CS61C: Machine Structures. Fall 2007, UC Berkeley
Patterson and Hennessey Kernighan and Ritchie
CS61C Fall 2007
MWF 3pm-4pm 2050 VLSB





DateOld News
12/13/2007 Final Office Hours

Greg will be holding what are likely to be the final office hours of the semester from 8am-10am Friday, 12/14 in 271 Soda. This is your last chance for both technical and grading issues.
12/12/2007 Exam2 Grading

If you accidentally interpreted the addresses in problem 1 as decimal instead of hexadecimal and lost significant credit because of it, please submit a regrade request.
12/11/2007 Grading Status Updates

Keep an eye on the Grading Status page. We are busily grading so don't be surpised to find your score changing, or extra autograder logs appearing. Please check the Grading Status page before you report any "problems."
12/10/2007 Grading Status Updates (Updated 11pm) 

Regrades are in progress.  Do not be surprised if you receive the relevant log file long before your score is changed.  If you receive such a log file, you will eventually receive a new score with a comment mentioning the regrade.  Questions like "when will my score change" will be ignored.
  • All assignments now have grading deadlines posted.
  • We have clearly spelled out the regrade policies for all projects and homeworks. You need to read this before submitting one!
  • Greg will be handling all homework regrades.
  • The proj5 autograder had some troubles and has been re-run several times, we will extend the regrade deadline if final logs are not availible later today.
  • Exam2 will be returned today in lecture.
12/8/2007 Final Week

Just to be clear on scheduling for this week:
  • There is lecture on Monday, 12/10. It will be the final lecture.
  • Homework 12, a survey, will be posted on Monday.
  • There is no discussion or lab. 
  • There is a final exam review on Tuesday 12/11 from 4-7pm in 10 Evans. From 7-8pm, TAs will be availible for individual questions.
  • Office Hours:
    • Greg will have his Tuesday & Friday office hours as normal, further hours may be availible by appointment if you have an emergency and he has a bad habit of hanging out in 271 at random times.
    • Peter will have office hours Tuesday 12/11 from 12-2pm in 511 Soda.
    • Alex will have office hours Monday as normal.
    • Valerie will have office hours Monday as normal.
  • Regrade deadlines can be seen on the Grading Status Page. This contains all the grading information we know to date. Do not e-mail us questions which this page answers. Please direct issues to the appropriate reader as listed on this page.
  • The final exam is Saturday 12/15 from 12:30-3:30pm in Bechtel Auditorium.
12/8/2007 Homework12

There will be an twelfth homework. It will be a survey, and it will be released on Monday 12/10.
12/7/2007 Exam2 & Exam3 Updates

We've posted links to our cache (not that kind) of old CS61C final exams. You may also wish to look back over old midterms, since the Exam3 is cumulative.

Exam2 e-mails should go out Monday. Depending on how long this takes they will be returned either Monday at lecture or Tuesday at the review. We apologize for the delay, a miscommunication among the staff prevented their return by today.
12/6/2007 Final Lectures

Be sure to attend the final two lectures. Tomorrow, Prof. Wawrzynek will cover parallel programming and HKN will collect course reviews. On Monday Greg will be giving a mystery lecture to wrap up the semester.
12/6/2007 Grading Status Updates

  • All quizzes have been graded.
  • All autograded homeworks have been graded.
  • Proj5 will be autograded after Saturday 12/8 at 2:45pm to accomodate people with slip days to use.
  • Any issues should be e-mailed to the person listed on the Grading Status page as being responsible for the assignment.
  • We're hoping to distribute exam2 results this weekend, and return exams early next week. We apologize for the delay.
12/3/2007 Proj5 Updats

A number of people are still unclear and justifiably worried about the accessCount and LRU parts of proj5. As such we've added a section entitled "AutoGrading Corner Cases" which contains answers to the most common questions. We've also updated the comments yet again to reflect these notes.

Finally, as a few people pointed out the LRU scheme in our solution had a bug and definitely did not work correctly. Good catch to those students. In addition, our solution was not properly initializing all the data words of the cache to 0 before running, again good catch to those students who noticed. Keep an eye out for the next autograder run today at 2:45pm.
12/1/2007 Proj5 Minor Update

For those people struggling to understand accessCount and when it should be incremented to match the autograder, we've updated the comments in proj5.c slightly.
12/1/2007 Proj5 Early Autograder

Given how few people actually managed to get something submitted for the early autograder run yesterday, we've decided to run it again on Monday 12/3 at 2:45pm.
11/30/2007 Final Exam Time & Place

The
final exam will be held Saturday, December 15th from 12:30-3:30pm in the Bechtel Auditorium. For those of you in CS61CL, contrary to BearFacts, your final is the same time and place as the rest of the CS61C students.
11/30/2007 Quiz37 Update

If you are confused as to the difference between 1MB (1,000,000B) and 1MiB (1,048,575B) you may wish to read this: Wikipedia: Binary Prefix. We will be lenient on the different when grading this quiz, and we will be more clear (as you should be) in the future.
11/30/2007 Proj5 Update

