wyvernSemi
commented
1 year ago Hi Jim,
regarding the -interceptcoutput option, I think it's okay to to have this all the time, as it can only cause a check when running VHPI and an additional thing needs to be added to the C code (setvbuf(stdout, 0, _IONBF, 0)) to make it work, which won't be done when not using co-sim code.
On the stream issues, I haven't had much time to look at this fully these last few days, but your last e-mail has given me an idea where I might look for a potential point of failure (though this doesn't fully explain why it works on some tests). Just to clarify the questions you asked, the code is compiled just once and then run for all the generic combinations unmodified. My test did have random data, but I tried with incrementing data, as per the Ethernet test, just in case it was some data dependency, but it failed in the same way. I'm hoping to get back to this next week, though I acknowledge that your availability is now limited for the next fortnight.
best regards,
Simon
On Fri, 3 Mar 2023 at 18:55, Jim Lewis @.***> wrote:
WRT -interceptcoutput, do we always want this on or just when doing co-sim? IE: if we always use it, will it impact simulation performance?
— Reply to this email directly, view it on GitHub https://github.com/OSVVM/OSVVM-Scripts/pull/41#issuecomment-1453967480, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFHTY32OFTLD6MJZC2ABGCLW2I5BVANCNFSM6AAAAAAVLYQ2OM . You are receiving this because you authored the thread.Message ID: @.***>
WRT -interceptcoutput, do we always want this on or just when doing co-sim? IE: if we always use it, will it impact simulation performance?