Merge pull request #105 from jzarnett/winter2024-updates

labs/Lab1.tex

@@ -169,8 +169,8 @@ \subsection*{Simple dependency calculations: \texttt{deps\_available.rs} (25)}
 \vspace{1em}
 Let's now combine the available and installed data. You are going to implement \texttt{deps\_available},
 which checks the declared dependencies of a package and prints out whether they are satisfied or not.
-(Please stick to the provided output format for the TAs' sake; I've provided the necessary \texttt{println!} 
-statements).
+(If you modify the output format it will cause the autograder to fail, which will make you sad when you see the grade, and us sad when we have to manually check the output.
+I've provided the necessary \texttt{println!} statements.).
 
 \begin{verbatim}
 $ deps-available 4g8
@@ -389,7 +389,7 @@ \section*{Clarifications}
 \begin{itemize}
 \item \emph{What about compilation warnings?} We encourage your code to be free of warnings but we won't dock marks if there is no underlying issue. You can also fix linter errors in the provided code.
 \item \emph{What about error handling?} You can pretend that failures never happen (famous last words), i.e. you need not implement code to retry after failures. If you do notice an error, print a message and exit. You can also assume that the inputs are correctly-formatted and solveable.
-  \item \emph{What can we change in the provided code?} Go ahead and change it as you like. You can add extra optional command-line arguments and commands if you'd like, but don't make them mandatory (i.e. the TA should be able to plug-and-play your code just like all the other solutions.) 
+  \item \emph{What can we change in the provided code?} Go ahead and change it as you like, except for the print statements (we need consistent output to autograde). You can add extra optional command-line arguments and commands if you'd like, but don't make them mandatory (i.e. the TA should be able to plug-and-play your code just like all the other solutions.) 
 \item Tell me more about the \texttt{how-to-install} algorithm.
 
 \vspace{1em} Here's an example (thanks to Alexander Kursell). $A$ depends on $[B, C]$; $B$ depends on $D \mid E$; and $C$ depends on $E \mid F$. Let's say the rules resolve $B$'s dependency to $D$ and $C$'s dependency to $E$.
@@ -403,6 +403,13 @@ \section*{Clarifications}
 \item Pull $D$ and $E$ from the worklist.
 \end{enumerate}
 You end up with install set $[A, B, C, D, E]$.
+
+\item \emph{What about packages that are in the "Provides" list?} Some packages referenced are not real packages, but are provided by another package. Your solution should still include the "provided" package as a dependency, but it should not appear in available packages.
+\item \emph{What about packages with ":any" (or other tags)?} Do not add special handling for these tags. The reference implementation treats them as part of the package name to simplify the assignment, so should you.
+\item \emph{How to determine the highest version number in how-to-install?} For each alternative, find the available version of the package that meets the version requirements. Once you have these versions, use them to determine the highest version alternative.
+\item \emph{Can I print duplicate dependencies?} No, the dependencies you calculate and print should be unique.
+\item \emph{Please ensure the verify server is set back to the default} We do not run a local verify server, so if you are using your own for testing, please ensure the default one is enabled in your submission.
+\item \emph{Does order of packages matter?} No
 \end{itemize}
 
 \paragraph{version 3 update:} fixed incorrect output for 3depict in v2 of this writeup, add note about redirection.
@@ -419,6 +426,8 @@ \section*{Clarifications}
 
 \paragraph{version 9 update:} 2023.
 
+\paragraph{version 10 update:} Winter 2024 clarifications.
+
 \end{document}
 
 % note for next year: we should define a testing api and call those methods from the print methods, saving the necessity to format properly

labs/Lab2.tex

@@ -152,7 +152,12 @@ \subsection*{What goes in a commit log}
 Write logs in files {\tt commit-log/separate-maps.md} and {\tt commit-log/concurrent-hashmap.md}.
 
 \section*{Clarifications}
-I'm sure I'll have some. But I don't have any yet.
+
+\begin{itemize}
+  \item \emph{I noticed a bug where the starter code double counts for consecutive lines.} Leave the double count there (and match the behaviour in your submission).
+  \item \emph{Ensure --num_threads 1 works} Make sure your assignment does something sane for different inputs of num_threads, including edge cases like 0 and 1. Note the default value for num_threads is 8, so leaving the argument off the command line will not test your single threaded case.
+  \item \emph{Does output order matter?} You need to print the same sections, in the same order, as the starter code. The printing of hashes/lists do not need to have the elements in the same order. Do not remove any prints.
+\end{itemize}  
 
 \bibliographystyle{alpha}
 \bibliography{Lab2}

labs/Lab3.tex

@@ -180,6 +180,8 @@ \section*{Clarifications}
 
 \paragraph{Do we commit compiled files to git?} (\verb+kernel.ptx+ and \verb+output/out_cuda.csv+) Doesn't matter, we'll regenerate.
 
+\paragraph{Do I need to implement it all on the GPU?} An optimal (and full mark) solution will have all of the compute in GPU kernel(s).
+
 \newpage
 \section*{Appendix: Taking arrays as arguments in C instead of raw pointers}
 This appendix is authored by previous ECE 459 student Aidan Wood, who was kind enough to allow us to include this in the description.

labs/Lab4.tex

@@ -140,4 +140,12 @@ \section*{Clarifications}
 
 \paragraph{How many packages?} This is set using a command-line option.
 
+\paragraph{What testscases are you going to use for performance testing?} We'll use a variety of test cases. The one in the Makefile is one of the harder ones. We've found increasing the idea count, the pkg gen count, and the student count to large numbers to also be difficult.
+
+\paragraph{What prints do I need to keep?} Everything from "Global checksums" to the end. Do not rename Global checksums.
+
+\paragraph{Do I make a pull request for you to mark?} No, please merge everything to main.
+
+
+
 \end{document}