1
00:00:00,040 --> 00:00:02,480
The following content is
provided under a Creative

2
00:00:02,480 --> 00:00:03,900
Commons license.

3
00:00:03,900 --> 00:00:06,340
Your support will help
MIT OpenCourseWare

4
00:00:06,340 --> 00:00:10,610
continue to offer high-quality
educational resources for free.

5
00:00:10,610 --> 00:00:13,310
To make a donation or
view additional materials

6
00:00:13,310 --> 00:00:17,200
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:17,200 --> 00:00:17,825
at ocw.mit.edu.

8
00:00:21,670 --> 00:00:25,170
PROFESSOR: So today, we're going
to have the final presentations

9
00:00:25,170 --> 00:00:26,990
from the student projects.

10
00:00:26,990 --> 00:00:28,400
So they did a lot of hard work.

11
00:00:28,400 --> 00:00:30,164
Some of them are
still working upstairs

12
00:00:30,164 --> 00:00:32,485
and will come down just
in time for presentation.

13
00:00:32,485 --> 00:00:37,993
So just a few slides on format
and how the judging will

14
00:00:37,993 --> 00:00:41,580
happen-- so each team will
have 15 minutes during which

15
00:00:41,580 --> 00:00:43,050
to tell about their
little project

16
00:00:43,050 --> 00:00:45,900
and to give us a little demo.

17
00:00:45,900 --> 00:00:47,720
At least two PlayStation
3s will be set up

18
00:00:47,720 --> 00:00:50,300
by the end of the class.

19
00:00:50,300 --> 00:00:51,940
One will have audio.

20
00:00:51,940 --> 00:00:55,440
The other will have video and at
least two teams will use them.

21
00:00:55,440 --> 00:00:57,447
The others will
SSH in and show you

22
00:00:57,447 --> 00:00:59,280
their programs running
on the PlayStation 3s

23
00:00:59,280 --> 00:01:02,150
that we have upstairs
on the network.

24
00:01:02,150 --> 00:01:04,740
One of the things that they'll
be graded on or scored on

25
00:01:04,740 --> 00:01:06,490
is performance.

26
00:01:06,490 --> 00:01:08,140
Some of the things
we'll do after class

27
00:01:08,140 --> 00:01:10,210
is performance profiling,
to try to understand

28
00:01:10,210 --> 00:01:12,660
how much scalability
you were able to achieve

29
00:01:12,660 --> 00:01:15,803
using parallel processor versus
two versus three, all the way

30
00:01:15,803 --> 00:01:18,270
up to six.

31
00:01:18,270 --> 00:01:20,980
So what we'll do is we'll
set up some time with the TAs

32
00:01:20,980 --> 00:01:22,550
to do the profiling after class.

33
00:01:22,550 --> 00:01:25,960
And you can update your
source code up until 9:00 AM.

34
00:01:25,960 --> 00:01:29,490
So you'll show us how
to do the profiling

35
00:01:29,490 --> 00:01:32,580
and then we can re-run that
at 9:00 AM Saturday morning

36
00:01:32,580 --> 00:01:36,224
and give you updated scores
to change your grades

37
00:01:36,224 --> 00:01:37,840
at the time of grades.

38
00:01:37,840 --> 00:01:41,000
There are six
scoring categories.

39
00:01:41,000 --> 00:01:44,277
Performance-- and this
first Performance category

40
00:01:44,277 --> 00:01:46,599
is essentially what your
performance metric is.

41
00:01:46,599 --> 00:01:48,140
You're going to
define it, hopefully,

42
00:01:48,140 --> 00:01:50,480
and then you'll tell
us how well you did.

43
00:01:50,480 --> 00:01:53,180
Project scope-- this is
essentially a measure,

44
00:01:53,180 --> 00:01:55,620
because each of your different
projects were different.

45
00:01:55,620 --> 00:01:57,084
You picked your own projects.

46
00:01:57,084 --> 00:01:58,548
We'll try to get
an understanding

47
00:01:58,548 --> 00:02:01,070
for how much you tried to
cover and what you actually

48
00:02:01,070 --> 00:02:02,340
did achieve.

49
00:02:02,340 --> 00:02:05,695
So we'll hopefully give
us a good idea of that.

50
00:02:05,695 --> 00:02:08,669
Completeness-- how
robust is your code?

51
00:02:08,669 --> 00:02:10,750
How much you actually
achieve relative

52
00:02:10,750 --> 00:02:12,840
to what you wanted to do?

53
00:02:12,840 --> 00:02:14,860
Are there data races?

54
00:02:14,860 --> 00:02:17,460
Is your code reproducible
with every run?

55
00:02:17,460 --> 00:02:19,994
Presentation and Demo-- so
we've decoupled it here,

56
00:02:19,994 --> 00:02:22,035
just because there are
different ways people will

57
00:02:22,035 --> 00:02:22,759
be doing a demo.

