{"id":1486125,"date":"2026-02-04T16:30:52","date_gmt":"2026-02-04T16:30:52","guid":{"rendered":"http:\/\/yPGaEcptUCpKhPCX2FGAYX"},"modified":"2026-02-04T16:30:52","modified_gmt":"2026-02-04T16:30:52","slug":"my-testing-shows-that-16-gb-of-system-memory-is-still-absolutely-fine-for-todays-pc-games-but-there-are-some-caveats-to-it-all","status":"publish","type":"post","link":"https:\/\/arcader.org\/news\/my-testing-shows-that-16-gb-of-system-memory-is-still-absolutely-fine-for-todays-pc-games-but-there-are-some-caveats-to-it-all\/","title":{"rendered":"My testing shows that 16 GB of system memory is still absolutely fine for today’s PC games but there are some caveats to it all"},"content":{"rendered":"
\n

With 32 GB DDR5 memory kits now all sporting colossal price tags, it might seem that doing a full upgrade of your gaming PC’s CPU, motherboard, and memory is out of the question. That’s because 16 GB of DRAM isn’t enough, right? Especially if you have a high-end graphics card or you play games with super-heavy graphics.<\/p>\n

A few weeks ago, I spent some time browsing through various forums and social media posts that are dedicated to asking questions about building or upgrading a PC, and almost all of the comments were adamant that nobody should be going with anything less than 32 GB of system memory in 2026.<\/p>\n

That kind of advice was fine six months ago, when DDR5 was genuinely very cheap, but now that a 16 GB kit of DDR5-6000 CL36 costs $200 at Newegg<\/a>, compared to $370 for 32 GB<\/a> of the same memory, it surely doesn’t make sense to spend the extra $170, when that can be used to get a better graphics card.<\/p>\n

Even though I regularly use a couple of test PCs with 16 GB of DRAM and have a PS5 with the same amount, I wanted to see how much of a difference 32 GB makes when you’re playing some of the most memory-hungry games around.<\/p>\n

To begin with, I checked the memory usage of 27 big games no older than four years, with the graphics settings maxed out at 4K, but with DLSS Performance enabled to avoid a frame rate slideshow, on a gaming PC with 64 GB of DRAM.<\/p>\n

From there, I selected eight to focus on: three that were using around 16 GB of memory, four that were above 16 GB but below 32 GB, and one that used well over 32 GB. Then it was a case of measuring the frame rates over five minutes of gameplay, with the test PC’s memory limited to 16 and then 32 GB, to see what would happen.<\/p>\n

High-end GPU results<\/span><\/h2>\n
\n
\n
Image 1 of 4<\/span><\/p>\n
\n
\n
\n

\"A<\/p>\n<\/div>\n<\/div>

Cyberpunk 2077 <\/span>(Image credit: CD Projekt)<\/span><\/figcaption><\/figure>\n<\/div>\n
Image 2 of 4<\/span><\/p>\n
\n
\n
\n

\"A<\/p>\n<\/div>\n<\/div>

Kingdom Come: Deliverance 2 <\/span>(Image credit: Deep Silver)<\/span><\/figcaption><\/figure>\n<\/div>\n
Image 3 of 4<\/span><\/p>\n
\n
\n
\n

\"A<\/p>\n<\/div>\n<\/div>

Black Myth: Wukong <\/span>(Image credit: GameScience)<\/span><\/figcaption><\/figure>\n<\/div>\n
Image 4 of 4<\/span><\/p>\n
\n
\n
\n

\"Fighting<\/p>\n<\/div>\n<\/div>

Stalker 2 <\/span>(Image credit: GSC Game Worlds)<\/span><\/figcaption><\/figure>\n<\/div>\n<\/div>\n<\/div>\n

I actually did all the tests twice: first with an RTX 5090<\/a> (which has 32 GB of VRAM) and then again with an RTX 5070<\/a> (12 GB of VRAM). The reason for this was to see what impact the video memory capacity had in relation to the amount of system memory, when the game is munching through all the DRAM you can throw at it.<\/p>\n

As well as recording the average and 1% low frame rates, I also monitored the average amount of system memory used during the five-minute runs and how much DRAM the game process itself was using. You can view the results for the latter by clicking on the ‘next’ arrow in the chart below.<\/p>\n

\n