Thank you for the well filled out report. I see the issue in the trace and will work on a bisect.
Might be related to https://github.com/doitsujin/dxvk/issues/3117
Is unrelated as that regressed all the way back in d9vk 0.20
Broke with the alpha test changes. Last known good commit is https://github.com/doitsujin/dxvk/commit/2c713a34c9088dd726fd4d10cd022ed0be49e2fb and earliest known bad commit is https://github.com/doitsujin/dxvk/commit/85cc87e42a9adaa631593abfd06ec281001eae63
I'm not able to test the commits in between as the trace crashes with them but should be the one of those https://github.com/doitsujin/dxvk/compare/2c713a3...85cc87
Thanks for the trace and bisect. The PR above should fix this.
With the same trace captured with wined3d, texture issues were not observed under dxvk-1.10.3 replay, only on dxvk-2.0 replay
Below are screenshots from frame 3633
Tested on an intel laptop with UHD Graphics 620, AMD RX 580, and a mobile GTX 1050 owned by a friend
Might be related to #3117
dxvk-1.10.3
dxvk-2.0
Software information
S4 League, runs on d3d9
System information
Apitrace file(s)