RTKConsortium / RTK

Reconstruction Toolkit
Apache License 2.0
245 stars 145 forks source link

rtk::CudaIterativeFDKConeBeamReconstructionFilter memory error. #623

Closed jxhno1 closed 1 month ago

jxhno1 commented 1 month ago

I used CudaIterativeFDKConeBeamReconstructionFilter to try a iter fdk case, a error occur when the data size is 1294x952x600, image but the data size 1294x952x300 is ok! then I put the code on the 4090 card, as GPU memory is 24G, but the error still here. then I try the cpu filter rtk::IterativeFDKConeBeamReconstructionFilter, data size 1294x952x600 run ok, so I think the CudaIterativeFDKConeBeamReconstructionFilter may have a bug in memory management?

SimonRit commented 1 month ago

Which RTK version are you using? There's been a bug fixed recently, it hasn't been released yet. See #612.

jxhno1 commented 1 month ago

itkRTK:5.2 RTK:2.1

jxhzk1314

@. | ---- Replied Message ---- | From | Simon @.> | | Date | 10/14/2024 23:09 | | To | @.> | | Cc | @.>@.***> | | Subject | Re: [RTKConsortium/RTK] rtk::CudaIterativeFDKConeBeamReconstructionFilter memory error. (Issue #623) |

Which RTK version are you using? There's been a bug fixed recently, it hasn't been released yet. See #612.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

SimonRit commented 1 month ago

Then there is another bug fix which explains it, #598. Please reopen if you still have the problem with the master version.

jxhno1 commented 1 month ago

Thanks,I will try.

jxhzk1314

@. | ---- Replied Message ---- | From | Simon @.> | | Date | 10/16/2024 13:28 | | To | @.> | | Cc | @.>@.***> | | Subject | Re: [RTKConsortium/RTK] rtk::CudaIterativeFDKConeBeamReconstructionFilter memory error. (Issue #623) |

Then there is another bug fix which explains it, #598. Please reopen if you still have the problem with the master version.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>