XMOS, THANK YOU for the great summary view, when I click on a Binary in the XDE.
This has real promise, and I'm hoping to get the best use of it.
I attach an image below, that shows the XN file, and the view I see. I would like to know the significance of the red entries on the right panel. Is there a reference doc, yet, on this? It seems not to be covered in the latest Tools pdf.
Can we have a graphic, or table, showing the runtime link and channel connectivity, as well as the understood load order and links?
Thanks,
--r.
Understanding the display of Binaries
-
- XCore Addict
- Posts: 146
- Joined: Thu Dec 10, 2009 10:17 pm
Understanding the display of Binaries
You do not have the required permissions to view the files attached to this post.
-
- XCore Addict
- Posts: 146
- Joined: Thu Dec 10, 2009 10:17 pm
Some of the glitz is going off this, already. Sorry!
I'm just outlining a board bring up app. I intend to run around the IO, reading all the uncommitted inputs, and test with VCC through a 1k resistor, looking for connectivity and shorts. All the time printing the port state. Once that is checked, I go to phase 2, where I drive outputs, one by one, checking for voltage, and smoke ;)
Simple enough.
So I started building a list of 1-bit ports for one node, and found that just creating a list of ports, and including a port within it (XS1_PORT_1A, in this case) caused that same 1 bit in all 3 packages to turn red, as well as P1:X0D00 , P2:X0D00, P3:X0D00, on the BInary/UserIOs listing, where P[1..3] are my package IDs in the XN file, and are displayed on the right.
(Image attached.)
Bear in mind that nothing yet is actually referencing the list of ports. I have a feeling this just a piece of code that did not get finished before release, but would welcome a revelation!
I would have expected the simple mention of a port in an unused array to have no effect. And why am I affecting 3 packages, with one port definition?
Best,
--r
I'm just outlining a board bring up app. I intend to run around the IO, reading all the uncommitted inputs, and test with VCC through a 1k resistor, looking for connectivity and shorts. All the time printing the port state. Once that is checked, I go to phase 2, where I drive outputs, one by one, checking for voltage, and smoke ;)
Simple enough.
So I started building a list of 1-bit ports for one node, and found that just creating a list of ports, and including a port within it (XS1_PORT_1A, in this case) caused that same 1 bit in all 3 packages to turn red, as well as P1:X0D00 , P2:X0D00, P3:X0D00, on the BInary/UserIOs listing, where P[1..3] are my package IDs in the XN file, and are displayed on the right.
(Image attached.)
Bear in mind that nothing yet is actually referencing the list of ports. I have a feeling this just a piece of code that did not get finished before release, but would welcome a revelation!
I would have expected the simple mention of a port in an unused array to have no effect. And why am I affecting 3 packages, with one port definition?
Best,
--r
You do not have the required permissions to view the files attached to this post.
-
- Experienced Member
- Posts: 69
- Joined: Mon May 17, 2010 10:19 am
Declaring a port without an "on" component causes that declaration to be enacted on every core in the system. If you change your initialised from "XS1_PORT_1A" to "on stdcore[0]: XS1_PORT_1A" you should see it affect one package only.
A port declaration is not a no-op. Merely declaring a port causes it to be initialised so it is considered to be used whether it's involved in any explict IO or not.
A port declaration is not a no-op. Merely declaring a port causes it to be initialised so it is considered to be used whether it's involved in any explict IO or not.
-
- XCore Addict
- Posts: 146
- Joined: Thu Dec 10, 2009 10:17 pm
Thanks m_y,
to run a full pintest in a three processor network, I can expect red on almost all the ports in the Binary display.
I presume I can take these as warnings.
Thanks for your help!
--r.
Ah. OK. That makes sense, as far as it goes....m_y wrote:Declaring a port without an "on" component causes that declaration to be enacted on every core in the system. If you change your initialised from "XS1_PORT_1A" to "on stdcore[0]: XS1_PORT_1A" you should see it affect one package only.
I see. So, if I use a technique such as that used in Henk's pintest ...m_y wrote:A port declaration is not a no-op. Merely declaring a port causes it to be initialised so it is considered to be used whether it's involved in any explicit IO or not.
Code: Select all
struct pinDescriptors ports[NPINS] = {
{ XS1_PORT_1A, 1, "PORT_1A", 0 },
{ XS1_PORT_1B, 1, "PORT_1B", 1 },
etc...
I presume I can take these as warnings.
Thanks for your help!
--r.
-
- Experienced Member
- Posts: 69
- Joined: Mon May 17, 2010 10:19 am
Yes.russf wrote:I can expect red on almost all the ports in the Binary display.