- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi, I have this
for (…) {
enquee()
enquee()
read()
read()
}
after a few thousands of iteration program get blocked on a read call,
but this way it work
for (…) {
enquee()
read()
enquee()
read()
}
the issue is that this last way is slower
any suggestion?
Link Copied
1 Reply
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@llero The issue may be related to performing a read() and the output queue being empty. I was just wondering what version of the NCSDK you are using? Please try using the latest NCSDK version (2.05) if you aren't already doing so. Some fifo issues were fixed with version 2.05 that were a problem with 2.04.
You can also try increasing the size of the fifos. The number of fifo elements can be set using Graph.allocate_with_fifos() or Fifo.allocate() and changing the number of elements for each fifo. This will allow you to be able to have larger fifo queues.
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page