We updated some of the comments in memory.h both to fix typos left over from old versions of the project, and to explain some of the struct fields a little better. In particular, the least significant bit of the cache tag, should appear at the least significant bit of the tag field, in order to avoid autograding problems.
11/28/2007 Final Exam Review

We now have a time and a room for the final exam review. We will post the actual final exam room shortly. As always, let us know about conflicts or problems ASAP.
11/28/2007 Proj5 Early Autograder

We apologize for the late announcement, but we will be running the proj5 early autograder Friday 11/30. Submissions before 2:45pm will definitely be processed, but we make no promises about later submissions. As before, any bugs in the autograder will result in a second run, most likely on Monday 12/3.
11/26/2007 Proj4 AutoGrading Update

During one of the autograder rebuilds, some files got left behind it would appear, resulting in the beq_tback test being run twice, once for 1 point and once for 2 points. We have corrected this and are re-running the autograder as of noon today. You results and score shouldn't change by much.
11/26/2007 Proj5 Update

There was a small error on line 18 of proj5.c. The code struct row * should have been struct set *, we missed this as we updated the project from past semesters.
11/24/2007 Regrade Requests

Project regrade requests for proj3-5 must consist of a list of the files, the line numbers within those files and the changes you wish us to make to those files to bring your submission in line with the autograder's expectations. Regrade requests of the form "can you look at my project" and "can you explain the autograder log to me" are not acceptable. We will, however, be more than happy to talk to you during office hours to help you formulate an acceptable regrade request. Figuring out how to make these requests and what changes you need to make is actually a great way to study for the final exam in this class.
11/21/2007 Thanksgiving

For those of you under a rock, Thursday and Friday this week are U.C. Berkeley holidays. That means no lab tomorrow, Thursday 11/22, no office hours and no lecture on Friday 11/23. There is lecture today, Wednesday 11/21 and then again Monday 11/26. Enjoy your holiday.
11/21/2007 Grading Status

There are a number of grading updates as we shuffle things around to compensate for the issues caused by losing a reader.  We'd like to apologize yet again for the inexcusable confusion this has caused in the grading process. 
  • HW8 will be AutoGraded.
  • HW11 will be AutoGraded.
  • Proj3 has been regraded (in progress as of 9am 11/21) to account for slip days. There will be no human reading of this project. Regrades should go to cs61c-ta.
  • Proj4 has been regraded (in progress as of 9am 11/21) to account for slip days and to fix a couple of minor autograder bugs. Regrades should go to cs61c-ta.
  • Proj5 will be AutoGraded only.
Note also that we have posted regrade deadlines for all of the assignments which have been or will be autograded. As we get more human read assignments back to you, the regrade deadlines for those will appear.

No regrades submitted 12/13 or later will be accepted for any reason whatsoever.
11/20/2007 HW11 Posted & AutoGrading

We've posted HW11 which will be due the week after Thanksgiving. Please make sure you conform to the template.txt file, as we will be autograding this, in the hopes of helping to catch up on the grading.
11/19/2007 Proj4 Graded (Updated 11/20/2007)

There was a bug in our solution processor causing it to ADD instead of OR for the oir instruction. This appears only to have affected the matrix test. We are currently re-running the autograder to account for this, and any slip days you may have used. You log e-mail may take up to 24 hours to appear.

In an effort to deal with the fact that we've had 1/3 as many readers as we should have for much of the semester, we will only autograde Proj4 and Proj5, no reader will be involved (they're busy with homeworks and earlier projects). Please check your e-mail and glookup. Regrade requests should be sent to cs61c-ta. Aside from obvious autograder mistakes, if you failed a wide variety of tests (for example because lui didn't work properly) and you can fix this in only one or two lines of code, you should include those changes, and what line numbers in what file to change. We reserve the right not to accept changes we deem, too large, but this may help you get points back.  Deadline for these requests is 11/28.

For those people who used slip days, they will be applied, give us a day or two to do it though. Consider also letting us know why you used a slip day, by posting to the newsgroup. Many people used them for this project and we'd like to know: was it too hard or did it simply overlap with midterms and projects from other classes? Most people we've heard from said the latter.
11/15/2007 Proj5 (Updated 11/14/2007)

The files and description for proj5 should be fully debugged, as we have completed our solution.

We've posted proj5. It will be due 2:45pm on Wednesday 12/5. Please note that we have not yet finished our solution, and there may be some updates forthcoming in the next few days. Please be sure to report any bugs you find in the assignment on the newsgroup.
11/14/2007 Grading Status Update

Okay, so we've gotten another reader, and we're working to clear up the grading situation. In particular one of our readers dropped off the face of the earth for a month and still hasn't completed any real grading. We're doing our best, so please be patient.

All of the quizzes to date have been graded completely. Many people's proj2 have been graded, and we're working on proj1 and proj3. Our new reader will be helping to clear the homework grading backlog.

We will be getting all grades back by 12/15 (the final) or heads will roll. All regrade requests of any kind must be submitted by 12/15, except for those assignments which are not graded by that deadline, for which we'll come up with another policy. Please submit the request to whoever actually entered your grade first, and in the event of no response contact your TA.
11/9/2007 Exam2 (Updated 11/14/2007)

The exam2 review will be held 4-7pm in 2040 VLSB.

