Testing TotalDepth With Recorded Data

TotalDepth is tested against a diverse data set of real world files. The test set is split into small/medium/large datasets.

Small Test Set

todo:Complete this

Medium Test Set

The Medium Size Test Set is 20,000+ files (130Gb+) of typical oilfield data. Here is the approximate breakdown of the test set:

File Type	Files	Total Size	Notes
BIT	~500	~1.5Gb	Largest file is around 6Mb.
LAS v1.2	~500	~1Gb	Largest file is around 16Mb.
LAS v2.0	~20,000	~30Gb	Largest file is around 250Mb (RP66V1 converted files are much larger).
LAS v3.0	A few	~0	Rarely present, their absence is not considered significant.
LIS	~2000	~2GB	Largest file is around 60Mb. Around half have TIF markers.
DLIS (RP66V1)	~800	~100GB	Largest file is around 4GB. About one quarter are corrupted by TIF markers.
DLIS (RP66V2)	0	0	Not present, their absence is not considered significant.
Other	Various	Various	Various files such as PDF, TIFF, miscellaneous binary files and unstructured ASCII. If present then their contents is not considered significant but archives containing these should be processed without drama.

The layout of the test set is typical of an oilfield repository, typically by well, with a well having an unspecified directory structure and a mix of file types in each directory.

The aim is that TotalDepth can process 97+% of the files in this archive in the formats that TotalDepth supports.

Large Test Set

todo:Complete this

Synthetic Test Set

TotalDepth can generate arbitrary sized files in a number of formats and this test set is used for performance analysis. This test set is not distributed with TotalDepth but the means to create it is.