# host machine development vs. on board development?

**URL:** <https://forums.developer.nvidia.com/t/host-machine-development-vs-on-board-development/45015>\
**Category:** Jetson TX1\
**Created:** [September 19, 2016, 3:49pm UTC](https://forums.developer.nvidia.com/t/host-machine-development-vs-on-board-development/45015 "2016-09-19T15:49:34Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![openr2](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@openr2](https://forums.developer.nvidia.com/u/openr2)\
**Post date:** [September 19, 2016, 3:49pm UTC](https://forums.developer.nvidia.com/t/host-machine-development-vs-on-board-development/45015/1 "2016-09-19T15:49:34Z")

</div>

Except for flashing the board and initial install of the OS and dev samples, do most people develop directly on the carrier board?

---

<div class="post-metadata">

**Author:** ![linuxdev](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/linuxdev/32/10243_2.png) [@linuxdev](https://forums.developer.nvidia.com/u/linuxdev)\
**Post date:** [September 19, 2016, 6:39pm UTC](https://forums.developer.nvidia.com/t/host-machine-development-vs-on-board-development/45015/2 "2016-09-19T18:39:42Z")

</div>

The JTX1 kernel compile is best via cross compile due to supporting both 32-bit and 64-bit (there are two compilers required, but installing the second compiler on the Jetson would require adding a foreign architecture…this tends to be messy at best for anything other than desktop 64-bit supporting 32-bit compatibility…ARM 64-bit and 32-bit mix in a single environment does not really exist).

If you just want to develop purely 64-bit user space, then building on the Jetson is probably easier. I don’t use JetPack due to using a Fedora host, but something like nsight eclipse somewhat blurs the distinction by doing some parts on host and still running on the Jetson.