Uncertain about the best approach, should you attempt to increase performance beyond normal limits?
Uncertain about the best approach, should you attempt to increase performance beyond normal limits?
I will edit my signature after this to reflect my new hardware, but it is: corsair cx650m pwr msi b660 pro-a intel i512400 (w/ coolermaster hyper212black) g.skillz ripjawsV 32G (two separate kits of 16g, so four sticks) running at 3200 xmp enabled wdblack 1tb sn270 Nvidia rtx 3060 xc 12gb aerocool strike-x case, mostly the panels are just grates, so air moves fairly freely, I have a fan at the bottom, a fan at the back, and the cooler fan, but it's much more just open-air as the top is just a grate and the side panel is like a grate. I won't go water cooler, but I can later buy a new case and get mega-fans, or I do have this old cooler-master case the size of a small fridge I can build it in, it's from the phenom-2 era, and just buy a bunch of fans. I'm trying to play Starmade in a specific way. Starmade is a procedurally generated universe game that works like minecraft in many ways, with chunks and all. I am getting low framerates, but only when in a very busy area, with thousands of blocks rendered in,,, but that's my need. I am trying to basically fly around a city that I uploaded from a blueprint (players can share blueprints of stations and ships and you can use them by simply importing the blueprints) and it's a mega-build. If you are just in one place, it's fine, but since the objective is to "drone race" around the city, I constantly move from one side of it to the other, and my framerate drops a lot when it needs to render in blocks. It helps to increase the "chunks visible" which technically just makes it so more "segments" can be drawn at once (segments are chunks or something, , anyway you can think of it like a block count), but I still struggle. I have the game at medium graphics (with some customizations to eliminate various features) and in a windowed 800*600 resolution, and the struggle remains. The game is very much an indie project, but critically is procedurally generated and all of the objects are made of individual blocks as in minecraft, so basically I don't know where the slowdown is. Is it my CPU? is it my Graphics? The game doesn't have any support for modern tech with optimizations of any kind, but it's really an essential game for my repertoire, and the objective is to "space drone race" which is basically niche, and there's not many games that let you have a simular simulation. (6dof newtonian thrust physics drone racing). here's a video taken on my old machine. I dont' know how to overclock, and I don't know what to overclock, and I don't know if I really can achieve anything with it, since so much of the clock is automatic,, and I"m not sure extra cooling will help either, since I don't think I'm really going outside of the smooth operating temps anyway. Component upgrades are a little bit out of the question,, it's not worth it for a few videos and otherwise the computer is pretty good. fyi I do my 6dof thrust games on this homemade controller: https://hackaday.io/project/180884-4-joy...controller
You can determine if the issue is a CPU or GPU bottleneck by using monitoring tools (such as MSI Afterburner) to observe usage. Based on the performance, it seems likely the CPU is the limiting factor, especially since it runs smoothly at low resolution and is a high-end build. The game appears not yet optimized for heavy simulations, and you have a powerful modern CPU—overclocking isn’t feasible. It’s possible the system is hitting its CPU limits, not the GPU. You’re likely seeing around 4 to 4.4 GHz during gameplay.
I had the task manager open and it was up close to 4. 50-60% utilization but after tuning the settings that's down a lot. It just seems to be that when it needs to load blocks in it's just too much and there's no feasible way to really minimize that any more than I did. But I want to take video and in my experience if I record gameplay that slows it down even more so I wanted to talk about it all. that would be sweet if the game could somehow be more optimized, but it's basically an old project and no one is working on it, it's in it's final form basically. There's always an imaginable future that it might get popular but it's more likely that a new project does it some time in the future, when this one would run smoothly on the 15th or higher gen cpu's anyway. It's always believable that mobile processors have some kind of innovation that makes super-core-count take over. I did see somewhere that you can change the number of threads assigned,, that would mean the i7 might make a difference, but I don't see myself upspending when I just bought the one I have, would running an i7 12700k on the coolermaster hyper 212 make sense?
Can you verify each core separately and confirm if any reaches 100%? I suspect the overall CPU usage for a gaming rig falls between 50-60%, but many titles struggle due to a single bottlenecked core. I think the game likely runs single-threaded, processing all blocks with one core. I’d want to explore how the game handles multiple cores. If it scales well, the 12700K offers significant gains, though I doubt it delivers much beyond that. In reality, you’d probably need something like a Mugen 5 for casual gaming and NH-D15 for more demanding work.
Acquire the 12700k right away. You're ready to proceed. Keep sending questions about it—do it now.