The good news: t-shirts are available! Come pick them up from Mollie's
office on the 3rd floor. They are $15 (no, we don't take credit cards).
The bad news: they appear to have botched the sides, as in the front is
the back and vice versa. While this annoys me to no end since I carefully
described which side was which to them, it's not that big a deal...the
t-shirt works. It's just more...unconventional now. =)
Thanks again for a great semester!
Nick
Yes, I do in fact mean _Thursday_, 5/23...it was the earliest they could
get them back. And I will try to compile such a list before then. =)
Nick
> -----Original Message-----
> From: Bryan Choi [mailto:bchoi@fas.harvard.edu]
> Sent: Monday, May 20, 2002 6:35 PM
> To: Nicholas C Murphy
> Subject: Re: [cs161-list] T-shirt logistics
>
>
> oh yea, and perhaps it'd be easier if you had a list with the
> sizes and who ordered what at mollie's office when we pick up
> our shirts on
> *thursday* (not tuesday)?
> just a suggestion.
>
>
> On Mon, 20 May 2002, Nicholas C. Murphy wrote:
>
> > The t-shirts are ordered! They should be available late Thursday
> > afternoon. So!...to get your t-shirt you need to take the
> following
> > steps.
> >
> > 1) Find/earn/blackmail $15 somewhere.
> > 2) Wait until Thursday. Try not to lose the $15.
> > 3) Late Tuesday afternoon (after about 3 pm or so), go to Mollie
> > Goldbarg's office (the same place you presumably got the
> source book).
> > 4) If there is a line there, wait in it.
> > 5) Give Mollie the $15.
> > 6) Mollie will give you a t-shirt.
> > 7) Wear the t-shirt with pride and glee. Sneer at the plebians
> > mindlessly scurrying around the Yard. Bask in your superiority.
> >
> > Please get only the size you said that you wanted in the email you
> > sent me. There are few if any extras.
> >
> > If you have any questions, please email me.
> >
> > Thanks,
> > Nick
> >
> > _______________________________________________
> > cs161-list mailing list
> > cs161-list(a)fas.harvard.edu
> > http://www.fas.harvard.edu/mailman/listinfo/cs161-list
> >
>
The t-shirts are ordered! They should be available late Thursday
afternoon. So!...to get your t-shirt you need to take the following
steps.
1) Find/earn/blackmail $15 somewhere.
2) Wait until Thursday. Try not to lose the $15.
3) Late Tuesday afternoon (after about 3 pm or so), go to Mollie
Goldbarg's office (the same place you presumably got the source book).
4) If there is a line there, wait in it.
5) Give Mollie the $15.
6) Mollie will give you a t-shirt.
7) Wear the t-shirt with pride and glee. Sneer at the plebians
mindlessly scurrying around the Yard. Bask in your superiority.
Please get only the size you said that you wanted in the email you sent
me. There are few if any extras.
If you have any questions, please email me.
Thanks,
Nick
Remember the assignment 3 performance graphs we showed in class?
It had a couple of last year's kernels on it. Next year, I'd like to
have a few of this year's kernels on it too. (A few, not all 12; I'm
thinking 3-5 or so.)
Because things change over time, I need to be able to patch and
recompile the kernels involved, so I need the sources. I'm intending
to do both VM and FS performance graphs next time around, so I'd be
interested in both assignment 3 kernels and assignment 4 kernels.
Also, if you souped up your VM system in assignment 5, you can send
that too and I'll make a separate category for it.
So, if you and your partner would like your kernel to be among the
chosen few, send me appropriate release(s) of your operating system
that I can stash in the CS161 staff's CVS repository.
Assignment 3 (VM system performance) kernels, souped up or not, should
have the following characteristics:
- stable. They should survive all of the tests we might want to use
for benchmarking, which includes triplehuge, triplemat, and
parallelvm. (Note that parallelvm requires 1M of RAM.)
- optimizable. Ideally, they should still work if compiled with -O2.
- unaffected by asst4. If what you send has some assignment 4 code
in it, #if 0 it or at least make sure it's not taking RAM for
buffers away from the VM system.
For the souped-up category, send whatever you like. For the non-
souped-up category, please send something reasonably close to what you
actually handed in for assignment 3. That is, make bug fixes, but
please don't do structural changes or additional performance
optimization; the graphs should be fair comparisons.
Assignment 4 (FS performance) kernels should be, likewise, stable and
optimizable, and reasonably close to what you actually handed in for
assignment 4.
(Please do include any bug fixes you know about. If you have one last
mysterious bug that you can't find, I'll help, and so forth.)
Note that it does *not* matter whether your kernel performs "well";
I'd rather have a range of performance than the five fastest kernels.
The performance graphs will be anonymous; your names and/or group name
won't be attached unless _everyone_ agrees. Or perhaps if you become
course staff. :-)
And, finally: I don't need these right away. Anytime in the next few
weeks is fine... or even later.
--
- David A. Holland / dholland(a)eecs.harvard.edu
Okay, I have orders from:
Georgi Matev
John Sheu
Kittle Leung
Fritz Behr
Sanjay Mavinkur
David Xiao
Josh Forman
Uri Braun
Yau Chin
Richard Eisenberg
Jesse Tov
Michael Love
Jess Bezanson
Dave Holland
Sasha Fedorova
James Chen
David Chen
Octavian Timaru
Jeff Shneidman
Charles Duan
Roger Huang
Bryan Choi
Me
Margo
If you're not on this list and want to be, or else are on the list and
don't want to be, send me an email.
Nick
I will take this time to congratulate you on finishing the most
wonderful, blissful, and relaxing class you have taken at Harvard.
Now is also a good time to inform you that grading is actually done
entirely on the basis of a combination of randomness and physical
attractiveness.
This means that Mike Vernal automatically failed.
You go sleep now.
Nick
Sometime before you begin working on the exam, please send a note to
cs161(a)fas.harvard.edu if you want a t-shirt (based on the last design I
sent out), and if so, what size. Note that if you send a note, you are
honor-bound to actually give me money!
It looks like the shirt should be around $10-$15.
Nick
I. It was pointed out that the rename code does not release its locks
when returning an error. This was an oversight; please assume it
releases the locks in each of the error return cases.
II. It was pointed out that in cases where the correct behavior is to
return an error, answers of the form "It will work correctly" and "It
will return an error" are equivalent and therefore ambiguous. In these
cases, please assume that "It will work correctly" means "It will
succeed (without corrupting the file system)."
--
- David A. Holland / dholland(a)eecs.harvard.edu
The pathnames supplied in multiple-guess question #19 are not
absolute. However, the code provided is defined in such a way that it
doesn't make sense to have non-absolute paths.
Proceed under the assumption that the paths are "/a/." and "/b", rather
than "a/." and "b", respectively.
--
- David A. Holland / dholland(a)eecs.harvard.edu
Multiple-guess question 19 says "Assuming that [...] a/. is not resolved
to be the same as a".
This clause was meant to be interpreted textually; that is, the lookup
routines contain no special-case code for "." (or "..").
"a" and "a/." do still (normally) refer to the same inode.
Also, "b does not" should be read "b does not exist".
--
- David A. Holland / dholland(a)eecs.harvard.edu
