Build for NVIDIA Terga TX2 Platform

Question

Build for NVIDIA Terga TX2 Platform

renjie130 opened this issue 7 years ago · 12 comments

Errors occured when we ran it on the NVIDIA Terga TX2 Platform. Can it run on Terga Tx2 ? Details are as follows. Thank you!

UNAME_MACHINE = aarch64
UNAME_RELEASE = 4.4.38-tegra
UNAME_SYSTEM = Linux
UNAME_VERSION = #1 SMP PREEMPT Thu Mar 1 20:49:20 PST 2018
configure: error: cannot guess build type; you must specify one
CMakeFiles/xproto.dir/build.make:108: recipe for target 'xproto-prefix/src/xproto-stamp/xproto-configure' failed
make[8]: Leaving directory '/home/nvidia/.hunter/_Base/8fee57e/3942445/a0ab86d/Build/xproto/Build'
make[8]: *** [xproto-prefix/src/xproto-stamp/xproto-configure] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/xproto.dir/all' failed
make[7]: Leaving directory '/home/nvidia/.hunter/_Base/8fee57e/3942445/a0ab86d/Build/xproto/Build'
make[7]: *** [CMakeFiles/xproto.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make[6]: Leaving directory '/home/nvidia/.hunter/_Base/8fee57e/3942445/a0ab86d/Build/xproto/Build'
make[6]: *** [all] Error 2

[hunter ** FATAL ERROR **] Build step failed (dir: /home/nvidia/.hunter/_Base/8fee57e/3942445/a0ab86d/Build/xproto
[hunter ** FATAL ERROR **] [Directory:/home/nvidia/.hunter/_Base/Download/Hunter/0.20.28/8fee57e/Unpacked/cmake/projects/xproto]

Answer 1 · 2018-04-17T04:06:17.000Z

Can you add log with HUNTER_STATUS_DEBUG=ON?

Answer 2 · 2018-04-17T04:41:44.000Z

I will try it. However, I do not think it works. I think it is the older version of xproto, which provided.
Thank you.

Answer 3 · 2018-04-17T05:16:04.000Z

I will try it. However, I do not think it works.

HUNTER_STATUS_DEBUG=ON will print extra info, I haven't said it wil fix the build.

Answer 4 · 2018-04-17T05:46:49.000Z

Got it. Thanks

Answer 5 · 2018-04-17T14:46:12.000Z

Can it run on Terga Tx2 ?

I don't have one, but if you send me one I can test it 😄. From a quick read, it looks like that board uses NVIDIA's Linux4Tegra (L4T), which is similar to Ubuntu, Right? If it is close to Ubuntu, I think it should work without too much trouble.

There are clang and gcc CI builds on Ubuntu trusty, for example. Note: Those currently only run CPU tests, since there isn't support for GPU testing on Travis:

https://travis-ci.org/elucideye/acf/builds/364195910

A little background might be helpful. For the end-to-end processing, we have: (1) ACF pyramid computation and; (2) multi-scale object detection. The gradient boosting detection stuff always runs on the CPU. The pyramids can either be computed using CPU SSE/NEON instructions via Piotr's toolbox SIMD code, or on the GPU using OpenGL ES 2.0 (or equivalent). The shader processing is implemented on top of ogles_gpgpu . The CPU code should run anywhere SSE/NEON is supported, and the GPU code should run on most platforms where OpenGL w/ GLSL 1.1 is supported. I say most, only because this is limited in practice to platforms where ogles_gpgpu currently builds. That is currently tested on versions of: Windows, OSX, Ubuntu, iOS, and Android. On other platforms, like Raspberry Pi, ogles_gppu might require some minor updates: cmake, opengl header paths, etc.

You seem to be hitting an issue with xproto specifically. It is worth noting that this isn't actually used in the acf library directly, but it is used in the tests and sample app via the aglet library (which uses glfw -> xproto). That lib is used to provide a portable OpenGL context to support cross platform OpenGL tests and applications. For a quick test, you could try to turn off the samples/tests by passing ACF_BUILD_EXAMPLES=OFF and ACF_BUILD_TESTS=OFF to cmake to see if the acf library builds OK.

If you share your build command and the full log we can probably track down the xproto issue.

You could also check to see if the CPU only acf library is building. You can do this by disabling the ogles_gpgpu dependency from the command line by passing BUILD_OGLES_GPGPU=OFF to CMake.

Answer 6 · 2018-04-18T02:35:50.000Z

@headupinclouds Thanks for your reply. I will try it and the the results will be posted.

Answer 7 · 2018-04-19T04:45:02.000Z

Got it.Thanks 2018-04-17 22:46 GMT+08:00 David Hirvonen <notifications@github.com>:

…

Can it run on Terga Tx2 ? I don't have one, but if you send me one I can test it 😄. From a quick read, it looks like that board uses NVIDIA's Linux4Tegra (L4T), which is similar to Ubuntu, Right? If it is close to Ubuntu I think it should work without too much trouble. There are clang and gcc CI builds on Ubuntu trusty, for example. Note: Those currently only run CPU tests, since there isn't support for GPU testing on Travis: https://travis-ci.org/elucideye/acf/builds/364195910 A little background might be helpful. For the end-to-end processing, we have: (1) ACF pyramid computation and; (2) multi-scale object detection. The gradient boosting detection stuff always runs on the CPU. The pyramids can either be computed using CPU SSE/NEON instructions via Piotr's toolbox SIMD code, or on the GPU using OpenGL ES 2.0 (or equivalent). The shader processing is implemented on top of ogles_gpgpu <https://github.com/hunter-packages/ogles_gpgpu>. The CPU code should run anywhere SSE/NEON is supported, and the GPU code should run on *most* platforms where OpenGL w/ GLSL 1.1 is supported. I say *most*, only because this is limited in practice to platforms where ogles_gpgpu currently builds. That is currently tested on versions of: Windows, OSX, Ubuntu, iOS, and Android. On other platforms, like Raspberry Pi, ogles_gppu might require some minor updates: cmake, opengl header paths, etc. You seem to be hitting an issue with xproto specifically. It is worth noting that this isn't actually used in the acf library directly, but it is used in the tests and sample app via the aglet library (which uses glfw -> xproto). That lib is used to provide a portable OpenGL context to support cross platform OpenGL tests and applications. For a quick test, you could try to turn off the samples/tests by passing ACF_BUILD_EXAMPLES=OFF and ACF_BUILD_TESTS=OFF to cmake to see if the acf library builds OK. If you share your build command and the full log we can probably track down the xproto issue. You could also check to see if the CPU only acf library is building. You can do this by disabling the ogles_gpgpu dependency from the command line by passing BUILD_OGLES_GPGPU=OFF to CMake. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#51 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ahp3ekWkRzBp-9L9B1mGs5a0HBrliKUsks5tpgA1gaJpZM4TXlzS> .

Answer 8 · 2018-04-22T20:27:43.000Z

Note: sample application added in #55, this will likely be merged to the installed library in a follow up after further tuning (see GPUDetectionPipeline)

Answer 9 · 2018-04-29T21:36:42.000Z

FYI: #53 (comment)

Answer 10 · 2018-05-11T02:24:39.000Z

@renjie130 : FWIW It looks like @D0CX4ND3R was able to build for this board, see #58

Answer 11 · 2018-10-27T13:23:01.000Z

Inactive, closing

Answer 12 · 2018-10-28T08:16:07.000Z

@renjie130 : FWIW It looks like @D0CX4ND3R was able to build for this board, see #58

Got it. Ths