1
00:00:00,040 --> 00:00:02,540
NARRATOR: The following content
is provided under a Creative

2
00:00:02,540 --> 00:00:04,000
Commons license.

3
00:00:04,000 --> 00:00:06,340
Your support will help
MIT OpenCourseWare

4
00:00:06,340 --> 00:00:10,710
continue to offer high quality
educational resources for free.

5
00:00:10,710 --> 00:00:13,310
To make a donation or
view additional materials

6
00:00:13,310 --> 00:00:17,192
from hundreds of MIT courses,
visit MIT OpenCourseWare

7
00:00:17,192 --> 00:00:17,817
at ocw.mit.edu.

8
00:00:20,876 --> 00:00:21,980
RODRIC RABBAH: OK, great.

9
00:00:21,980 --> 00:00:24,800
So a lot of effort into
making this course, and we

10
00:00:24,800 --> 00:00:26,742
hope you've had fun.

11
00:00:26,742 --> 00:00:31,120
It's been a lot of support
from Sony, IBM, Toshiba.

12
00:00:31,120 --> 00:00:34,880
A lot of on demand help
from people at IBM and Sony,

13
00:00:34,880 --> 00:00:38,317
trying to debug problems,
even up to the last minute.

14
00:00:38,317 --> 00:00:39,650
So we're very thankful for that.

15
00:00:39,650 --> 00:00:41,820
And we hope in return that
you really enjoyed the course,

16
00:00:41,820 --> 00:00:42,870
and it was really fun.

17
00:00:42,870 --> 00:00:45,870
So looking forward to
the course evaluations.

18
00:00:45,870 --> 00:00:47,740
So just want to tell
you a little bit

19
00:00:47,740 --> 00:00:49,260
about some opportunities ahead.

20
00:00:49,260 --> 00:00:53,450
I've passed out these brochures
for the cell broadband engine

21
00:00:53,450 --> 00:00:56,080
challenge, Beyond Gaming.

22
00:00:56,080 --> 00:01:00,040
And this is a IBM sponsored
event, where you really

23
00:01:00,040 --> 00:01:03,150
can compete in two different
challenges, and in one of them

24
00:01:03,150 --> 00:01:07,650
you can win really
seriously cool t-shirts.

25
00:01:07,650 --> 00:01:13,240
That's actually lifted
directly from the online link.

26
00:01:13,240 --> 00:01:17,090
So the competition starts
February 5 at midnight.

27
00:01:17,090 --> 00:01:20,310
And you have to pass an online
quiz, of which I expect all

28
00:01:20,310 --> 00:01:22,770
of you should be able to pass.

29
00:01:22,770 --> 00:01:25,220
And hopefully, you will
compete and at least get

30
00:01:25,220 --> 00:01:27,300
a seriously cool t-shirt.

31
00:01:27,300 --> 00:01:30,680
The second challenge, which is
potentially more interesting,

32
00:01:30,680 --> 00:01:33,200
just because there's
a $10K grand prize--

33
00:01:33,200 --> 00:01:37,500
there's also a $7K second
prize, and some serious money

34
00:01:37,500 --> 00:01:39,740
for at least three
or four prizes.

35
00:01:39,740 --> 00:01:42,580
You can look at information
in the brochures.

36
00:01:42,580 --> 00:01:44,080
According to the
information, you'll

37
00:01:44,080 --> 00:01:47,326
find it requires some serious
coding skills on your part.

38
00:01:47,326 --> 00:01:48,700
But I certainly
expect all of you

39
00:01:48,700 --> 00:01:51,800
to be able to be well
primed at this point,

40
00:01:51,800 --> 00:01:54,700
especially since you've
learned about cell hands on,

41
00:01:54,700 --> 00:01:58,550
and you've discovered all
the pitfalls on your own.

42
00:01:58,550 --> 00:02:02,610
It would make for a very good
competition on your part.

43
00:02:02,610 --> 00:02:03,960
There are different categories.

44
00:02:03,960 --> 00:02:06,100
Sorry, it's three
categories, not four.

45
00:02:06,100 --> 00:02:09,374
Application solutions-- so some
of things, some of the people

46
00:02:09,374 --> 00:02:11,540
have done more application
oriented thing could fit,

47
00:02:11,540 --> 00:02:13,800
or you can sort of take
what you've learned,

48
00:02:13,800 --> 00:02:15,720
apply it to financial
services if you

49
00:02:15,720 --> 00:02:17,750
have interest in
that, medical imaging,

50
00:02:17,750 --> 00:02:22,230
electronic design, automation,
things of that sort.

