1. Home
  2. Field-Programmable Gate Arrays
  3. Net name convention for Xilinx UCF files.

Net name convention for Xilinx UCF files.

Hi, all:

I am writing some UCF files for synthesis. In XST I set the hierarchy seperator to / and bus symbol to []...

Now after I write the UCF files and build, the ISE6.1 complain can't find the net-names.

How do I find the correct names of the various symbols in my Verilog codes? Is it possible to write a "netlist" in the NGC files?

Best Regards, Kelvin

In core.v Module main;

testctl u_testctrl( ... .srck_i(p_srck_i), //testctrl input

.arstn_testrx_i(arstn_testrx), //testctrl input .clk36_pll_i(clk_36), //testctrl input );

... endmodule

in core_ucf.ucf NET "u_testctrl/clk36_pll_o" TNM_NET = "u_testctrl_clk36_pll_o"; TIMESPEC "TS_u_testctrl_clk36_pll_o" = PERIOD "u_testctrl_clk36_pll_o" 27.8 ns HIGH 50 %;

Error in core.bld ERROR:NgdBuild:756 - Line 87 in core_ucf.ucf': Could not find net(s) 'u_testctrl/clk36_pll_i' in the design. To suppress this error specify the correct net name or remove the constraint.

Reply to
One Day & A Knight
Loading thread data ...

codes?

FPGA Editor helps. You can also use the Constraints Editor.

Also try: "*u_testctrl/clk36_pll_o" instead of: "u_testctrl/clk36_pll_o"

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Martin Euredjian

To send private email:
0_0_0_0_@pacbell.net
where
"0_0_0_0_"  =  "martineu"
Reply to
Martin Euredjian

Hi,

Your .ucf having "pll_o" and ERROR shows "pll_i". I think you have copied wrong lines. thats ok.

XST tools will not change the instance names during synthesis. so, whatever you type in RTL should be valid in PAR too.

I am doubting, the hiearchy seperator only. Because, XST by default takes "." as hiearchy seperator. Pls check that again, whether you have forced correctly to "/"

and you can generate netlist after Translation phase.

it is "ngd2ver" application from Xilinx tools.

Get back for any clarification.

Regards, Muthu

Reply to
Muthu

Yeah, that is indeed a copy error.

Now I switch to a single "clk_36"...and it seemed to work. "clk_36" is an implicitly defined wire label.

I do remember there is a document said all the names and definitions in the source code will remain unchanged, but how come it doesn't in my case.

In the source code, there is no simple *u_testctrl?clk36_pll_o. All I found from the ngd2ver is a totally new definition u_testctrl/Mmux_clk36_pll_o_Result1_1. This is quite strange to me.

And I am sure there is no command called ngc2ver now.

Thank you Muthu and Martin.

Best Regards, Kelvin

find

codes?

27.8

specify

Reply to
One Day & A Knight

In my experience, the synthesis tools do a relatively good job of keeping the net names, or something very close, but sometimes it just isn't possible. Often, some nets are duplicates of other nets, optimized out, merged, or G knows what else... Try to follow the netlist logic to see whether this is truly the net that you want (you should be able to tell by examining the other nets that join it to form logic functions). Personally, I use X's constraints editor initially to find the net, and then edit my UCF manually or with scripts. The synthesis tools rarely cause me problems by spuriously changing net names on me on subsequent syntheses barring major changes in my code.

--
Pierre-Olivier

-- to email me directly, remove all _N0SP4M_ from my address --
Reply to
PO Laprise

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.