- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I have the UP Xtreme system with two Movidius chips (UP AI Core XM 2280). I am looking for the datasheet, and a programmers guide for the Myriad X VPU 2485 chip. At this moment, I am not into playing with the existing AI examples. My interest is to use these chips as (low power) accelerators for radar (MIMO) signal processing algorithms. Is there any way to communicate with these chips, and use them for number crunching? I search the Intel site (as a register user) but found nothing about the architecture of the Myriad IC, and nothing about such APIs.
Any help is welcome.
Regards,
Darek
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Darek,
The datasheet and manual for AI Core XM 2280 are available in the Downloads section at the following link:
https://www.aaeon.com/en/p/ai-edge-computing-board-ai-core-xm-2280
The relevant materials for UP Extreme Series and UP AI Core X Series are available in the Downloads section at the following pages:
https://up-shop.org/up-xtreme-series.html
https://up-shop.org/up-ai-core-x-series-myriadtmx.html
Information about Movidius architecture is available at the following page:
https://en.wikichip.org/wiki/movidius/microarchitectures/shave_v2.0
The following page, https://builders.intel.com/ai , contains links to solutions catalog, documents library, blog, podcasts and videos, serving as a one-stop location for AI ecosystem innovation with Intel® architecture. I would encourage you to explore the variety of solutions available there.
Regards,
Munesh
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Munesh,
Thanks for you answer but I had these documents already. The AI Core XM 2280 module is based on the Myriad X VPU 2485 device. My questions were about how I can "talk" to these chips on the XM 2280 module. How I can open a communication channel to the device, prepare and send my data, and finally, retrieve results ( for example via the Myriad Development Kit (MDK) ). I am fine with low-level access to Shave & Leon cores as well :).
I did not find (yet) the datasheet of the Myriad X VPU 2485 chip itself.
I am worried that the MDK is not available for mortals .... (https://community.intel.com/t5/Intel-Distribution-of-OpenVINO/Where-can-I-get-the-Myriad-Development-Kit-MDK/td-p/693555)
Regards,
Darek
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Darek,
Inference of deep learning models on the Intel® Vision Accelerator Design with Intel® Movidius™ VPUs is supported by HDDL plugin.
The related configuration parameters are available at the following pages:
Regards,
Munesh
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page