Tonex capture bug

AmpliTube TONEX, AmpliTube 5, AmpliTube Custom Shop, AmpliTube Joe Satriani, AmpliTube Brian May, MESA/Boogie®,Orange™, Fender™, Hendrix™, Metal, AmpliTube SVX, and more for Mac/PC

Tonex capture bug

Postby jedsas » Sat Mar 04, 2023 11:54 am

I have been considering upgrading my computer but have found a worrying computer specific bug in Tonex which is a bit of a showstopper.

I currently have two Dell laptops which are very similar - Precision 3510 and Precision 3520. The 3520 is newer and has a later processor so should be faster. I would like to retire the 3510 which has been my main computer.

As a test, I installed Tonex on my newer laptop and ran a capture. The default training ran noticably faster than the old laptop, so I tried an advanced one. This was not successful. I click the start button and a flashing message saying creating network shows. After a few seconds, the stop button changes colour to orange, but from that point, the training never starts. I have left this for up to an hour where the old one starts to show progress in about a minute.

To try to fix this, I wiped the SSD and reinstalled Windows 10 and updated it to the same version as my working one - 21H2. The sound card drivers are the same version. The Tonex is 1.1.3. The same problem occurs with this very simple set up. As a note, when it is doing this, the processor load drops to <1% so it has stopped trying whatever it is supposed to be doing.

From reviewing the FAQ, the issue is not noise level, because the controls were not changed between one laptop and the other and the interface, RME802, stores any programable configurations between reconnections.

The later laptop has an Nvidia M620 GPU which I believe is a Maxwell architecture, so this should not be an issue.

The problem seems to have followed the laptop hardware. Is there any way that I can generate a debug log to send to IK?
jedsas
 
Posts: 21
Joined: Fri May 08, 2020 1:26 am

Re: Tonex capture bug

Postby Smaxo303 » Sat Mar 04, 2023 2:45 pm

As far as I understand, Tonex uses the graphics card (GPU) to calculate and optimize the AI ​​network. I think the CUDA cores are used to be precise. The M620 should be able to do this, but it may not be particularly fast.
The M620 has relatively little memory, could also be an issue.

https://www.notebookcheck.net/NVIDIA-Qu ... 796.0.html

It's best to open a ticket, there is may be a list of tested graphics cards or a way to send them more Informationen about your setup.

I make sometimes convolutional _ networks on the Jetson platform.
https://developer.nvidia.com/cuda-gpus
Smaxo303
 
Posts: 25
Joined: Fri Feb 24, 2023 9:51 pm


Return to AmpliTube Guitar Amp & FX modeling