We've posted some basic information to the new Exam2 page. The exam is a week and a half away!

In particular the second Midterm will be held on Tuesday, November 20th, 2007 from 7-9pm in 2050 VLSB, the same room the lectures are in. If you have a conflict it's quite late to notify us, but much better now than a week from now.
11/8/2007 Proj4 Blocks Update

The way our that Block.s printed information cause many people with excellent proj4 implementations to fail several autograder tests. We have updated Block.s, and will use the updated version for the next early run, mentioned below.
11/8/2007 Proj4 Early Autograder

The first early autograder run is complete, and you should be able to see your results. Please report any suspected bugs in the autograder ASAP. We will re-run the early autograder not before 12am (midnight) Friday night. Results should be availible early on Saturday.
11/6/2007 Proj4 Early Autograder (Update 8:30pm 11/7/2007 & 10:30pm)

We have gotten the autograder running. Results should appear in your inbox in a few hours. It is still very likely that we will re-run the autograder on Friday after 2:45pm. As of now, the proj4 autograder will not run properly on the instructional cluster.  We are working with instructional support to resolve the problem, and we hope to be able to run the autograder by Friday during lecture.  Submissions received by 2:45pm Friday 11/9/2007 will be autograded and results should be availible later that evening. Our sincerest apologies to anyone who rushed to meet the deadline today.

If you submit your proj4 solution, however partial it may be, by 2:45pm Wednesday 11/7/2007, we will run a partial autograder on it, and send you the results. We will probably not be able to return results before 7 or 8pm as running ModelSim on many projects may take a while.

The cutoff of 2:45pm is to strongly encourage everyone to attend lecture that day. As a note, if significant bugs are found in the autograder, we will re-run it Thursday around 3pm.
11/3/2007 Quiz27 Solution Update

It was pointed out to us that our solution to Ex1 was wrong, and that the solution to Ex2 needed some explanation. We've updated the solution to compensate. We apologize for any confusion.
11/1/2007 Proj4, Blocks.v & CPUTestBench.v Updates

We have added a number of helpful notes to the Proj4 description. Nothing major, but if you find yourself confused consider refreshing the project description and checking if we've already answered your question.

There were some minor typos in the comments of Blocks.v. We've fixed them.

There was also a small bug in how CPUTestBench.v printed BreakCodes. We have fixed this too.
10/31/2007 HW8 & Verilog Circuit RulesTweaks

In response to some last minute questions (we suggest you start a little earlier next time) we've clarified a few points of HW8:
  • HW8 Ex1: You may only use primitive gate instantions in your code.
  • Verilog Circuit Rules: we have added simple explanations of the three subsets of Verilog, and general rules about when you may use them.
10/30/2007 Proj4 Tweaks

We wrote the solution to proj4, and in the process tweaked a few minor things about the project. Below is a list of changes you might be interested in. If you haven't started the project yet, don't bother reading this list.
  • We added a note to the effect that you may freely change the interface of the Control.v module (but you shouldn't need to change any others,let us know if you feel otherwise).
  • We added a 1bit register to the Blocks.v file for those who might want it.
  • We've added a hint to the effect that your first test program should test nothing but the break instruction.
  • We noted that proj3 and proj4 implement nearly identical instruction sets, which you might find useful during testing.
  • We've fixed a number of small bugs in the given code, from reordering the ports for Control.v to fixing a few typos in Blocks.v.
  • We will be running the autograder (or a subset thereof) at least once before the deadline allowing you to see your results and get in some extra debugging.  We are aiming to do this on the evening of Nov7th more details as follows.  The more or your processor you've completed by then, the more useful your results will be.
10/29/2007 AutoGrading

We ran quite a few autograders this evening: quiz23-25, hw7 and proj3. You should check the Grading Status page, and contact cs61c-ta@imail if there are any autograder mistakes on quiz23-25 or hw7. As usual major autograder mistakes on proj3 should be reported to cs61c-ta@imail, but slip days and full grading will be accounted for by the readers.

In other news at least some of the reading situation is being sorted out slowly but surely.
10/29/2007 Verilog Circuit Rules & Green Card

We've split up the "Verilog Crib Sheet" into the Green Card, which is a short reference and the Circuit Rules, which detail the things which will not be allowed in CS61C. Note that the Circuit Rules may be further restricted on certain homework, project or exam problems, but we will be very clear about this. If you are ever in doubt, feel free to double check with us.
10/29/2007 Calendar Updates

We've shifted the lecture schedule listed on the calendar so that it matches reality over the last few weeks.
10/26/2007 Proj4 Update

We updated the Block.v file for proj4 to ensure that the register file will work properly when a register is read and written on the same clock cycle.
10/25/2007 MARS Update

For proj4, we have updated mars-cs61c. In particular you will find that there is a new dropdown next to the Save Dump button. This dropdown selects the format of the dump file. Binary should be used for proj3 and Hex Listing for proj4. The instructions in both projects have been updated accordingly.
10/25/2007 Calendar Changes

We've finally posted proj4. In order to bring the project and lecture schedules back into line, we're making the following deadline changes:
  • proj4 will be due 11/14/2007 instead of 11/7/2007.
  • proj5 will be due 12/5/2007 instead of 11/28/2007.
  • quiz25 will be due Monday 10/29/2007 instead of Friday 10/26/2007.
10/25/2007 Verilog Crib Sheet (Updated 10/29/2007)

We've separated and renamed the two components, into the Verilog Circuit Rules & Green Card.

