Patch match::Problem

[christo...@gmail.com]

Hello,

I'm new to Colmap, and first of all thank you for this great piece of software, I wrote a shell script instead of using the gui and it works very well. 

but I get the following warnings, is that normal?

Undistorting image [469/470]

Undistorting image [470/470]

Writing reconstruction...

Writing configuration...

Writing scripts...

Elapsed time: 3.885 [minutes]

Reading workspace...

Reading configuration...

Configuration has 470 problems...

==============================================================================

Processing view 1 / 470

==============================================================================

Reading inputs...

PatchMatch::Problem

-------------------

ref_image_id: 194

src_image_ids: 195 157 196 158 156 197 155 159 193 154 92 93 94 91 192 153 95 88 89 90

PatchMatch::Options

-------------------

max_image_size: -1

gpu_index: 0

depth_min: 1.14431

depth_max: 2.33945

window_radius: 5

sigma_spatial: 5

sigma_color: 0.2

num_samples: 15

ncc_sigma: 0.6

min_triangulation_angle: 1

incident_angle_sigma: 0.9

num_iterations: 5

geom_consistency: 0

geom_consistency_regularizer: 0.3

geom_consistency_max_cost: 3

filter: 0

filter_min_ncc: 0.1

filter_min_triangulation_angle: 3

filter_min_num_consistent: 2

filter_geom_consistency_max_cost: 1

write_consistency_graph: 0

PatchMatch::Run

---------------

out of memory in /home/christoph/software/colmap/src/./mvs/gpu_mat.h at line 152

StereoFusion::Options

---------------------

max_image_size: -1

min_num_pixels: 5

max_num_pixels: 10000

max_traversal_depth: 100

max_reproj_error: 2

max_depth_error: 0.01

max_normal_error: 10

check_num_images: 50

cache_size: 32

Reading workspace...

Reading configuration...

WARNING: Ignoring image DSC03027.jpg, because input does not exist.

WARNING: Ignoring image DSC03028.jpg, because input does not exist.

WARNING: Ignoring image DSC03029.jpg, because input does not exist.

WARNING: Ignoring image DSC03030.jpg, because input does not exist.

WARNING: Ignoring image DSC03031.jpg, because input does not exist.

that list goes on until 

WARNING: Ignoring image DSC03494.jpg, because input does not exist.

WARNING: Ignoring image DSC03495.jpg, because input does not exist.

WARNING: Ignoring image DSC03496.jpg, because input does not exist.

my question is what to do now? and how to avoid that warning?

cheers christoph

 

[Johannes Schönberger]

Hi,

The errors says:

out of memory in /home/christoph/software/colmap/src/./mvs/gpu_mat.h at line 152

Please refer to: https://colmap.github.io/faq.html#reduce-memory-usage-during-dense-reconstruction

Cheers,
Johannes

> --
> You received this message because you are subscribed to the Google Groups "COLMAP" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to colmap+un...@googlegroups.com.
> To post to this group, send email to col...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/f68206b0-9063-4b11-839d-eecd1f3a0ab4%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Christoph Macho]

Thank you for replying, I already solved the problem.

As you mentioned you will bypass when shutting down x-server. Just out of  curiosity ,isn't there a possibility to avoid that error since cuda6 as far as I have  read,  it will use CPU memory if running out of  GPU memory. As I have plenty of system memory I am a bit surprised running into trouble??

Thanks a lot christoph

On 23 Oct 2017, at 11:12, "Johannes Schönberger" <js...@demuc.de> wrote:

Hi,

The errors says:

 out of memory in /home/christoph/software/colmap/src/./mvs/gpu_mat.h at line 152

Please refer to: https://colmap.github.io/faq.html#reduce-memory-usage-during-dense-reconstruction

Cheers,
Johannes

 On Oct 22, 2017, at 9:57 PM, christo...@gmail.com wrote:
 
 Hello,
 I'm new to Colmap, and first of all thank you for this great piece of software, I wrote a shell script instead of using the gui and it works very well. 
 but I get the following warnings, is that normal?
 Undistorting image [469/470]
 Undistorting image [470/470]
 Writing reconstruction...
 Writing configuration...
 Writing scripts...
 Elapsed time: 3.885 [minutes]
 Reading workspace...
 Reading configuration...
 Configuration has 470 problems...
 

 Processing view 1 / 470

[Johannes Schönberger]

I don't know from where you have the information that in CUDA 6 the GPU can use CPU memory. This is not the case. There is something called unified memory in CUDA 8 and later but that still doesn't solve this problem. In any case you need sufficient GPU memory for the COLMAP dense reconstruction and it cannot be shared with the CPU.

