have tried to use one benchmarking utility, "Mark3 or 5" and I was realy confused. Reading a few post, I am still at loss. I would like to bench mark my cpu and quake2 game while playing so I can tweak for more performance. I see post where guys have done this but no more info other than scores. Any tips for a lost soul needing some advice?

  • 2 Contributors
  • 4 Replies
  • 102 Views
  • 1 Day Discussion Span
  • Latest Post Latest Post by SillyBilly

Dani AI

Generated

A short, dummy-proof workflow that builds on @OillyComputers' demo idea and gives repeatable, usable numbers for both Quake II and the CPU.

  1. Prepare the system: update GPU drivers, set Windows power profile to High Performance, close unnecessary background apps, and disable V-Sync/frame limiters for raw measurements. Record GPU/CPU driver version and OS build before starting.

  2. Measure interactive gameplay (easy, fast): use a simple overlay tool (FRAPS or MSI Afterburner/RivaTuner OSD) to show live FPS and log telemetry. Pick a 30–60 second play sequence that can be repeated (same map area or same actions). Run that sequence 3 times and note min/avg/max FPS and any stutters.

  3. Measure repeatable performance (consistent comparison): use Quake II demos/timedemo playback or record a short in-game demo and run it back for consistent results. Timed demos remove human variability so changes (driver, resolution, settings) show up clearly.

  4. Isolate CPU behavior: run a CPU benchmark or stress test (SuperPi/wPrime/Cinebench/Prime95) separately to verify single-thread and multi-thread performance, and to check for throttling. While gaming, monitor CPU clocks, core usage, and temps with a hardware monitor (HWINFO/MSI Afterburner).

  5. Tweak one variable at a time: change only one setting per test (resolution, texture detail, anti-aliasing, CPU clock/multiplier). Keep a simple log: test ID, date, drivers, single change, results (fps/temp/clock).

Troubleshooting notes: if FPS does not change after CPU tweaks, the GPU is likely the bottleneck; if GPU usage is low but one or more CPU cores are at 100%, the CPU is limiting. If CPU temps climb and clocks drop, watch for thermal throttling and undo overclocks. Inconsistent runs usually stem from background processes or differing player actions — use demos for repeatability.

This approach produces clear, comparable numbers and helps separate whether tweaks affect the CPU, GPU, or overall system behavior.

Recommended Answers

All 4 Replies

Member Avatar for OlyComputers

There are two kinds of benchmarks in use today:

Synthetic benchmarks run a series of tasks and score how well they perform.

Real-world benchmarks monitor frame rate and other performance indicators during actual use of the system (usually game-play) and give you the informations (FPS) or a score.

There is a third kind that I'll call "theoretical" benchmarks, for lack of a better term. These aren't really benchmarks in the traditional sense but they examine you're hardware and estimate how it would perform based on it's specifications.

Real world benchmarks are the most accurate but synthetics are better rounded. Certain games will work different aspects of a system more, so you may perform great in a game that leans on memory or CPU, and not so well in a game that leans on the video card more. Or good on a game that that has a lot of pixel shadowing and bad on one that requires a lot of physics effects. Because of this a synthetic benchmark that tests all these things can give you a weighed score, as well as detailed information about individual tasks.

If you have specific questions about benchmarks in general, ask them here. If you have questions about a specific benchmark you might want to check out their sites FAQ's page first.

Member Avatar for SillyBilly

Thanks for the info,,, Do you have any suggestions on a bench mark program while playing quake2?

Member Avatar for OlyComputers

I don't do a lot of benchmarks, and when I do they're always synthetic. The thing about quake benchmarks (as I understand it) is that it usually doesn't actually monitor frame rate in YOUR game, but rather uses a "Demo" feature to run through a level and then gives you the results. This is because of the differences in people style. Somebody could get a huge score for sitting still, while somebody who's blowin S**t up with grenages will get slugish frame rates.

Try google "how to run a quake demo" and you should find the info.

PS: I believe some games will keep track of your FPS during gameplay, or that there are third party apps to do it, but I've never tried.

Member Avatar for SillyBilly

Wow, interesting, but still do not know which to use, being first time and all, I would love it to be dumby proof! lol... Thanks again, an interesting subject.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.