We're created a Verilog Crib Sheet, which we hope you'll find helpful. In addition to being a short, printable resource which you might use on an exam, it is the definitive reference on which constructs are allowed in circuits, and which are restricted to testbenches.  As with all such resources you'll find a link to this on the Resources page.
10/24/2007 Assignment Deadline Extensions

The three assignments due today, 10/24/2007 at 2:45pm have been extended due to the number of random technical problems on the instructional computing machines. Please use glookup -t to confirm your submissions have been logged, just in case.
  • Quiz24 & Proj3: Both have been extended to Wednesday 10/24/2007 at 7:45pm (same day, just 5 hours later).   Note that proj3 slip days will still be computed from the normal 2:45pm deadine.
  • hw7: Extended to Friday, 10/26/2007 at 2:45pm (48 hour extension).
10/23/2007 Proj3 Updates (Updated 10/24/2007 11:15am)

One enterprising student discovered a weird bug in sim.c that seems to only appear when testing your project using the Microsoft Visual Studio C libraries (not gcc on the instructional machines) it turns out we were opening the .dump files as if they were text! sim.c has been updated. If this causes a problem with your project somehow, relax, and don't waste a slip day. Let us deal with it during grading instead.

We have posted the output from our proj3 solution while running sample.dump, for those of you who are paranoid about formatting errors.  It is availible as sample.output in the project files directory.

There is a post below which mentions that you should check for lw and sw accessing memory illegally, but the truth is that any memory access should be checked. However, you needn't check for a vaid PC because the code in computer.c will do this for you as it calls contents() to get the instruction from memory (line 64).
10/23/2007 Grading Status Page

Rather than answer questions all the time, we've created a grading status page to let you know how the readers are doing with your assignments. There are some questions marks as the information gets pulled together properly. We're hoping to improve on this dismal situation quite soon.
10/23/2007 Exam1 Grades (Updated 10/23/2007 7:45pm)

Exam1 scores have been entered into glookup, and grade breakdowns mailed to your course mailbox. Regrade request deadline is 11/2 (no exceptions), instructions are in the e-mail. If you have not received a grade breakdown e-mail, send mail to cs61c-ta (Greg Gibeling) ASAP. We have also posted summary statistics (mean, standard deviation and histogram) to the Exam1 page.

The answers, and grading key have been posted to the Exam1 page.

Barring unforseen problems exam1 grades will be in glookup tomorrow along with an e-mail to your course account showing your grade breakdown. Paper exams will likely be returned in lab on Thursday, at the end of your lab section. This news item will be updated as we know for sure.
10/22/2007 Proj3 Error Message (Updated 10/23/2007 10:30am)

It was pointed out that our sample.s and sample.dump files illegally accessed memory in a way which would likely overwrite $0. We've fixed that. We also added an alignment check to contents() in computer.c.

In the event that a sw or lw instruction accesses memory which does not exist you should display exactly the same error message, and call exit(0), as in the recently updated contents() function in computer.c.
10/18/2007 Lab8 & HW7

Because we cut the parallel sorter exercise from lab8 in order to reduce the length of the lab and allow people to focus we have made that exercise into HW7. This homework should be very easy after you've completed lab8 and we will be grading it very loosely: any reasonable attempt at a solution will be accepted. HW7 is intended to give you more practice with Verilog and ModelSim and give you a chance to setup up ModelSim from home if you intend to use it for proj4.
10/16/2007 ModelSim Tutorial (Updated 10/18/2007 8:00am)

We made some minor edits to the tutorial, including listing the sections which you need to read (see the introduction) and fixing the website URL, which has since been bought by an company unrelated to ModelSim

We have posted the tutorial for ModelSim to the Resources page, the Verilog simulator which you will be using on Thursday in lab. You may wish to wait until after lecture tomorrow, but you should read this tutorial before coming to lab, as it will save you time to have even glanced through it. While reading: you don't not need to know anything about RDL or Xilinx libraries. This tutorial is more widely used than this class.
10/15/2007 Exam1

The first midterm for CS61C Fall 2007 will be held tonight Tuesday 10/16/2007 in 155 Dwinelle from 7-9pm. Seating is assigned, so be prepared to come 5min early, and bring your Cal Student ID card with you.

The midterm is open book, open notes and you may use any written material you like during the test. Electronic aids of any kind of strictly prohibited. The midterm will cover material from the start of class through lecture on 10/5/2007.
10/15/2007 Proj3 Update (Updated 1:00pm)

There was a second fix to this line, and we've improved the notes in the project description about the exact ouput from disassembled by giving a short example.