Best,
Johannes

> On Oct 23, 2017, at 12:18 PM, Christoph Macho <christo...@gmail.com> wrote:
>
> Thank you for replying, I already solved the problem.
> As you mentioned you will bypass when shutting down x-server. Just out of curiosity ,isn't there a possibility to avoid that error since cuda6 as far as I have read, it will use CPU memory if running out of GPU memory. As I have plenty of system memory I am a bit surprised running into trouble??
> Thanks a lot christoph
> On 23 Oct 2017, at 11:12, "Johannes Schönberger" <js...@demuc.de> wrote:
> Hi,
>
> The errors says:
>
> out of memory in /home/christoph/software/colmap/src/./mvs/gpu_mat.h at line 152
>
> Please refer to: https://colmap.github.io/faq.html#reduce-memory-usage-during-dense-reconstruction
>
> Cheers,
> Johannes
>
> On Oct 22, 2017, at 9:57 PM, christo...@gmail.com wrote:
>
> Hello,
> I'm new to Colmap, and first of all thank you for this great piece of software, I wrote a shell script instead of using the gui and it works very well.
> but I get the following warnings, is that normal?
> Undistorting image [469/470]
> Undistorting image [470/470]
> Writing reconstruction...
> Writing configuration...
> Writing scripts...
> Elapsed time: 3.885 [minutes]
> Reading workspace...
> Reading configuration...
> Configuration has 470 problems...
>
>

> Processing view 1 / 470
>
>

> To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/285b138a-f0a1-412c-b0e5-5ec56164194f%40gmail.com.

[Christoph Macho]

Hello,

Exactly unified memory that was the point. Under https://devblogs.nividia.com/parallel forall/unified-memory-in-cuda-6/

I read about that, but anyway what would be a decent amount of GPU memory for a scenery with 500 pictures each 6000 to 4000 pixel and will make it a difference if downsampling  images ? 

cheers chris

>  To unsubscribe from this group and stop receiving emails from it, send an email to colmap+unsubscribe@googlegroups.com.

>  To post to this group, send email to col...@googlegroups.com.
>  To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/f68206b0-9063-4b11-839d-eecd1f3a0ab4%40googlegroups.com.
>  For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups "COLMAP" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to colmap+unsubscribe@googlegroups.com.

> To post to this group, send email to col...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/285b138a-f0a1-412c-b0e5-5ec56164194f%40gmail.com.
> For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "COLMAP" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/colmap/5PcP_fLfagY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to colmap+unsubscribe@googlegroups.com.

To post to this group, send email to col...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/983D1007-A051-4802-A241-22D8A1E09772%40demuc.de.

[Johannes Schönberger]

For GPU memory usage recommendations please see https://colmap.github.io/faq.html#reduce-memory-usage-during-dense-reconstruction

> > To unsubscribe from this group and stop receiving emails from it, send an email to colmap+un...@googlegroups.com.

> > To post to this group, send email to col...@googlegroups.com.
> > To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/f68206b0-9063-4b11-839d-eecd1f3a0ab4%40googlegroups.com.
> > For more options, visit https://groups.google.com/d/optout.
> >
> > --
> > You received this message because you are subscribed to the Google Groups "COLMAP" group.

> > To unsubscribe from this group and stop receiving emails from it, send an email to colmap+un...@googlegroups.com.

> > To post to this group, send email to col...@googlegroups.com.
> > To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/285b138a-f0a1-412c-b0e5-5ec56164194f%40gmail.com.
> > For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to a topic in the Google Groups "COLMAP" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/colmap/5PcP_fLfagY/unsubscribe.

> To unsubscribe from this group and all its topics, send an email to colmap+un...@googlegroups.com.

> To post to this group, send email to col...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/983D1007-A051-4802-A241-22D8A1E09772%40demuc.de.
> For more options, visit https://groups.google.com/d/optout.
>

> --
> You received this message because you are subscribed to the Google Groups "COLMAP" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to colmap+un...@googlegroups.com.

> To post to this group, send email to col...@googlegroups.com.

> To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/CAJ%2B7db62F5YBFgJkVy4v3z4QR_2sGMiLB1oo3TGDL7q4PDuq6A%40mail.gmail.com.

[Christoph Macho]

Thanks for the information.

I just had the situation of PatchMatch running for 3 days and created about 420 scenes out of 470 and then crashed due to memory issues. How can that happen? Do I need to run again all the scenes or are the stored somewhere in a database?

Cheers Christoph

[Johannes Schönberger]

The results are stored on disk and will be skipped when you restart.

To view this discussion on the web visit https://groups.google.com/d/msgid/colmap/17f1c685-4e93-497e-81fd-aabff525f413%40gmail.com.

[Christoph Macho]

Wow, you are fast 😀,

Thank you very much, I will try to resume!