58
00:02:22,759 --> 00:02:25,050
So you can sort of earn points
in different categories.

59
00:02:25,050 --> 00:02:26,633
If you have a really
good presentation

60
00:02:26,633 --> 00:02:29,874
but you might not be flashy,
we'll take that into account.

61
00:02:29,874 --> 00:02:31,790
And then, Performance
Scalability, which we'll

62
00:02:31,790 --> 00:02:32,530
do off and on.

63
00:02:32,530 --> 00:02:36,700
So each one of these categories
score on a 10-point basis

64
00:02:36,700 --> 00:02:39,310
from one to 10, 10
being really good,

65
00:02:39,310 --> 00:02:41,740
and for a total of 60 points.

66
00:02:41,740 --> 00:02:44,690
So for our judging
panel, unfortunately, we

67
00:02:44,690 --> 00:02:47,320
couldn't get virtual Saman
to really work, so Saman

68
00:02:47,320 --> 00:02:49,064
is with us on the phone.

69
00:02:49,064 --> 00:02:50,980
We have Michael Gordon
and Bill [? Theeves, ?]

70
00:02:50,980 --> 00:02:52,800
two graduate students
in Saman's group.

71
00:02:52,800 --> 00:02:56,410
We have myself, Professor
Steve Ward from MIT,

72
00:02:56,410 --> 00:02:57,890
and Professor Weng
Fai Wong, who's

73
00:02:57,890 --> 00:03:00,014
visiting from the National
University of Singapore,

74
00:03:00,014 --> 00:03:02,810
doing a sabbatical year here.

75
00:03:02,810 --> 00:03:04,630
Student presentations
will dynamically

76
00:03:04,630 --> 00:03:08,200
reorder based on who's
available and who's ready to go.

77
00:03:08,200 --> 00:03:10,740
This is a rough ordering
that was picked randomly.

78
00:03:10,740 --> 00:03:11,770
So [? Pumuk is ?] here.

79
00:03:11,770 --> 00:03:12,790
He'll go first.

80
00:03:12,790 --> 00:03:15,405
And then, I see Linear Algebra
Pack so they'll go after that.

81
00:03:15,405 --> 00:03:17,530
The Ray Tracing team is
still in my office working,

82
00:03:17,530 --> 00:03:19,490
so we might have
to bump them down.

83
00:03:19,490 --> 00:03:21,430
Soft Radio Team is here?

84
00:03:21,430 --> 00:03:22,100
Yep.

85
00:03:22,100 --> 00:03:23,764
Speech is here.

86
00:03:23,764 --> 00:03:24,264
Backgammon?

87
00:03:24,264 --> 00:03:24,764
Yeah.

88
00:03:24,764 --> 00:03:30,180
Those guys were ready
having breakfast at 9:00.

89
00:03:30,180 --> 00:03:33,220
The seventh team, Molecular
Dynamic Simulator--

90
00:03:33,220 --> 00:03:35,190
unfortunately, they're not here.

91
00:03:35,190 --> 00:03:38,780
That team is not
physically in Boston,

92
00:03:38,780 --> 00:03:40,250
so they'll present
tomorrow morning

93
00:03:40,250 --> 00:03:45,180
but will not be considered as
competitors in the competition.

94
00:03:45,180 --> 00:03:46,582
There is a countdown clock.

95
00:03:46,582 --> 00:03:47,905
Bill is running his laptop.

96
00:03:47,905 --> 00:03:49,810
You can see it directly
in front of you.

97
00:03:49,810 --> 00:03:52,520
It'll start at 15
minutes and every minute,

98
00:03:52,520 --> 00:03:56,380
it'll tick down, 14, 13,
all the way down to zero.

99
00:03:56,380 --> 00:03:58,820
And you can go into
negative times,

100
00:03:58,820 --> 00:04:01,001
but we'll start
subtracting points.

101
00:04:01,001 --> 00:04:06,690
So the more you run over, the
more points you start losing.

102
00:04:06,690 --> 00:04:08,190
So try to keep under
the time limit.

103
00:04:08,190 --> 00:04:09,550
What about questions?

104
00:04:09,550 --> 00:04:11,300
If there are questions,
we can take those

105
00:04:11,300 --> 00:04:12,258
after the presentation.

106
00:04:12,258 --> 00:04:15,914
If it's a short question--
yeah, let's leave all questions

107
00:04:15,914 --> 00:04:18,159
until after presentation
today, unless there's

108
00:04:18,159 --> 00:04:18,950
something pressing.

109
00:04:18,950 --> 00:04:20,769
I'll try to moderate.

110
00:04:20,769 --> 00:04:24,150
So if we're ready to
start, come up here.