We've made a minor bugfix to line 101 of computer.c in order to avoid segfaults when using the -m command line option. The two base addresses on this line have been changed from 0x0040000 to 0x00400000. Please make sure you have the latest version.
10/13/2007 Quizzes, Reading & Schedule(Updated 10/14/2007 8:30pm & 10/15/2004 1:30pm)

The reading has been adjusted yet again. Quizzes 21 and 22 will be up as soon as we're sure we've got the material sorted out correctly. We apologize for the changes, we're trying to improve this section of the course over past semesters.

We adjusted the readings for this coming week. In particular there are a few more for tomorrow, which you should try to get through. Obviously we're posting them a bit late, and there's a lot, so you may not get through them all before wednesday.

We've updated the lecture schedule, reading lists and posted the quizzes for next week. We apologize for the last minute change, but we miscalculated lecture positions before the semester started, there is a lot of information to be learned next week and it's important for us to get it right.
10/12/2007 Proj2 AutoGrader & Network Outtages

Some off confluence of events on Thursday appears to have left some of the autograder files in an odd state, and our re-run that evening 1) had errors in the test output and 2) did not regrade all students. Rather than worry about how and why, we're rerunning the autograder now. Sometime tomorrow or Monday please double check that 1) you received an e-mail from the autograder, and 2) the autograder test cases are behaving properly.
10/11/2007 Proj3 Update (Updated 10/13/2007 10:00am)

As a student pointed out to us, you'll need to manage the memory for any string returned from disassembled().  We've added a note to the webpage and proj3.c.

We updated computer.c for project 3, it had a bug in it which basically guarentees that your project 3 would not run. As a consequence it's a fair bet that few people have started project 3. I assure you that the solution material for project 3 will be covered on the midterm.
10/10/2007 Quiz Grading

In order to actually have time to properly grade homeworks and assignments, and to get assignments back sooner we will no longer be automatically reviewing imperfect quiz scores. Quiz16 and later will be autograded and we will only review your score if you have a specific complaint, and can explain why you deserve points back.
10/10/2007 Exam1 Webpage

Above is a link to the exam1 (e.g. the midterm) webpage. Please read it before you come to the exam. Furthermore if you would like to request that the Friday review cover certain topics, please post to the course newsgroup. The more such posts, the better the review will be.
10/10/2007 Project2 AutoGrading (Updated 10/11/2007 4:05pm)

We are re-running the autograder. You will receive your new score in ~18 hours as the proj2 autograder takes quite a while.  The re-run is because the test.1.hex_over.out file generated by our solution was wrong (the solution had a bug) and because we forgot to test for certain register initialization conditions: namely that you initialized registers properly before use.  Many scores will go up because of the first, and a very few will drop because of the second.

Project2 has been autograded, and the results have been sent to the readers. It's unlikely you'll see reader adjustments before the midterm. Rest assured, as you examine your score, that everyone did pretty poorly on proj2. Our solution is the only implementation to get a full 8/8. Below are common mistakes which will result in an adjustment of some kind:

  • Assuming %c/%d/%u/%x take a pointer to the data instead of the actual data as the argument.
  • Calling or jumping to onedig, or any other given code. You were to write your own solution, copying would have been okay.
  • Forgetting to save some of the $s0-$s7, $sp, $fp or $gp registers.
  • Overwriting the pre-guard, which is before the output buffer.
  • Overwriting the after-guard, which is after the output buffer.
  • Forgetting a null terminator on the output.
  • Off by one on the number of characters output (e.g. the output count shouldn't include the '\0')
Below are common errors which will not result in changes by your reader:

  • Arithmetic exception because you used add instead of addu.
  • Not handling any of the error cases we listed in the assignment.
10/9/2007 Project3 Update

Project3 has been updated to explain what to do when disassembled encounters an unsupported instruction (it should call exit(0) to exit), and we have updated the code in proj3.c slightly to match. We have also reinforced the need for you to format your output precisely as specified, otherwise you will lose points.
10/9/2007 MARS Update (Updated 10:45pm)

We patched MARS to recognize the '\0' and "\0" escape sequences.

We made a small update to MARS so that it can be used to run programs from the command line without an X11 (GUI) server. This means that you can now type mars-cs61c snprintf.s, for example, to run your project. This will work whether you have loged in at a terminal or using SSH.

You can check if you have the current mars-cs61c.jar at home by checking the file size against ~cs61c/bin/mars/mars-cs61c.jar. File size is a crude indicator of file identity. Furthermore the latest revision was uploaded 10/9/2007 at 8:29am, so if your jar is newer than that, you're fine.
10/8/2007 MARS Bugs (Updated 3:40pm)

Please make sure you are using mars-cs61c, which is our modified version of MARS. It includes some features (memory dump) required for proj3, and some bug fixes (PC relative branches and andi/ori/xori with 32bit values). Furthermore, please post any suspected or known MARS bugs to the course newsgroup, or e-mail Greg. Don't e-mail the MARS maintainers, they are unlikely to be able to help you very quickly, whereas we can patch the code quite easily.

