Pictures are spread over 3 pages. Page 1,
2, and 3. If you
want to have a full sized copy of the picture (2048x1536 resolution) just
It was a great semester, thanks for all your hard work! We hope that it was
a worthwhile experience. Please remember to e-mail any comments to the address
||Muwhahaha! Tremble in fear, mortals! Grading is done!|
- We've entered the grades into Bearfacts so they should appear whenever it decides that you deserve to see them.
- Glookup has been updated with final scores for all of John's section. Kurt's and Jack's should follow soon.
- The group entry in glookup has been left blank (in John's section at least) because it is partially determined by your group members' evaluations of you as well as the TA's discretion.
- Finally, please send mail to cs152-staff AT cory DOT eecs DOT berkeley DOT edu if you have any ideas on how to improve the class. Also please let us know about any boneheaded mistakes that we made, or horrible design decisions in the overall course. Remember, if you don't let us know, then the students of next semester will be doomed to suffer the same fate as you.
||Make sure that you email your review of your group members for the whole semester to your TA as they described in their emails to you.
||MT2 Solutions are up.|
- Your TA has emailed you instructions about how to complain about grades, Please make sure that you
read that email.
- Update: I wrote down the wrong point values for some of the questions. Part 1E is out of 8 points total.
Part 1F is also out of 8 points. The maximum perfect score for each question should be moved from 5 to 8, but
the partial credit scores remain the same.
||More on the final report|
Your final report should include: a description of your project, going into slightly more detail than
your oral presentation did. Feel free to talk about the history of your processor or anything else you feel is
relevant. There should also be an extensive performance section, with more quantitative
numbers than you had presented in your oral report. Some good topics to include are how your choice of
modules improved (or deproved) your performance. If you can quantify these by varying parameters (such how
how big your ROB might be), that would be good as well. The accuracy of your predictors for certain tests is
also a good number. We are looking for in depth understanding of the tradeoffs that a processor designer has to make.
You should also include a section on testing, and test results, as well as your online notebooks.
Your website should also have your supplemental material. (zip up what you would have submitted via the submit program
and upload those to your website). Supplemental material should include verilog files, schematics, etc.
Your group evals are due by e-mail to you TA by Saturday midnight as well.
You will need to post your final reports on your website by Saturday midnight. Pictures and final racing
results will be posted soon!
Presentations before lunch will be in 606 Soda. Presentations after lunch will be in 651 Soda. We will have a projector and a computer.
||Problems on board?|
This file will allow you to link 4 DLLs together to create a divide by 16. There may be a bug with the files on
the instructional accounts which do not allow you to overwrite a certain Xilinx file(despite your comment to divide by 16),
so you effectively can only divide by 2. If you suspect that your design doesn't work on board due to a clock issue,
you may want to try downloading this file. Thanks to Yang and Ilya for discovering this.
||Notes for the presentation|
Sample Presentation here
This presentation is very important and is a significant part of your grade, so make sure you do a good job.
||Notes for John's Sections|
For the final hand in, aside from the guidelines laid out in the final
project and the example lab report, make sure that you include some form
of automated testing. A lack of automated testing will cause you to lose
points. Also make sure all the elements of your memory-mapped
I/O module perform correctly. Including output in simulation.
We will be doing individual checkoffs
on Monday, so e-mail your TA to set up a time. The presentations on Tuesday will
go from 8:30-5. Each group needs to sign up for a 30 minute (23 minutes presentation, 6 minutes Q&A,
1 minutes set up) slot. A signup sheet will be posted on Dave's door around 4pm today.
Good presentations will cover the
specific sub-projects you chose to implement, and how they affected your processors performance.
Be sure the include analysis of your performance. One example might be to report your prediction
accurancy for various benchmarks. Detailed descriptions of your project datapath are not
appropriate for a 20-minute presentation. However, high-level data paths might be appropriate.
A sample powerpoint presentation will be posted soon. Everybody in your group must talk during the presentation.
||There has been an update to the memory-mapped I/O module in the final project.|
||Improved mipsasm and boot scripts|
- We've written a web-based mipsasm
- This version has some assembler directives that the old version lacked.
(Like the ability to re-define the base of the address space
on the fly and the ability to get the # of instructions between any
two labels -- very useful for the boot0 blocks.)
- There are undoubtedly some bugs in it. If you find any, email them to Kurt.
- Here's a sample boot script. It has the
ability to run different tests based on the dipswitches. Right now, it runs
basic and corner from lab 4, but feel free to add your own tests. This will
save you some time -- you won't have to re-synthesize every time you want to run
a different test. (Note: I haven't tested the assembly yet, so there may be
bugs. Use at your own discretion.)