UIforETW VS shadow

You can use UI4ETW[0] to capture an event trace when the mouse freezes. You may need to install the Windows SDK or Visual Studio to interpret the results, though.

[0] https://github.com/google/UIforETW

Event Tracing for Windows (WPA/xperf) works quite well for profiling, and it's really easy to set up with UIforETW. Symbol loading works with MSVC-generated symbols (just set the target/debug/ folder as a symbol path), though it's rather slow.

There's a couple methods to achieve this but I like to recommend this project since it provides an easy to use GUI and (iirc) doesn't require the installation of Windows ADK.

If anyone could use UI for ETW (https://github.com/google/UIforETW/releases) to grab an ETL file during the micro-stutters and share it with me, that'd be amazingly helpful. An ETL is like a whole-machine tracefile that'll show exactly what chunk of code is using the time when a stall happens.

Install UI for ETW (https://github.com/google/UIforETW/releases) and collect an ETL file while the stuttering is happening. Upload them to somewhere and send it to JeffHill once you got them.

If you capture an ETL file using UI for ETW (https://github.com/google/UIforETW/releases) and link that ETL file in the report, I'm happy to take a look and see what might be going on?

Could you install UI for ETW and collect an ETL file? https://github.com/google/UIforETW/releases

If you install UI for ETW (https://github.com/google/UIforETW/releases) and capture an ETL trace during gameplay, I'd love to open it with dev tools and find out what's going on! It sounds like you've done all the obvious things, so I'm pretty curious what's actually going on with your system. With a rig like that you should have no problems running Dota at a pretty high frame rate.

If you're seeing poor performance in Dota, one thing that's incredibly helpful is to install UI for ETW (https://github.com/google/UIforETW/releases) and collect an ETL file during the problem. We can read these files to see which parts of Dota are misbehaving or what else on your system might be interacting with Dota.

Cool, will be interested to see how this develops! tokio's loom framework has been a big help in testing some tricky concurrency code I've worked on.

Folks interested in this space might also be interested in the system I spend most of my time working on: Shadow. It also performs deterministic simulation of a network of hosts, but it intercepts network and system interactions at the syscall level via seccomp. As such it can work with binaries compiled from ~any language, usually without any code modification or special compilation. https://shadow.github.io/

Nice writeup! I've also had to dig a bit into this area in my work on shadow.

It is no longer active. If you are asking about Shadow, check out https://shadow.github.io

I do share this hesitation. I think for simple cases iterators are usually fine, but I've definitely run into cases where an iterator adapter caused unexpected performance problems. e.g. https://github.com/shadow/shadow/pull/2543

Yes. Though I'm not sure I see the connection to the OP...?

The example I'm most familiar with, because I work on it, is Shadow. We used ptrace for a bit but now use seccomp.

https://github.com/shadow/shadow/

For anyone interested in following current development on Shadow, we've been publishing a series of updates. Most recent: https://github.com/shadow/shadow/discussions/1274

The previous update has links back to the whole series; I stopped including it in the most-recent update since it was getting a bit cumbersome: https://github.com/shadow/shadow/discussions/1060