Biggus

Seems to be working well for me. I'm only really using the Standards though. The rest of the weapons I haven't really played with too much. WP seems to be working, so there's a mix of red and white smoke at the impact sites.

My tracks are too long to share, but in MP with a cold and dark spawn I'm still seeing it every sortie.

I can't help you with the sketch but you should be aware that the source you used for DCS BIOS has been dead for years at this point and you should be using the DCS-Skunkworks fork which can be found here: https://github.com/DCS-Skunkworks/dcs-bios You'll need to remove all your current DCS BIOS stuff including that library and then install the new stuff.

I hope one day to see an update to FSR.

Seconding that it's absolutely not a VRAM limitation issue. Was a problem for me at 8GB, has been a problem for me at 16GB. It's been an issue reported by people with 4090s in the past. It's also somewhat infuriating that it isn't consistent. With my 3070, I'd get it more often than I do with the 9070XT. It's now on roughly 30% of the instant actions I load into, vs about 80% prior to the upgrade. Sometimes people get lucky with masking the issue by dropping clutter, forest details and textures, but it's definitely not a normal performance issue due to limited VRAM and is absolutely a bug. It's a pity because it's a lovely map otherwise, I just haven't been able to use the damn thing in what, two years?

Not using VR. Every one of those performance settings checked and tested multiple times in different combinations.

I will add my hardware to my last post, although based upon seeing users with 4090s experiencing the same thing, I'm doubtful that it's a GPU issue.

Here's a quick demo with my current setup. FPS limiter set to 65fps, but the effect is still the same. Behavior is identical with all modules. I'm using some mods in this track to change the trees from high to low and I'm also using reduced texture sizes and don't really have the patience to return to a completely stock install for this test, because the result is the same. SA_frame_drop.trk Edit: My rig is a 5800X on a B450M PRO VDH Max board with 64GB at 3600, a 9070XT and DCS is installed on an M.2 drive. The issue also existed with a 3070. Current texture mods include resizing everything above 4096x4096 and replacing the high trees with low tree models. These mods do not seem to mitigate it. If I find time, or end up being sufficiently annoyed by this bug enough to make time when I can, I'll return to a standard setup and post more videos and tracks, but it feels utterly pointless because I suspect that it's something that only hits a tiny percentage of the user base.

It was still happening for me in the last patch. I haven't tested since then, I'll try and grab a track soon. I don't believe it's very widespread at all and would be surprised if it could be replicated given that people reporting it used a pretty wide range of hardware.

Probably not terribly useful information if it's fixed internally, but I'm getting it on cold start missions too. Scared the heck out of me the first time.

Sounds like a problem with your STM32 to me if it's doing it even without being on a PCB.

I don't think it's a specific file or texture as it will compress everything with the original parameters. I'll see if I can find a specific parameter that is triggering the error though.

Thank you for getting back to me. After checking with a script comparison tool, the only differences I can see are the ones that are cut and paste from the original one to the modified one you shared a couple of pages ago and the CHANGE-ME value to 2048. The lines change because the lines are changed in your example I quoted yesterday. I'm quite stumped.

After having used the original script for awhile, I've been really happy with it. I was hoping to exclude the terrains and cockpits using this set of parameters but I've been getting an error when I attempt to use what's in the quote above. Any idea of how to fix it? dcs-texture-optimizer.ps1

Just adding because I'm seeing some odd behavior when enabling AFMF and Super Resolution in the AMD software. It takes a big hit in frames from 900ft down, just like the 3070 did. Running vanilla MSAAx4, there's a tiny drop but it's barely noticeable. I'll experiment with Optiscaler too over the next few days just for the hell of it.