There was a second, related, set of bugs which our fixes introduced.  If you had trouble assembling immediate instructions with negative immediates, please close and reload mars-cs61c.
10/4/2007 HW6 Update

We made some minor changes to HW6.
  • The questions remain the same.
  • To help try and get all the grades we can back to you ASAP, we're using quiz-style autograding (based on template.txt).
  • We have added several hints and clarifications, in an effort to avoid confusion.
9/30/2007 Proj1 Grading (Updated 10/1)

I'm getting tired of saying this, but proj1 was autograded a third time.  At least one of the test files had been corrupted.  It is unlikely that this re-run will significantly affect your score.

We re-ran the autograder last night, and as always if the comment for your score says "a reader will review" then it's quite likely your score is too low, and a human will look over it. Please relax about your score until that message changes.

There were two common causes of low scores, all of which will be corrected by readers:
  • You allocated the name for the root node statically (a constant instead of malloc). This is fine, but will cause the autograder to give you 0/8 because it's a little stupid.
  • You used different error messages than the oracle. This is fine, but again the autograder isn't smart enough to know that.
Furthermore, please do not be alarmed that your project has memory leaks. All solutions do because there is no exit command, and thus no chance to free all memory. The trick is that you should have the same number of leaked memory blocks as the oracle.

You can see the memory leak tester code at ~cs61c/files/ims.
9/29/2007 Errata Page

In response to some of the errors, and inconsistencies students have found in the course textbooks we have created an Errata page. It will contain a listing of all those problems which we cannot actually fix (bugs in code we don't have source, or errors in textbooks). Please let us know if you believe you have found things to be added to this list.
9/29/2007 MARS Bugs

A few students tried to get clever about the quiz13 question2, by using MARS to generate the machine code, instead of doing it by hand, or perhaps just to check their answers. Unfortunately, MARS had what we in this class will consider a bug in how it converted branches to machine code. The resulting program would run correctly, but the hexadecimal machine code representations of branch instructors were wrong. Full details would involve a description of "delayed branches" and can be obtained by asking on the newsgroup.

In order to fix the bug, we have installed a second copy of MARS on the instructional servers named mars-cs61c, in which this bug is fixed. Running mars will run normal MARS, as downloaded from the MARS website. Running mars-cs61c will run the version with our bug-fixes. You can download our version to your computer at home, from ~cs61c/bin/mars/mars-cs61c.jar. Please post any questions to the newsgroup.
9/27/2007 Project 2 Slight Fix

There was a bug in the snprintf.s code we gave you. Line 73 should have been addi $t0,$0,-39, rather than addi $t0,$0,39 (notice the negative sign on the constant?). You can either download the new version or just make the change yourself.

PS: If you download the new file, and have already modified yours, the utilities diff and patch might come in handy, if you don't feel like trying to copy-and-paste all the changes around properly. It's totally up to you though.
9/26/2007 Midterm 1 Review Session

We have a time and a place for the midterm1 review session: Friday, October 12 4-9pm in the Woz, the Wozniak Lounge 430/434 Soda Hall, right after class that day. You can see this event on the Calendar page. You should have most, if not all, assignments graded and back to you by then, we hope.
9/25/2007 Autograder E-mails (Updated 9/26)

We have finally sent all the autograder reports to your course e-mail addresses. You can use pine on the instructional servers to read your e-mail and see which test cases passed and which failed.

Alternatively you can set up a .forward file so that all future e-mail is forwarded to another address, simply put the address to forward all your mail to into a file named .forward in ~/ under your account.  You can find more information about e-mail setup here, including spam filtering!
9/25/2007 Proj1 Grading (Updated 9/26) 

We finally ran the autograder for proj1. Most people didn't get full credit. Keep a few things in mind before you get indignant:
  1. We grade hard.
  2. Your error messages might not have matched ours, this is fine and will be adjusted by the readers.
  3. Projects have slip-days (see the Info page) a zero for lateness will be double checked by the readers, and your slip days used instead.
  4. Don't bother worrying too much until the reader adjusts your score (if necessary).
  5. If you think your score is too high, let us know ASAP, first person to report a problem like this will keep their "too high" score.
  6. We forgot the test for memory leaks, and will be re-running this autograder on 9/29, your score may drop a bit if you have memory leaks
9/19/2007 Proj2 Update

We caught some bugs in project2 shortly after posting it, namely we hadn't finished updating it to use the MARS MIPS simulator instead of the old SPIM simulator. If you downloaded a file "snpf-main.s" from this site, please delete it, and go get the new "snprintf.s" instead.
9/19/2007 Project Slip Days

We just wanted to remind you that you have a total of 3 project slip days to use this semester. Check the course Info page for the details.
9/17/2007 Proj1 Oracle (Updated 3:30pm)

There was yet another bug in the oracle, in that the move command did not track parent pointers, when moving a file to a subdirectory. Making automated test cases out of these bugs would be a good idea.

