Note: Class taken remotely during COVID-19.
tl;dr: this class is A LOT OF WORK for just 2 units. esp remotely. last few projects were difficult and we had to submit long reports AND record videos for each project. grading wasn't particularly harsh though.
This is the first time in a while where Sarrafzadeh is in charge of this course. However, you will not see the prof at all aside from OH. The TAs are in charge of your course. Honestly the TA does not matter too much as projects are all graded by the same TA for all students regardless of TA section.
I only attended my section for the first two weeks and did not go to any of the rest of the TA sections. I didn't find the sections very helpful and find learning on my own, i.e. googling, for the projects was faster.
There were in total 5 projects all in Verilog. All done individually(which kinda sucks). You have to download the Xilinx ISE in some form and I strongly recommend just dualbooting linux instead of using a virtual box which takes a lot of memory and is quite slow. First three projects weren't too difficult, but last two projects were VERY TIME CONSUMING. They're not incredibly hard, it just takes time to understand the vague specs, generate 100000(im exaggerating but you get the point) testcases, screen shot all the waveforms, write the long report, record the video...
Each project took me approximately a day(writing report+video generally took me around 3+ hours). It is quite a lot of effort for a 2 unit class.
The TAs aren't particularly harsh on the grading, but just make sure you follow the instructions carefully. Missing an explanation could cause you 5+ points out of 100.

Did I learn a lot about Verilog from this class? YES.
Was it worth the time? I am not sure.

This is a review of the M152A class, which he's in charge of but the TA's run. As all labs, your experience mainly depends on your TA. I got a pretty good TA. His exams and grading policy were all very fair. While the concepts themselves aren't that difficult (as long as you kind of get what happened in M51, you should be find). However, the major difficulty is the shoddy equipment + software you have to work with. It takes FOREVER to compile (you spend most of your lab time staring at the screen waiting for it to compile). Also - the software and/or transfer from the program to the hardware is buggy. Sometimes you can code/design it correctly, and have the waveform test show correct results, but the FPGA doesn't work somehow so you'll have to redesign it so some bug in the system will work. This is incredibly problematic for projects that require that the timer we build is accurate. Mainly we had a TA who understood this and wasn't incredibly stringent on a little deviation.