SystemVerilog Classes 8: Constraints

[@anastasiacunningham]

This course series was really concise and informative! Thank you!

[@Duramen131]

What a good primer series. Definitely one I will recommend to new verification engineers. Thanks for posting it

[@hamzaelhawary6808]

Indeed this was very useful. Simple and to the point. Thank you

[@gantisubash7901]

The course was very informative and well delivered. Thanks!

[@tientranmanh798]

Thank you so much. Your video is extremely helpful for me.

[@kakakapo8225]

It is incredibly useful , Thank you Brian !

[@rgarciaf071]

Really good and concise!

[@astghikavagyan1119]

Thank you for a very useful class!

[@mohammedkhatib6290]

Thank you Brian! it was so useful :)

[@LowMelaninLife]

It was very very useful, thank you very much! :D

[@yufan-yang]

Hi, Thank you Brian ! it is really useful. Would you mind to share the ppt ? Thanks

[@S_P_S]

That was a wonderful explanation. But I have a doubt. In the conditional constraints part, the first method (without the if-else method) suggests if the first part occurs then the second part only occurs. Is it reversible?

[@mr_official_tech7734]

Can you please explain how to achieve randomization without using Rand key????

[@ishwarbannur7228]

use builtin construct $random to assign a random 32 bit variable

[@briandickinson7771]

If a property of a class is declared without the rand modifier, it can still be randomized by passing it as an argument to the randomize method. However ONLY the properties passed to the randomize method are randomized. So if you want to randomize both rand and non-rand properties, you may need two calls (see below). Also a non-rand property does not have a rand_mode switch. e.g. class randclass rand bit[1:0] randp; bit[1:0] staticp; endclass randclass c1 =new(); initial begin ok = c1.randomize(); // randp randomized ok = c1.randomize(staticp); // only staticp randomized ...

[@thamizhazhagi2627]

vlsicsinfo.blogspot.com/2020/07/how-to-randomize-non-random-variable-in.html?m=1

[@gurunamsinghpanwar5080]

what is the benefits of distribution constraint and what is the benefits of giving the weight to our random number in constraints .can you explain this..

[@briandickinson7771]

Without distribution, randomization probability is uniform, i.e. every possible value is equally likely. With dist, you can change the probability of values to make them more or less likely to appear in randomization. Good example would be the parity of a data packet. If you declared a single bit switch to control parity (1 = good, 0 = bad), and randomized the switch without constraints, then you would get 50% chance of bad parity. Maybe for your verification you only need 10% of packets to have bad parity. Therefore you could use a dist constraint to set the probability of the switch to be 90% good(1) and 10% bad(0) as follows: {switch dist {0:=1, 1:=9};} Of course you'd probably want to use an enumerate for the switch type, but that's a different story..

[@attila3028]

It is useful, thank you.