I have a RTX5090 and I installed CUDA 12.8 ToolKit but RipX says GPU Hardware Acceleration is not available. I uninstalled it and installed CUDA 11.0 ToolKit, but it gives an error when it starts processing, I found the following error via the log file:
2025-08-04 09:23:38,319: Traceback (most recent call last):
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\lib\runpy.py", line 193, in _run_module_as_main
2025-08-04 09:23:38,319: "__main__", mod_spec)
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\lib\runpy.py", line 85, in _run_code
2025-08-04 09:23:38,319: exec(code, run_globals)
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\demucs\separate.py", line 280, in <module>
2025-08-04 09:23:38,319: main()
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\demucs\separate.py", line 138, in main
2025-08-04 09:23:38,319: from dora.log import fatal
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\dora\__init__.py", line 66, in <module>
2025-08-04 09:23:38,319: from .explore import Explorer, Launcher
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\dora\explore.py", line 27, in <module>
2025-08-04 09:23:38,319: from .shep import Shepherd, Sheep
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\dora\shep.py", line 25, in <module>
2025-08-04 09:23:38,319: from .distrib import get_distrib_spec
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\RipScriptLib\dora\distrib.py", line 12, in <module>
2025-08-04 09:23:38,319: import torch
2025-08-04 09:23:38,319: File "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\CUDA\torch\__init__.py", line 137, in <module>
2025-08-04 09:23:38,319: raise err
2025-08-04 09:23:38,319: OSError: [WinError 126] The specified module cannot be found. Error loading "C:\Program Files\Hit'n'Mix\RipX DAW\Resources\Python\CUDA\torch\lib\backend_with_compiler.dll" or one of its dependencies.
I tried manually updating the RipX built-in environment, but found that the built-in Python version is 3.7.2, torch-2.7.0+cu128 minimum required Python version is 3.9.I hope the built-in Python can be updated and support NVIDIA Blackwell GPUs, thanks!
I’m afraid RipX DAW’s GPU hardware acceleration is not currently supported with the 5000 series of Nvidia Geforce GPUs because the minimum version of CUDA they support is 12.8. This has been added to our list to fix in a future update.
Any idea of when we will get a update. My old GPU died, now I have a blackwell GPU and now it takes close to a hour to rip a track. The whole system is high end so its not the rest of the HW.
We've just released version 8 of RipX DAW which supports the 5000 series of Nvidia Geforce GPUs. We recommend uninstalling any earlier versions of RipX before installing this one.
Hi, I’m using RipX DAW Pro 8.0.3 on Windows with an RTX5070Ti (CPU: Ryzen 7 7800X3D). This is a clean install — I’ve never installed any earlier versions of RipX on this PC.
A test track takes around 3 minutes to rip on my M4 Mac mini, and about 2:45 on this Windows PC. This is a bit disappointing, considering the result on Windows is achieved using a GPU that costs significantly more than an M4 Mac mini.
RipX shows “Ripper GPU Hardware Acceleration On (CUDA)”, but during ripping Windows Task Manager shows:
Ripper GPU usage stays around 0.1–0.3%
RipX DAW GPU usage fluctuates around 2–7% (about 2.4% when RipX is idle)
I also tried installing CUDA Toolkit 13.1.1 and 12.8.0 (restarting RipX each time), but there was no difference in ripping time.
Is such low GPU usage — and performance only slightly faster than the M4 Mac mini — expected for the RTX5070Ti?
@337w76 The GPU hardware acceleration is only used during the instrument separation (before Ripper opens). There's a limit to how much of a GPU the machine learnt separation can take advantage of. The Apple M chips are extraordinarily powerful for their size and cost.