Diligent students continue to find bugs in our project1 oracle. For example mkdir a followed by mv a a would result in the subdirectory a actually moving into itself, and thus disappearing. The oracle now prints an error instead, as your code should. Our deep copy was not properly setting the parent pointers, leading to some behaviour while using cd in a copied directory. Finally, we had yet another string handling bug in dirmain, which meant that if you hit <enter> on an empty command line, the previous command would be re-run. All of these have now been fixed. Our apologies for all the problems.
9/15/2007 Other Servers

Instructional Computing maintains a number of servers. The main server for the course is nova.cs.berkeley.edu, but since that is often heavily loaded, we suggest that you use others when you can. Any of the sun/solaris servers on that list will run all of the course software properly, including precompiled code which we distribute. The intel/solaris servers are the second choice, and should be mostly the same, though they will not run our precompiled code. The rest of the servers are quite useable but you should test your code on one of the solaris servers before handing it in, just to be sure.

You should also feel free to used cygwin to work on your windows machine at home, or, of course, any linux machine or Mac will already have most tools installed. Again be sure to test your code again on one of the solaris servers, before submitting it. C code, unlike Java, can have bugs (buffer overflows and pointer mistakes) which do different things on different machines, like crashing on one, but magically working on another. The only way to avoid this is to fully debug your code, and test it.
9/12/2007 Autograding Info (Updated 9/13)

To stem the tide of confusion, we've updated the autograding policy. In the future, starting with quiz7, everyone will receive a grade at the same time. Those which need review by a reader will be marked as such in the comments. We've updated the info page to match. We apologize for all the confusion, we hope this will stem the tide of questions like "why don't I have a grade for xxx".

We've added a description of how the autograding process works to the Info page. We apologize for not making this clear earlier.

We cannot stress enough that a blank score, or a score with a comment "a reader will review this" means that your grade is not yet set in stone. Please don't bother asking us about it until you have a score, and the comment does not mention a need for review by a reader.
9/12/2007 Projec1 Oracle (Updated 5pm & 9:40pm)

There was a second bug in our solution, whereby copying an empty directory would result in a segfault. Thankfully we caught it before anyone noticed.

There was a bug in our solution to the project, running cp existingdir newdir would copy the directory, and only the first file within it. Good catch to those students who noticed it. Please let us know if you ever see bugs in our solutions.

We have released a compiled version of our project 1 solution. You can use this to check on the proper behavior in the event that our project spec is wrong. also use this to test any automated test cases you develop for your project. We apologize for not releasing this sooner.

This version will run on nova.cs.berkeley.edu, but not your home computer. Sorry.
9/12/2007 Quiz4&5 Regrades & Late Quizzes (Updated 12:30pm) 

The autograder for quizzes 4 & 5 was seriously brain damaged, and accepted some very wrong responses and probably gave some very wrong grades. Quizzes 4 & 5 have been regraded. Thanks to the student who brought this mistake to our attention.

Having to regrade these quizzes means that every has a grade now for quiz4 & quiz5. If you have a grade of 2, your score will not be examined unless you wish to complain or point out an autograder mistake. If your score is less than 2, a reader will examine it. Normally you would not have been assigned a score of less than 2 until a reader had examined your submission. The first student to report receiving a grade higher than they deserve, will keep that high grade, as an incentive to report grading errors quickly.

Students who submitted a late version of any quiz have received 0s. A reader will review this to check if you submitted an on time version with a correct answer, in which case you will receive credit for your on-time version.

Please use the course newsgroup to ask general grading questions.
9/11/2007 Quiz Solutions & Grading

We've posted solutions to all the past quizzes on the Calendar page, and we've started autograding all submissions. If you don't yet see a score for the past quizzes when you run glookup it means that either you didn't get the right answer and we're going to check if you need partial credit, or we couldn't figure out how to automatically grade your solution, so a human will assign your score shortly. If we misgraded your quiz (gave you 100% when you got it wrong), please let Greg know ASAP.
9/11/2007 Connecting to the Newsgroup

Instructional Support Group has an excellent, and quite comprehensive tutorial on connection to the newsgroups here on campus. If you use a UNIX based reader on your class account, make sure you set the name you post under to be your own, otherwise all of your posts will appear to come from "Class Account", which makes it rather hard to follow a discussion. You can check your name with finger | grep <login>, your name is the "In Real Life". You can change your name by running ssh update and changing your Gecos information.
9/11/2007 Alternate Debugger: ddd

Adam Megacz kindly wrote up a tutorial on the ddd debugger, which you might find more or less useful than gdb. You might want to spend a few minutes playing with ddd to find out.
9/11/2007 Proj1 Clarifications

Testing: We added notes about how to run automated tests against your project, such as the kind you wrote for homework #2. Example: We added an example transcript of commands and their responses as you would see them when running the project to help clarify the spec.
9/8/2007 Proj1 Clarification

We've added a slight clarification to the mv and cp command details: "Checks to ensure that the move does not replace an existing file."
9/8/2007 Website Updates

For those of you with quick eyes or a stale browser cache, we've updated the location of some of the files on this website. We updated all the links we could find, and everything should be put back together. If you have any website troubles, e-mail cs61c-ta <at> imail.eecs.berkeley.edu.
9/8/2007 Proj1 Buffer Overflow

The file dirmain.c contained a buffer overflow error. You probably didn't even notice it on the instructional machines, but it may have caused problems on other systems, or other C compilers. We'll discuss this class of errors, and the specific error in discussion next week, since it's a common mistake. For now, get the updated dirmain.c, and don't worry too much about it.
9/4/2007 Quiz Submissions

When you submit a quiz, your submission should consist of exactly one file (template.txt) and no others. You should fill your answers into template.txt. Any other files in your submission will be completely ignored for quiz4 and later, though we will be lenient for quiz3 and earlier and go the extra mile to try and figure things out. If you have submitted quiz3 using a different filename for some reason, please go back and resubmit.  A note about this has been added to quizzes 3 and 4.

