Killing Chips

27 views
Skip to first unread message

Andrew Hansen

unread,
May 1, 2018, 12:21:57 PM5/1/18
to dynapse-users
Hi all,

I've noticed that when using netparser, building adjacent connections across chips (example: U0->U3) seems to kill the receiving chip. Any reason why this happens? Is it a hardware issue?

I've developed my own module, and it still seems to be a recurring issue, here's example input from my module. This kills chip 3. Any advice/explanation would be helpful.

CAM:  0001001010010001100000000010110000
SRAM:1000110000000100100000000000100001

Best,
Andrew

carsten.nielsen

unread,
May 3, 2018, 4:17:08 AM5/3/18
to dynapse-users
Hi Andrew,

I don't think these SRAM bits are valid. Are you programming the SRAM bits directly or using the libcAER convenience function writeSramN?

Cheers,
Carsten

Andrew Hansen

unread,
May 16, 2018, 8:02:10 PM5/16/18
to dynapse-users
Thanks for the response Carsten, I figured it out. It had to do with an error in a CAM-clash check.

Best,
Andrew
Reply all
Reply to author
Forward
0 new messages