Hello,
Whenever I process an image from Lightroom in Topaz photo AI, I will get a crash and windows (w11) will turn my graphics card (AMD rx570) off. This happens especially (but not always) when returning to Lightroom using upscaling.
I have tried to monitor the incidents with the adrenaline software running and I notice that the VRAM is filled to the rim quickly.
I have also tried to limit the voltage to the GPU, which does not prevent the issue.
My System:
Operating System
Windows 11 Pro 64-bit
CPU
Intel Core i7 14700 44 °C
Raptor Lake 10nm Technology
RAM
32,0GB Unknown @ 2394MHz (40-39-39-77)
Motherboard
Micro-Star International Co., Ltd. PRO B760M-P (MS-7E02) (U3E1) 45 °C
Graphics
LG ULTRAFINE (3840x2160@60Hz)
LG Ultra HD (3840x2160@59Hz)
Intel UHD Graphics 770 (MSI) (not used)
4096MB ATI Radeon RX 570 Series (Gigabyte) 52 °C
CrossFire Disabled
Steps to reproduce issue:
- export an image directly into Topaz (edit in Topaz photo AI)
- Do denoise, sharpening and upscale (20% of the crashes when I fiddle arround to much with the sliders)
- Return to Light room (80% of the crashes)
Topaz Photo AI v3.1.2 on Windows
From the log file:
[2024-07-31 16:20:26.326, 411.65 ms] [359c] Info | [Benchmark] Time to encode: “411.626 ms”
[2024-07-31 16:20:26.326, 28.40 μs] [359c] Info | Cache compression finished.
[2024-07-31 16:20:26.326, 16.50 μs] [359c] Info | Compression info: 10.6MB => 10.3MB Saved 302.1KB
[2024-07-31 16:20:26.326, 785.81 ms] [51cc] Info | Cache insert hash: 18446744071969705326
[2024-07-31 16:20:42.539, 0.00 μs] [28ec] Info | [AIE] ONNX error at 211: D:\a_work\1\s\onnxruntime\core\providers\dml\DmlExecutionProvider\src\ExecutionProvider.cpp(896)\onnxruntime.dll!00007FFDC0B54919: (caller: 00007FFDC0B54D77) Exception(2) tid(28ec) 887A0006 De GPU reageert niet op meer opdrachten, waarschijnlijk vanwege een ongeldige opdracht van de aanroepende toepassing.
[2024-07-31 16:20:42.636, 97.79 ms] [28ec] Info | [AIE] Error debug information
[2024-07-31 16:20:42.636, 26.10 μs] [28ec] Info | [AIE] Image infer_input Size: 128x128x3
[2024-07-31 16:20:42.636, 7.80 μs] [28ec] Info | [AIE] Image infer_output Size: 512x512x3
[2024-07-31 16:20:42.636, 7.10 μs] [28ec] Info | [AIE] Unable to run model with index 1 it had error:
[2024-07-31 16:20:42.637, 79.10 μs] [28ec] Info | [AIE] Image Model inference failed: unable to run model with index 1
[2024-07-31 16:20:42.637, 16.39 s] [4c74] Info | [AIE] Error handling block: unable to run model with index 1
[2024-07-31 16:20:42.637, 879.50 μs] [28ec] Info | [AIE] ONNX error at 211: D:\a_work\1\s\onnxruntime\core\providers\dml\DmlExecutionProvider\src\ExecutionProvider.cpp(896)\onnxruntime.dll!00007FFDC0B54919: (caller: 00007FFDC0B54D77) Exception(4) tid(28ec) 887A0005 Het GPU-apparaatexemplaar is onderbroken. Gebruik GetDeviceRemovedReason om de juiste actie te bepalen.
[2024-07-31 16:20:42.638, 984.70 μs] [28ec] Info | [AIE] Error debug information
[2024-07-31 16:20:42.638, 15.50 μs] [28ec] Info | [AIE] Image netInput Size: 128x128x3
[2024-07-31 16:20:42.638, 9.90 μs] [28ec] Info | [AIE] Param pb: 0,468827
[2024-07-31 16:20:42.638, 10.60 μs] [28ec] Info | [AIE] Param pc: 0,000012
[2024-07-31 16:20:42.638, 5.50 μs] [28ec] Info | [AIE] Param pn: 0,386305
[2024-07-31 16:20:42.638, 5.60 μs] [28ec] Info | [AIE] Unable to run model with index 0 it had error:
[2024-07-31 16:20:42.638, 17.40 μs] [28ec] Info | [AIE] Image Model inference failed: unable to run model with index 0
[2024-07-31 16:20:42.639, 251.70 μs] [28ec] Info | [AIE] Model Backend state is invalidated due to previous errors
[2024-07-31 16:20:42.639, 7.80 μs] [28ec] Info | [AIE] Unable to run model with index 0 it had error:
[2024-
