4.6.4 slower than 4.6.3?

I have recently noticed that point cloud densification seem to take longer time that I expected. I did check an old project with 260 images and about 23 million points. With 4.6.3 densification took 19m:26s. Similar project with same settings with 21 million points took 52m:48s with 4.6.4!

Only difference that I can figure out is the Pix4D Mapper version. Computer is exactly same.

Hi Matti,

I had a look and there should be no differences between PIX4Dmapper version 4.6.3 and 4.6.4 when it comes to processing time.

Would it be possible to share the .log files or pdf quality reports for the same project but processed on different computers?

Best,

Blaz,

Here are the reports and log files.

Computer is same.

Something is now slowing the processing.

Thanks,

Matti

2020_12_10_Kaskinen_tehdas.log (2.13 MB)

2020_12_10_Kaskinen_tehdas_report.pdf (1.82 MB)

2021_06_01_Kaskinen.log (2.16 MB)

2021_06_01_Kaskinen_report.pdf (1.92 MB)

Hi Matti, thank you for the files.

I see what you mean, the hardware and the settings are exactly the same and I would not expect to see such a difference when processing on different PIX4Dmapper versions.

Is this something you observe also on other projects? Or is it only related to this one?

Is it possible that something else was running on the computer at the same time you were processing the project in PIX4Dmapper? Could you try to rerun the project after restarting the computer to make sure there are no background processes running?

Best,

Blaz,

I did notice at some point that processing is taking more time as it used to be. I think it is more clear with bigger projects than this this example.

I do not see any other processes taking much resources of processors or memory.

It may be that there is something wrong with my Computer.

If I have time I may try to install 4.6.3. and test with it.

Thanks

Matti

Let me know if you do any additional tests. I tried to run a few projects with both 4.6.3 and 4.6.4 but did not notice a significant increase in processing time.

Best,