<div dir="ltr"><div>Great. I have managed to reproduce the error with a smaller volume and I can confirm that a more meaningful message ("Description: ITK ERROR: CUDA ERROR: out of memory") is obtained after this change</div><div><a href="https://github.com/SimonRit/RTK/pull/477">https://github.com/SimonRit/RTK/pull/477</a></div><div>Would be great if you can confirm that it works for you too before I merge.</div><div>Simon<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Wed, Feb 23, 2022 at 3:19 PM Vincent Libertiaux <<a href="mailto:vl@xris.eu">vl@xris.eu</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div>
<div>On 23.02.22 11:32, Simon Rit wrote:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div>Hi Vincent,</div>
<div>Thanks. I cannot reproduce the issue on my laptop is I only
modify the size of the reconstructed volume with "--dimension
2500,2500,10". With this argument, do you still have the same
issue?</div>
<div>I am suspecting that a problem occured before (e.g., memory
allocation failing) and I would suggest to add some
CUDA_CHECK_ERROR in the lines before, e.g., right after
cudaMalloc3DArray <a href="https://github.com/SimonRit/RTK/blob/master/src/rtkCudaUtilities.cu#L129-L133" target="_blank">here</a>.</div>
<div>Simon<br>
</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Tue, Feb 22, 2022 at 3:28
PM Vincent Libertiaux <<a href="mailto:vl@xris.eu" target="_blank">vl@xris.eu</a>>
wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Hi
Simon,<br>
<br>
thanks for the quick reply ! Here is a copy of the projections
file header:<br>
<br>
ObjectType = Image<br>
NDims = 3<br>
BinaryData = True<br>
BinaryDataByteOrderMSB = False<br>
CompressedData = False<br>
TransformMatrix = 1 0 0 0 1 0 0 0 1<br>
Offset = -204.69999999999999 -204.69999999999999 0<br>
CenterOfRotation = 0 0 0<br>
AnatomicalOrientation = RAI<br>
ElementSpacing = 0.20000000000000001 0.20000000000000001
0.20000000000000001<br>
DimSize = 2048 2048 101<br>
ElementType = MET_FLOAT<br>
ElementDataFile = LOCAL<br>
<br>
The geometry file is attached. It is not a traditional CT
geometry but <br>
a tomosynthesis one with a linear displacement of the source.<br>
<br>
Best regards,<br>
<br>
Vincent<br>
</blockquote>
</div>
</blockquote>
<p>Hi Simon,</p>
<p>thanks for having a look. I indeed managed to successfully
reconstruct the volume by using smaller dimensions. It is the
first time that particular message showed up to indicate the
issue.</p>
<p>Thanks again for your time,</p>
<p>Vincent<br>
</p>
</div>
</blockquote></div>