Closed Rob174 closed 3 years ago
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓
┃ Name ┃ Access time (ms) ┃ Shape of the image ┃ Memory size (MB) ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩
│ 027481_0319CB_0EB7 avg time │ 834.939632 │ (10600, 18441) │ 781.8983999999999 │
│ 016505_01F10F_CE84 avg time │ 876.056842 │ (10655, 18706) │ 797.2497199999999 │
│ 016753_01F88A_4864 avg time │ 970.7034199999999 │ (10648, 18759) │ 798.9833279999999 │
│ Global avg time │ 893.8999646666665 │ - │ │
└─────────────────────────────┴───────────────────┴────────────────────┴───────────────────┘
and with h5py:
┌───────────────┬───────────────┬──────────────┬───────────────┬──────────────┐
│ Name │ Access time │ Access time │ Shape of the │ Memory size │
│ │ write (ms) │ read (ns) │ image │ (MB) │
├───────────────┼───────────────┼──────────────┼───────────────┼──────────────┤
│ 027481_0319C… │ 585.144 │ 0.0 │ (10600, │ 781.8983999… │
│ avg time │ │ │ 18441) │ │
│ 016505_01F10… │ 599.21039999… │ 0.0 │ (10655, │ 797.2497199… │
│ avg time │ │ │ 18706) │ │
│ 016753_01F88… │ 571.9643 │ 998400.0 │ (10648, │ 798.9833279… │
│ avg time │ │ │ 18759) │ │
│ Global avg │ 585.43956666… │ 332800.0 │ - │ │
│ time │ │ │ │ │
└───────────────┴───────────────┴──────────────┴───────────────┴──────────────┘
The read speed is most of the time less than 1ns (2.3 GHz processor) which is a bit too quick. Problem with the method to get the read speed (difference between time in ns) ?
We can take as an approximation 1ms
as it is the upper bound of the values
So, h5py is quicker
What is the time required to open one raster image versus to open it if it is in a hdf5 file ?