| Register | FAQ | Calendar | Search | Today's Posts | Mark Forums Read |
|
#1
| |||
| |||
| Hi, How to understand the part: ---------------- constant simulating : boolean := false -- synopsys translate_off or true -- synopsys translate_on ; --------------- This is part of the code generated by System Generator 10.1i. It can compile through by Modelsim. I am curious about "or true" line. It is like broken code, but it works. Could you explain its meaning to me? Thanks in advance. library IEEE; use IEEE.std_logic_1164.all; use IEEE.numeric_std.all; package conv_pkg is constant simulating : boolean := false -- synopsys translate_off or true -- synopsys translate_on ; constant xlUnsigned : integer := 1; constant xlSigned : integer := 2; .... .... end conv_pkg; |
|
#2
| |||
| |||
| fl wrote: > This is part of the code generated by System Generator 10.1i. It can > compile through by Modelsim. I am curious about "or true" line. It is > like broken code, but it works. Could you explain its meaning to me? > Thanks in advance. -- modelsim sees: constant simulating : boolean := false or true; -- = true -- quartus/ise sees: constant simulating : boolean := false; -- = false |
|
#3
| |||
| |||
| Mike Treseler wrote: > fl wrote: > > > > This is part of the code generated by System Generator 10.1i. It can > > compile through by Modelsim. I am curious about "or true" line. It is > > like broken code, but it works. Could you explain its meaning to me? > > Thanks in advance. > > -- modelsim sees: > constant simulating : boolean := false or true; -- = true > > -- quartus/ise sees: > constant simulating : boolean := false; -- = false Seems to me that ModelSim does the right thing. The expression on the RHS is evaluated (and we should know that 1 or 0 is 1) at analysis time and assigned to the constant. ISE and Quartus appear to be wrong. -a |
|
#4
| |||
| |||
| On Thu, 31 Jul 2008 14:45:27 -0700 (PDT), Andy Peters <google@latke.net> wrote: >Mike Treseler wrote: >> fl wrote: >> >> >> > This is part of the code generated by System Generator 10.1i. It can >> > compile through by Modelsim. I am curious about "or true" line. It is >> > like broken code, but it works. Could you explain its meaning to me? >> > Thanks in advance. >> >> -- modelsim sees: >> constant simulating : boolean := false or true; -- = true >> >> -- quartus/ise sees: >> constant simulating : boolean := false; -- = false > >Seems to me that ModelSim does the right thing. The expression on the >RHS is evaluated (and we should know that 1 or 0 is 1) at analysis >time and assigned to the constant. > >ISE and Quartus appear to be wrong. They are all three right, because the synthesis tools must obey the "synthesis translate_on/off" pragmas. This constant can be used to generate different behaviour in synthesis and simulation. Which is not usually a good idea; though there are occasionally good and safe uses for it. (For example, to greatly speed up a serial interface to reduce wasted simulation time, or to bypass a very long initialisation delay for some external component such as a PLL, or the DLL in DDR memory) - Brian |
|
#5
| |||
| |||
| "Brian Drummond" <brian_drummond@btconnect.com> wrote in message news:n0n4941trp10iib4lfmcbsfgoc2euraa3e@4ax.com... > On Thu, 31 Jul 2008 14:45:27 -0700 (PDT), Andy Peters <google@latke.net> > wrote: > > This constant can be used to generate different behaviour in synthesis > and simulation. > > Which is not usually a good idea; though there are occasionally good and > safe uses for it. (For example, to greatly speed up a serial interface > to reduce wasted simulation time, I usually use simply it to surround... - Code that writes to files during simulation for logging purposes during debug. - To add signals that are not intended to be synthesized which are of type 'real' to make working with designs that use the fixed point package easier to debug. In either case, I'm not creating different functional behaviour, just adding additional debug capability. KJ |
|
#6
| |||
| |||
| One can use the following pragma to remove code that is only used for simulation purposes and not translated to logic during synthesis: -- pragma translate_off ... -- pragma translate_on I think you need to make sure the the pragma is placed correctly around the "or TRUE" as follows: constant SIMULATING : boolean := FALSE -- pragma translate_off or TRUE -- pragma translate_on ; During simulation the pargma is not seen and the constant becomes TRUE. But synthesis would ignore "or TRUE" and the constant becomes FALSE. Basically, I think the purpose of this constant is to EXCLUDE code for simulation, like the way you exclude code for synthesis using the translate_off/translate_on pragmas. Example could be, debugging logic that you do not want to be there in your simulation, but you want them for synthesis. FPGA examples could be the XILINX ChipScope and ALTERA SignalTap. I do not want or care about these to be in my simulation, but I need them for debugging in the lab. g_SYNTHESIS_ONLY: if ( not SIMULATING ) generate -- Put debugging code that does not show up in simulation i_chipscope_icon icon port map ( ... ); i_chipscope_ila ila port map ( ... ); end generate g_SYNTHESIS_ONLY; Cheers, -- Amal |
![]() |
| Thread Tools | |
| Display Modes | |
In an effort to better serve ads to our visitors, cookies are used on objectmix.com. For more information, check out our Privacy Policy.