Remember: one file, template.txt, containing all your answers, no other files.
8/31/2007 Office Hours

All of the GSIs have now posted their office hours. You can find them on both the Staff and Schedule pages.
8/31/2007 Project1 Due Date

We have moved the project1 deadline to 9/19/2007, mostly so that we can give you more homework and labs to help get everyone up to speed and ensure things go smoothly. We'll also be adding a list of the tests you will be expected to pass to the assignment shortly. Feel free to get started early however, as the projects will be densely packed in the middle of the semester (see the Calendar).
8/29/2007 HW1 Typos

There were a few typos in the web page for HW1, first the source code for bitcount.c had some minor typos (the actual file was okay, but the listing in the assignment web page was wrong).  Also, we forgot to mention that you are not allowed to use arrays or strings in exercise 2, we've added a note about that.
8/29/2007 Accounts, Quiz1 & Clickers

For those people who were unable to get accounts yesterday, we will have account forms today after lecture. As a result of this and for the people who have been having trouble with submit, we will be pushing the Quiz1 deadline back to 7pm tomorrow, meaning you will get a chance to submit during lab. That way your TA can help if things go wrong.

Greg will hold some impromptu office hours outside the back of 2050 VLSB after lecture, to distribute account forms, deal with any odd CPS clicker problems and answer your various questions.

Please also consider taking a look at the Info page and the course newsgroup (see the Resources page) to answer your questions.
8/28/2007 Assignments

HW1: has been posted to the Calendar page and is due 2:45pm before class on Wednesday 9/5, no late credit, no exceptions.

Quiz1: has been fully posted to the Calendar page, including submission instructions and is due 2:45pm before class on Wednesday 8/29. There has been some confusion on the difference between pre-lecture quizzes which are submitted using your cs61c-** account, and the in-class peer instruction questions using the clickers, the Info page is a good place to find this kind of information.

Lab1: has been posted to the Calendar page. You should read it over, and make sure your account works before thursday's lab. In particular you may want to read some of the EMACS and GDB tutorials on the Resources page before lab, as that will save you quite a bit of time. Notice that the Calendar has tutorial sessions listed for Wednesday, Thursday and Friday this week on logging in EMACS and VI. These are run by CSUA and are strongly recommended for those unfamiliar with the tools we use.
8/28/2007 Staff E-mails

The e-mail addresses on the staff page were slightly incorrect (cs61-ta instead of cs61c-ta) and have been fixed.
8/28/2007 Gen1 CPS Keypads

The slightly older generation 1 CPS keypads should work just fine for this semester. Please make sure to register yours before lecture tomorrow, as we will be using them for the first time.
8/27/2007 Reading & Quiz Changes

We've added sections 5.1 and 5.2 of Kernighan & Ritchie to the reading list for wednesday, and swapped the order of the first two quizzes to better match the lecture and first lab. We apologize for any confusion this may have caused.
8/27/2007 CPS Enrollment

To participate in the in-class peer review questions, which you must do as part of your participation grade, you will need to enroll your CPS response pad using the einstruction website. Simply follow the instructions, the class key is "K32356C544", and the class name is "CS61C". Note: please do not enter your actual UC Berkeley Student ID when prompted for a student ID, instead enter the last two letters of your class account username. For example, cs61c-ta would enter "ta" for their student ID on the einstruction website.
8/23/2007 The First Week

For the first week of lecture, we expect you to have purchased both of the textbooks (see above left), and a CPS-RF transmitter which you can get from the Cal Bookstore.  There will be lecture, discussion and lab during the first week.  We will also be handing out the first homework and project by Wednesday, so be prepared to get started early.  There is reading for the first discussion section, but not the first lecture.  We will cover all of these announcements in the first lecture for those who have not read the website, so feel free to ask questions then.

Note also that the first two lectures will be given by the Head GSI, Greg Gibeling, because Prof. Wawrzynek is out of town.  As such, Prof Wawrzynek will not be having office hours this week, and you may wish to direct course questions to Greg.
8/21/2007 Welcome

Welcome to the CS61C website for Fall 2007.  Above, you will find links to all the major course related pages.  We're still adding content to the site, particularly the calendar. In the meantime please read the course information

We expect you to check this webpage often, we will assume you have seen any announcements within 24 hours of their posting (news about exam or assignment changes in particular).  Any problems with this site should be reported to your Head GSI, Greg Gibeling.


Site Maintained by: Greg Gibeling (gdgib <at> berkeley.edu)