# Why does trying to render with Cuda on a remote server fail?

I’m trying to do a remote render on a server with linux, when I do renders by Cpu it’s fine, but when I want to use cuda I have errors. The script that I use is this

These are my Gpu

When calling the render

blender -b finaltestcuda.blend -E CYCLES -P script.py -f 1


This error appears

I don't know your specific error code, getting this ***** to run on server took me forever. What helped was getting the latest Daily Build and installing it instead of the default blender package. Also try a similar code to mine if nothing works (mine works for two Titans XP's, you'll have to rewrite it for your GPUs I guess). If I remember correctly enabling all devices enabled CPU and caused blender to only use that.

def gpuSettings():
bpy.context.scene.render.engine = 'CYCLES'
bpy.data.scenes["Scene"].cycles.device = 'GPU'
bpy.data.scenes["Scene"].cycles.feature_set = 'SUPPORTED'

# Switch render engine to cycles

for device in cprefs.devices:
if device.name == 'TITAN Xp':
device.use = True
else:
device.use = False

#print('Compute_device_type: ', cprefs.compute_device_type)
for device in cprefs.devices:
if device.use == True:
print('USED: ', device.name)
else:
print('NOT_USED: ', device.name)
print('Will be used: ' + bpy.data.scenes["Scene"].cycles.device)
print('Feature set: ' + bpy.data.scenes["Scene"].cycles.feature_set)
print('Render engine: ' + bpy.context.scene.render.engine)
return