51
00:02:22,230 --> 00:02:24,114
Operating systems--
I'm not sure if that's

52
00:02:24,114 --> 00:02:26,030
particularly relevant
to a lot of people here.

53
00:02:26,030 --> 00:02:27,770
I'm not sure if that's
in their interest area,

54
00:02:27,770 --> 00:02:28,710
but programmability.

55
00:02:28,710 --> 00:02:31,990
So certainly the
streaming framework

56
00:02:31,990 --> 00:02:34,919
that one of the software
radio projects tried to do

57
00:02:34,919 --> 00:02:36,460
would fit onto that,
some of the work

58
00:02:36,460 --> 00:02:38,460
that the Ray Tracer
team did, in terms

59
00:02:38,460 --> 00:02:41,760
of coming up with their
own interface for graphics

60
00:02:41,760 --> 00:02:43,610
could certainly fit.

61
00:02:43,610 --> 00:02:47,420
So I urge you to consider
competing in these.

62
00:02:47,420 --> 00:02:51,250
There's probably a lot of fame
and fortune to be had here.

63
00:02:51,250 --> 00:02:53,170
So we had our own competition.

64
00:02:53,170 --> 00:02:56,440
And what you're going to get are
some seriously cool trophies,

65
00:02:56,440 --> 00:02:58,650
since we don't have t-shirts.

66
00:02:58,650 --> 00:03:04,060
So Mary was very--
could have had t-shirts.

67
00:03:04,060 --> 00:03:07,880
I didn't want to compete
with IBM directly.

68
00:03:07,880 --> 00:03:12,320
So for the winning team, you
will get this seriously awesome

69
00:03:12,320 --> 00:03:16,940
looking trophy, which has a
little imprint on the front

70
00:03:16,940 --> 00:03:22,890
that says first place, 6.189.

71
00:03:22,890 --> 00:03:26,480
And you're also going to
get a $150 gift certificate,

72
00:03:26,480 --> 00:03:29,410
I think we announced this at
the beginning it of the course.

73
00:03:29,410 --> 00:03:31,310
I'm also working
with people at IBM

74
00:03:31,310 --> 00:03:33,650
on arranging a one
day trip to IBM T.J.

75
00:03:33,650 --> 00:03:36,260
Watson Research Center,
for the winning team

76
00:03:36,260 --> 00:03:39,800
to present their winning
project, do a demo,

77
00:03:39,800 --> 00:03:42,750
and then meet with people
at IBM who are doing really

78
00:03:42,750 --> 00:03:44,750
cool, interesting research.

79
00:03:44,750 --> 00:03:47,120
And it should be a fun day.

80
00:03:47,120 --> 00:03:48,746
And certain people
at IBM have been

81
00:03:48,746 --> 00:03:50,495
very excited about
hearing what's going on

82
00:03:50,495 --> 00:03:52,750
in this project, in this class.

83
00:03:52,750 --> 00:03:55,430
And so it'll be a
good, fun opportunity,

84
00:03:55,430 --> 00:03:57,400
I think, all around.

85
00:03:57,400 --> 00:04:01,170
OK, so I'm going to work my
way up to the first prize.

86
00:04:01,170 --> 00:04:03,800
I think in terms of the
complexity of the project

87
00:04:03,800 --> 00:04:05,990
handled, certainly
Speech Synthesis probably

88
00:04:05,990 --> 00:04:09,350
took on the most
complex-- certainly

89
00:04:09,350 --> 00:04:11,380
in terms of the
software that they used,

90
00:04:11,380 --> 00:04:13,622
and then maybe that
bogged them down a little.

91
00:04:13,622 --> 00:04:15,080
So there's probably
something to be

92
00:04:15,080 --> 00:04:17,459
said in terms of starting
from scratch versus taking

93
00:04:17,459 --> 00:04:19,380
complex infrastructure
and re-engineering it

94
00:04:19,380 --> 00:04:20,390
for parallelism.

95
00:04:20,390 --> 00:04:22,220
I think you've
discovered firsthand

96
00:04:22,220 --> 00:04:23,870
how hard that really is.

97
00:04:23,870 --> 00:04:27,030
And this is a big problem
for parallel computing.

