65ORG16.b Core

Topics relating to PALs, CPLDs, FPGAs, and other PLDs used for the support or creation of 65-family processors, both hardware and HDL.
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

Interesting, I set up a top level constraint for the O2IN clock at 12ns, and it passed! Now it says best achievable case is 11.68ns. Will try 11ns next...
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

Is this something to be concerned about? Been giving me the warning since making the adjustment to the Z flag.

Code: Select all

WARNING:Xst:2677 - Node <XLXI_1/_ALU/Z> of sequential type is unconnected in block <top_level>.


It did fail @11ns. You mean this?

Image
User avatar
Arlet
Posts: 2353
Joined: 16 Nov 2010
Location: Gouda, The Netherlands
Contact:

Post by Arlet »

If you click on the paths, you get more detailed information.

But 12 ns is pretty good. With the 18 ns, I thought something may be wrong, but 12 ns is fine. That's going to be pretty much impossible to improve.

And, yeah, in order to see the top speed, you'll have to play with the constraints. If you put the constraint at 20 ns, and the tools reach 19.9, they're not going to try any harder, so you'll never see the top speed.
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

Is this something to be concerned about? Been giving me the warning since making the adjustment to the Z flag.

Code: Select all

WARNING:Xst:2677 - Node <XLXI_1/_ALU/Z> of sequential type is unconnected in block <top_level>.
__________________________________________________________________________________________________________________________


Also interesting to note, when I set optimize for speed to high. Max delay is 12.002ns. When I set it back to normal, it passes again with Max delay of 11.551ns...

Next on the agenda is to trim BCD out, which lucky for me I have your original .cpu and .ALU files from 10/27/2010 to compare against. But there are a couple places where I'm not sure whether I should trim them out of not and would appreciate your help.
User avatar
Arlet
Posts: 2353
Joined: 16 Nov 2010
Location: Gouda, The Netherlands
Contact:

Post by Arlet »

The warning is about the unconnected 'Z' output in the ALU. It's no problem, but
you can get rid of it by editing ALU.v to remove all references to 'Z'.
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

I think I've got the ALU.v trimmed of BCD.
Looking at cpu.v right around line 543 that contains:

Code: Select all

AXYS[regsel] <= (state == JSR0) ? DIMUX : { ADD[dw-1:4] + ADJH, ADD[3:0] + ADJL };
Should I change it back to:

Code: Select all

AXYS[regsel] <= ADD;
User avatar
Arlet
Posts: 2353
Joined: 16 Nov 2010
Location: Gouda, The Netherlands
Contact:

Post by Arlet »

No, it should be:

Code: Select all

AXYS[regsel] <= (state == JSR0) ? DIMUX : ADD;
That's needed for the RDY support.
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

OK, excellent, I think i got it!
A few more changes to the comments and I'll post them on github!

Will let you guys know when I'm done :D
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

Was abit premature, but after a few more corrections, I got it to pass synthesis ok. No speed gains by removing BCD...

An area to observe for correctness which was a little bit of a challenge is around line 98 in the ALU, and lines 1096-1112 in the cpu. I think they're ok so I'm going forward.... Should I add my name to the copyright? heh :lol:
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

This github is giving me a headache! I just want to upload 2 files... I don't have time to read all this stuff!
User avatar
BigEd
Posts: 11463
Joined: 11 Dec 2008
Location: England
Contact:

Post by BigEd »

you've made a fork, so now you can visit a file you want to change, hit Edit This File, and paste in your new contents. Then put in a commit message and hit Commit Change.

Edit: for example, this is your cpu.v file
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

man I'm slow with this stuff. Does look neat though. Can see code in github's editor... But I'm not just adding comments and such, I've modded both entire files.
User avatar
BigEd
Posts: 11463
Joined: 11 Dec 2008
Location: England
Contact:

Post by BigEd »

right, so just select-all and cut, then paste in your new version.
ElEctric_EyE
Posts: 3260
Joined: 02 Mar 2009
Location: OH, USA

Post by ElEctric_EyE »

sec, do I need to be using explorer? I'm using chrome right now. I don't see those options...
User avatar
BigEd
Posts: 11463
Joined: 11 Dec 2008
Location: England
Contact:

Post by BigEd »

Hi EEye
no, any browser, any OS. Let me walk you through it.

these are text operations you're about to do, in a text box - rather like these comment editting boxes. You're not going to upload a file, you're going to post new content to github, and then github's version of the file will be changed.

so, you visit the file in github, in your fork, and you hit the edit button, and the file content in the text box is now open for edit. You could now make a small change or two, and hit the commit button, preferably with a descriptive comment in the comment box.

for big change, I'm suggesting you wipe out all the text in the text box, by selecting all of it: Edit-> Select All -> Delete, or Right Click->Select All->Delete. And then visit your favourite editor (Notepad, Wordpad, ISE), pull up the new file contents, select the whole file text using Edit->Select All ->Copy. Finally, back to the browser and the github text box, paste in your updated content: Edit->Paste or Right Click->Paste.

(Or you might be the kind of person who uses Control-A, Control-C, Control-X and so on)

Does that start to make more sense? You're editting text, not uploading files. But when you hit submit, github will be up to date, and you can download a new version (and so can I!)

Cheers
Ed
Post Reply