You wish ... It depends a LOT on the programming.
For Jedi Survivor I had to increase the page file size on my primary SSD in order for the game to run properly (I use Windows 11 and have 32 GB RAM).

The disk cache usually only affects files that have to be written, not the ones you read. That too depends on the software you are using. A good example are picture viewers. Even in Windows 11 times it makes a huge difference if you activate preload or not and cycle back and forth through your pictures. With preload deactivated, it should be about the same speed if the files really got cached in RAM, but it's not.

Ok, sounds like Windows is utter garbage if that is the case. I have 64GB of RAM and Linux is aggressively caching anything I read. Right now 40 gigabytes of cached data in RAM. I can easily grep large repos without ever hitting the disk, for example.

If you wanted to avoid caching, that is what requires programming. You can open with O_DIRECT, and thankfully almost no software does.

Turns out you can run it on a pretty modest supercomputer

Damn, my Ryzen 5 5600X is outdated now, huh...?
I wish system requirements would stop going up, I mean games look pretty enough and are big enough already.
What's all that new required power for?

I will play this later I guess...
I didn't expect to run it at Ultrawide or anything, just though I would still be meeting recommended specs... Depressing, since new CPU means new motherboard.
Maybe in 2025...?

I've got an i7-11700k, and that's not able to run this game at recommended settings? That's a mighty bitter pill to swallow.

There's only really a handful of games that have such high requirements. Very few games have full ray tracing, and most of the newer stuff on GOG is D3D11

I've never had a top of the line rig, but I suspect it would feel a lot like being all dressed up with nowhere to go

not really bothered about hdd drop as i had not used one for gaming since 2016. and people who still game on hdd are weird! so optimizing the game loads for sata ssd is perfectly acceptable in this decade.

the CPU requirement is interesting. 12700/7800* are top of the line CPUs so for ultra or 1440p&144fps gaming I can see them being the recommended. but 1080p 60fps? thats... worrisome. Ultra with all pedestrians, traffic maxed out? heck. demand threadripper, but high settings (based on the base game) needing such a powerful cpu? This will cause trouble as most of us are still on much older cpus:
https://store.steampowered.com/hwsurvey/processormfg/

*I presume its 7700x instead.

Well, data can be cached in the memory of course, but JS was programmed differently, for consoles which don't have much RAM. Try to run it on Linux with only a 4 GB page file, see what it does. By now however they might have changed it, I played the original release version. Still it would be interesting.

What I'm saying is that the caching is done by the kernel, entirely transparently and automagically. It's not something you program in. Every program benefits from it..

Now as I said, there is a way to program it OUT if you really wanted to.. but the consensus is that it is almost never the right thing to do, and some systems go as far as using syscall filters or patching the kernel to disable the O_DIRECT flag altogether so stupid programmers can't ruin performance with it.

Someone who really knows what they are doing *can* benefit from disabling caching, e.g. if they are going to preload exactly what they need ahead of time (and they know they won't be short on RAM) or if they are going to load something that will only ever be read once (it may not be useful to spend cache on such a thing and possibly push other, more useful things out of cache).

However, in any case, disabling kernel caching should not create a situation where increasing page file size makes the program run better. So if that helps, there must be something else up... such as the program simply needing more RAM than you have.

What happens when you're low on RAM but have a pagefile is that the OS pushes least recently accessed pages into swap, freeing up RAM for more frequently accessed pages. This will redcue stutter and improve performance.

This would be a prime example of the program benefiting from caching, and if instead of adding more page file, you added that much RAM, the end result should be just as good (or better).

I just looked into it, and everything I've found suggests that the game is simply a RAM hog. E.g. here: https://forums.guru3d.com/threads/importance-of-pagefile-for-gaming-smoothness.447687/

"it's not the Page file that is reducing traversal stutters, it's the increased use of system RAM by Windows. Windows will not tap into all your RAM unless a page file is present because it doesn't want to run out of memory."

So the game benefits from entirely normal caching done by the OS, but if you don't have enough RAM, you need a page file to free up RAM (or to trick Windows' stupid algorithms to utilize all of it, assuming it is severely underutilized otherwise). I have not used a page file or a swap partition in a decade and no game ever complained .. it is pretty much completely unnecessary on Linux if you're not short on RAM. Anyway, I would try that game if it were on GOG, but I doubt it would behave different from any other game.

Linux will happily utilize more or less 100% of RAM and you don't need a page file to trick it to do that. Btw, a common complaint from Windows users when they first tried Linux was that it uses all your RAM.. that's a feature! Unused RAM is a wasted resource.

Not quite true, or better: not the full story. Programs can control that directly. It's possible to completely deactivate the page file in Windows, but some programs require one to work, but I haven't seen that behaviour in almost 20 years, I htink the last one was Outlast or at least a game from that time. It's usually about 32 Bit games that can only use a limited amount of RAM.

Since JS can also run with 8 GB, 32 should be more than enough. If it fills up 32 and still has problems, then there's something fishy going on.

Ok, if that is really the case then my initial assessment holds: Windows is utter garbage. And people are writting utter garbage for it :P

Btw, can you name which calls programs use to control it directly? I would like to take a look at how these are implemented in Wine.

EDIT: I just found a beautiful comment in Wine's source code. "Titan Quest refuses to run if TotalPageFile <= TotalPhys"

And of course I have run the game with zero swap. What gives? Garbage programming. It doesn't actually need any and will happily run as Wine lies that we have swap even if we don't... It doesn't sound like you understood anything what I said, nor did you read the posts I linked. Time to move on, I guess.

Unfortunately the 7700x is below recommended requirements.

The store page for the DLC has been updated since I posted my screenshot of the fictional 7800X..

Recommended is now 7800X3D.

Ok

Wait a 7800X3D ?
The Original Cyberpunk 2077 doesn't even use all cores of my 5800x without an external fix.......that is prettty insane.
If I'm gonna have 30 -40 Frames again on a 3090 because they can't multicore well and have to use an external fix.....uff.

Well I'm definitely gonna wait a few month after release.

It seems to be one of those games that respond well to the extra L3 cache. I looked up a benchmark, and it seems you get an 18% fps boost (1080p @ ultra)

https://www.tomshardware.com/reviews/amd-ryzen-7-7800x3d-cpu-review/4

Edit: The 7700X is no slouch, though. There's a bit of a bottleneck if you're pairing it with an RTX 4090, but it's still a fast processor