CS 250: VLSI Systems Design

Annoucements:

5-23-97
Semester lab grades can be seen here. Please check for grading mistakes and always mail both Will and Randy about them.
5-16-97
For those who are turning in their final project late, please be sure to get it in by Tuesday, noon.
5-15-97
Get new cadencesucksbig from .../final/bin. Also, you may recopy Cadence library rcore if you have problems extracting.
5-15-97
End of semester request for student feedback by Prof. Wawrzynek. Reponse would really be appreciated!
5-15-97
Get new .cdsinit from .../final/dotfiles.
5-15-97
Get new full.VCMD from .../final/netlist.
5-14-97
Bug fix.
5-14-97
Appendix A of Lab 9 reviews some of the things covered in the last lecture.
5-12-97
The files in .../labs/final/netlist/ Makefile and nocntl.VTRAN have been updated. The check target in the Makefile should now work.
5-11-97
The file c_cpucntl.inc in the RTL has been updated. Download it from .../labs/final/.
5-9-97
I forgot to mention something rather important in lecture. You need certain dotfiles in your root directory for cadence to extract correctly. Some of you may already have good ones, but just to be sure, please get all the dotfiles from .../labs/final/dotfiles.
5-5-97
You should probably get the new RCORE RTL from Will for the final project. To do so, run /disks/test0/a/cns/cs250/labs/final/get_rcore from your NOW account in an empty directory. If (and only if) you are done with lab 8, please mail Will with the subject line `lab8done' to get the password that you must enter. Will is trying to get his procmailrc setup to reply automatically, but it refuses to work, so please be patient as he hunts around for procmail experts.
4-23-97
Nasty Bug in RTL: This bug rarely shows up, but it is really awful when it does. Both modules rcore and sram receive a signal called clkin, which they both invert to make phi. This is bad because, while both modules use phi as its clock, the two phi's are not synchronized to one another. Because the behavior is unpredicatable on Verilog, the simulator may fail when it is uninstrumented, and then may work fine when you add in statements to trace variables to find out what went wrong. The way to fix this is to generate phi as the inverse of clkin in the module all, and then pass phi to both rcore and sram.
4-18-97
Prof. Bose will no longer be responding at mmc.net since he no longer works there. A new email address will be given very soon. Meanwhile please send midterms, hwks, and questions to cs250-ax@durer.cs
4-6-97
One more lab7 mistake: signals branch and branchb should have been in On_critical instead of Op_critical.
4-4-97, 4:35pm
Update to lab7 just made again. Please reload cpucntl.script.
4-4-97, 11:40am
Update to lab7 just made. Please reload cpucntl.script.
3-31-97
Regarding lab 7: It will be out whenever Will finishes it, which is probably not earlier than Thursday.
3-27-97
An important message from Prof. Bose regarding the midterm cheating.
3-21-97
The computer account form is processed. It is the same password as your instructional account.
3-21-97
The license file has been updated.
3-20-97
Lab 7 will be going out on Monday. Enjoy the Spring Break.
3-20-97
Please turn lab1-6 in by next Friday (March 28) if you have not already.
3-20-97
The cs250 class has access to the NOW cluster. Here is how.

List of past announcements

Lecturer:

B. K. Bose

GSI(s):

Will Chang (willyc@cs.berkeley.edu)

Lecture TA:

Michal Siwinski (cs250-ax@durer.cs.berkeley.edu)

Time and Place:

Lecture
Tu 11:00-12:30, F 9:00-10:30 310 Soda Hall

Lab Lecture
Th 11:00-12:30 Place TBA

Office Hours
TBA (Prof. Bose)
TBA (Will)

Labs:

The lab exercises are designed to teach a commercial CAD tool design flow, reinforce concepts from the lectures, and look at processor design at the circuit and layout level. The framework for the labs is a custom MIPS implementation. This implementation includes an ISA simulator, an executable RTL description, a transistor-level netlist, and complete layout. As a course project, students will be provided with an incomplete netlist and layout and asked to first complete, and then improve the design.

The project is spread out over the semester and interleaved with tutorial material. A weekly lab lecture will cover the basic principles behind the CAD tools and review processor design. Tools demos will be held in the lab lecture or made available online.

Schedule:

Lab. Date Description Due Soln
1 22-Jan Verilog Tutorial: Shift and Add Multiplier Design 1-Feb
2 30-Jan MIPS CPU Pipeline Tutorial: Bypass Control Logic 13-Feb
3 5-Feb Synthesis, Static Timing, and Place&Route Tutorial: Shift & Add Multiplier 6-Mar
4 1-Mar Schematic Capture and Verification Tutorial: Multiplexer 6-Mar
5 6-Mar Layout Editor and Design Rule Checker Tutorial: Multiplexer 13-Mar
6 13-Mar Shifter Implementation 20-Mar
7 3-Apr Synthesizing RCORE Control Logic: Part I 10-Apr
8 16-Apr Synthesizing RCORE Control Logic: Part II 24-Apr
9 24-Apr Final Project 1-May, 15-May

Grading:

Each lab is graded on a scale of 1 to 10 and assigned a weight proportional to the number of weeks it is scheduled for. Lab grades are 45% of your final grade

Feedback:

In each lab, you will be asked to turn in feedback on working conditions such as computer resources, the time it takes to complete the assignment, cad tools, and the effectiveness of the assignment. This constitutes 10% of your lab grade.

Machines:

List of machines available in 347 and 349 Soda Hall.
Free machines. Thanks to Jay Lorch.

News:

ucb.class.cs250
comp.lang.verilog
comp.cad.cadence

Lab Lecture Notes:

Design Representations and Flow
Little T0 Tour

Interesting Links:

Cadence
VLSI Libraries Incorporated
PowerPlay

Other Semesters:

Spring 96