Messages - Page 2
If your Java program crashes because of lack of space - here is what to do:
Use the following option "-Xmx7500M", e.g.,
java -Xmx7500M myprogram 2000000000 4
Eric
Oblig3 was published last night - an a minor correction caused publication of version 2 today. The only change is the max time for factorization - we have increased the limit to take older, slower hardware into consideration.
PLEASE NOTE: Do NOT change the precode - just put it into a separate file, compile and use it.
Eric
Concerning those that took the course INF2440 last year: Any of the five Obligs that were approved last year can be transferred to this year. To have it transferred, you must resubmit it in devilry again this year with a note that it was approved in 2018, so that it can become official for IN3030 (which, technically, is a different course than INF2440). Once submitted, we will verify that it was approved by checking devilry3 (the Beta-version of the renewed devilry) and then approve the oblig for this year.
Eric
On Wednesday and Friday this week, we'll do some live coding of the algorithm we'll use in oblig 3. The code will be published, but you'll learn more if you attend.
Here's some details with a nice animation:
https://en.wikipedia.org/wiki/Sieve_of_Eratosthenes
Magnus
Lecture material uke5, uke6, uke7 uploaded.
Also uploaded: Hoare's seminal 1974 article on Monitors - done right.
Eric
1. The three variant that you must implement for A x B are:
- A x B using the classic method
- A x Transpose(B)
- Transpose(A) x B
2. For each variant, show the results in table form and in a graph. Each graph should show speedup of the PARALLEL version relative to SEQUENTIAL version of the algorithm.
3. Furthermore, show a graph with execution times for all six. Show the performance difference between the fastest of the six and the five others. If which is fastest varies depending on the matrix size then use the one that is fastest for the larger arrays.
Have fun :-)
Eric
The line "You must use the published precode that we provide on the course web site to fill the two arrays to be multiplied." in the oblig text should have read:
"You must use the published precode that we provide on the course web site to fill the two arrays to be multiplied, and then save each resulting matrix using the save method with the appropriate mode."
You should run the save method six times, one for each mode.
This is mandatory and helps the TAs (and yourself) check the correctness of your multiplication. Notice that there is a small example in the precode that you can use for verification of your multiplication.
Magnus
We see that there are some questions and posts on piazza which contain useful info that we would like to share so we copy them here:
Q: Are we suppose to just use array.sort on the sorting part where we have to use Array.sort and print out the last k number of sorted part ?
A: You use Arrays.sort only for testing that your own solution is correct and to report how much time Array.sort takes (Include the time measurements in the report).
Q: P? hvilke n-verdier kan man forvente ? f? speedup > 1?
A: Du kan i hvert fall forvente det ved n = 100 mill , men sannsynligvis f?r man det ogs? ved n = 1 mill.
- Maria
-----
Jeg tok en kjapp sp?rrerunde i begynnelsen av gruppetimen i dag. De aller fleste f?r speedup > 1 mellom n=1M og n=10M. Det kommer ann p? hardware og andre faktorer.
- Magnus
 ...
This week we will have a kahoot on both the wednesday and friday group sessions. The questions will be relevant to the lecture you had on friday. There will be a small prize for the winner!
Magnus & Maria
Det er n? ledig plass p? emnet, p?melding via studentweb.
At the link below you'll find both Maria and my live-code (separate folders). We'll publish the code from later live-code sessions and most likely the pre-code for some of the assignments in this repo.
https://github.uio.no/magnuesp/IN3030-v19
Magnus
Piazza signup: piazza.com/uio.no/spring2019/in3030
Piazza IN3030 home:piazza.com/uio.no/spring2019/in3030/home
Signup and use it!...
Hi.
I have put ukeoppgaver for uke1/week1 on the web site, see on the left side of the home page.
At the TA sessions (gruppetimene), the TAs will show how to do parallelprogramming in Java using Live Coding. After the sessions, you are encouraged to do the uke1 problems as to ensure that you can do parallelprogramming.
If you are unable to show up for the TA sessions, then try doing the uke1 problems.
Have fun!
Eric