Data Analysis¶
Tomographic reconstruction of 32-ID TXM data is typically done on machine txmthree.xray.aps.anl.gov: Intel Xeon i7 Silver 4214, 192GB RAM, NVidia Quadro RTX 4000 (8GB), 4TB SSD. For processing huge data volumes the instrument relies on the micro tomography computing infrastructure located at 2-BM:
Station |
Name |
Product |
Part list |
Model |
Quote |
2-BM |
tomo 1-2 |
MNJ15421064 |
Supermicro 740GP-TNRT cluster |
||
2-BM |
disk array |
MNJ15508749 |
SYS-220U-TNR Storage |
Reconstruction on CPU and GPU at the APS¶
Your raw data are automatically copied from the detector to the analysis computer (txmthree in this example) under the folder /local/ssd/data/YYYY-MM/PI_lastName. The SSD storage is limited, therefore after processing it is recommended to move data to a slower but larger HDD storage /local/data/YYYY-MM/PI_lastName.
Open tabs for reconstruction of 1 slice, full volume, google slides logging, Fiji, and data folder with reconstrutions.
[usertxm@txmthree,~]$ ~/start_rec.sh
The tabs will have appropriate conda environment activated and show examples of running reconstructions.
Reconstruction on CPU and GPU at your home institution¶
Download and install either tomocupy cli tool (reconstruction on GPU) or tomopy cli tool (reconstruction on CPU). Syntax is similar for both packages. List of commands:
$ tomocupy -h