GPU prices are finally getting normal, and in recent weeks you may have found yourself browsing graphics card reviews to see which tops the charts. After all, the best graphics cards live and die based on gaming performance, right?
These criteria are far from a definitive answer, and in most cases, the conversation stumbles away from the games you actually play and the experiences they offer.
I'm not saying we should get rid of the baby with the bath water. GPU benchmarks are pretty big, and I don't think anything should change about how we (or change) how we perform GPU reviews. However, now that your graphics card is upgradable, it's important to keep all the performance numbers in mind.
Games not standards
Evergreen Computing DTs coordinator made a joke about GPU pricing the other day: I need a new GPU so I can get 9000 frames in Vampire Survivors. Silly, but there is a noticeable difference there. When it comes to performance, it's important to realize that four times as many people are playing TerrariaorStardew Valley than playing Forza Horizon 5orCyberpunk 2077 at any one time.
Because the network is different, these games dominate the number of players, and GPU-limited games are usually over-represented.
The standard games that we and others have chosen as standards are not the answer. It provides a way to push the GPU to its limits in order to compare it to the competition and previous generations. The problem is that the modular suites have a pronounced margin, which could indicate performance that falls outside the graphics card's review.
Standards are often misleading
Benchmarks can be horribly misleading if it comes to the latest graphics cards. Each benchmark requires at least an average frame rate, which is a worrying number either way. Brief frame rate increases are over-represented in the average low 1% and 0.1% low which is the average of the lowest 1% and 0.1% of frames. However, these numbers don't just show how often these frame rate drops occur in any case.
Although the frame timeline can show how often a drop in frame rate has occurred, it is only the theme of the game that the benchmark focused on. I hope you see the trend here: more data points are trying to show real world performance. The benchmarks show relative performance, but they don't say much about playing the game.
The RTX 3090 Ti is 8.5 percent faster than the RTX 3090 inRed Dead Redemption 2, for example. That's right, and it's important to keep an eye on the differences between the cards when playing. Although the RTX 3090 Ti is technically faster, it does not affect game performance in any way.
Another example of F1 2022 is that the game shows significant performance differences between the resolution and all the raised settings (as we saw in our GPU review), but falls back on some GPU-intensive graphics options, and the game is so CPU limited that it offers nearly identical performance between 1080p and 4k. There is no need to upgrade the GPU there.
A strict GPU hierarchy is an abstraction for using your graphics card for what you first bought. Standards are crucial to establishing differences, but they do not explain whether these differences hold true.
How to perform an informed GPU upgrade
Before you upgrade your GPU, you should always look at the benchmarks, as often as you can. However, you should not put your money to the bottom until you have answered these questions.
- What games do I want to play?
- What decision do I want to play with?
- Are there other components I need to upgrade?
- What is my budget?
Relative performance is vital to understanding what you're getting for your money, but it's not entirely better in a PC environment. Depending on the games you play, the resolutions you play, and potential bottlenecks in your system, you might buy a more expensive GPU and get exactly the same performance as a cheaper tool.
There's a lot to worry about buying something cute, even if it doesn't offer a huge advantage. If you have the means, there is no shame in owning something as powerful as the RTX 3090 even if you only use it to play Vampire Survivors. However, don't expect to notice any difference as you play.
This article is part of our ongoing bi-weekly ReSpec column that includes a comprehensive online discussion, tips, and report.
- A GPU used for mining can save you a lot of money, but should you buy one?
- Nvidia is trying to save itself from the collapse of cryptocurrency, but it may not be able to do so
- Major Intels Arc GPUs may be nearby
- Sony's new InZone gaming headphones raise the volume for PS5
- Sony's first gaming monitor costs under $1,000 and it's all on HDR