98
00:04:27,030 --> 00:04:28,800
There's a lot of legacy
software out there

99
00:04:28,800 --> 00:04:30,425
that eventually might
have to be ported

100
00:04:30,425 --> 00:04:32,010
to parallel architectures.

101
00:04:32,010 --> 00:04:32,990
How do you do it?

102
00:04:32,990 --> 00:04:35,510
You pointed out, code
is not well documented.

103
00:04:35,510 --> 00:04:36,620
Tools are lacking.

104
00:04:36,620 --> 00:04:37,560
So what do you do?

105
00:04:37,560 --> 00:04:40,900
So you've stumbled on a really
seriously important problem

106
00:04:40,900 --> 00:04:44,830
facing a lot of people
in the industry.

107
00:04:44,830 --> 00:04:46,290
The most synchronization issues.

108
00:04:46,290 --> 00:04:51,370
We believe the guys who are
doing the Battery Simulations

109
00:04:51,370 --> 00:04:53,576
and modeling sort of
the chemical reactions

110
00:04:53,576 --> 00:04:55,700
certainly ran into a lot
of synchronization issues,

111
00:04:55,700 --> 00:05:01,240
just because of the dependent
nature of the computation.

112
00:05:01,240 --> 00:05:04,040
Next up on the list would be for
the most parallel programming

113
00:05:04,040 --> 00:05:04,980
challenges.

114
00:05:04,980 --> 00:05:08,720
So not only did they run into
issues with synchronization,

115
00:05:08,720 --> 00:05:10,670
but certainly discovered
problems with races.

116
00:05:10,670 --> 00:05:12,961
And I think they're the only
team who have successfully

117
00:05:12,961 --> 00:05:15,860
hung their PlayStation
3 several times,

118
00:05:15,860 --> 00:05:21,050
and required a reboot
quite a few times.

119
00:05:21,050 --> 00:05:21,550
OK.

120
00:05:21,550 --> 00:05:23,590
So next, best effort awards.

121
00:05:23,590 --> 00:05:26,050
These guys should really be
highlighted, because they

122
00:05:26,050 --> 00:05:27,956
worked largely on their own.

123
00:05:27,956 --> 00:05:29,011
Is that true?

124
00:05:29,011 --> 00:05:29,510
Yes.

125
00:05:29,510 --> 00:05:33,090
Yeah, and that's certainly true.

126
00:05:33,090 --> 00:05:36,620
And there was certainly
a lot of people

127
00:05:36,620 --> 00:05:38,760
who were impressed by the
Global Illumination stuff

128
00:05:38,760 --> 00:05:41,410
yesterday, and sort
of mentioned that.

129
00:05:41,410 --> 00:05:43,820
And what I'm hoping to do
is take the judges' feedback

130
00:05:43,820 --> 00:05:45,930
and sort of summarize
that for each team,

131
00:05:45,930 --> 00:05:48,260
so you can have some
constructive criticism back

132
00:05:48,260 --> 00:05:49,810
that might be helpful.

133
00:05:49,810 --> 00:05:51,250
Molecular Dynamic
Simulation work

134
00:05:51,250 --> 00:05:52,930
that you heard
about today, I think

135
00:05:52,930 --> 00:05:56,240
certainly a valiant effort
on his part, as well.

136
00:05:56,240 --> 00:05:58,880
The most complete project,
I think this would probably

137
00:05:58,880 --> 00:06:01,290
be Backgammon Tutor's project.

138
00:06:01,290 --> 00:06:04,550
I think in terms of scope
it was well defined,

139
00:06:04,550 --> 00:06:06,550
in terms of what they
actually completed.

140
00:06:06,550 --> 00:06:08,830
I think they came out
with a really cool game.

141
00:06:08,830 --> 00:06:11,120
My money was on them all
along to actually win

142
00:06:11,120 --> 00:06:11,900
the first prize.

143
00:06:11,900 --> 00:06:15,780
But I think the Ray Tracing team
actually blew everybody away

144
00:06:15,780 --> 00:06:17,730
with their demo.

145
00:06:17,730 --> 00:06:21,030
And I think they
were a clear winner.

146
00:06:21,030 --> 00:06:24,760
They really tackled sort
of their paralyzation,

147
00:06:24,760 --> 00:06:25,540
descendizing.

148
00:06:25,540 --> 00:06:27,530
They had a really cool
frame buffer demo,

149
00:06:27,530 --> 00:06:29,238
so they figured out
how to actually write

150
00:06:29,238 --> 00:06:32,110
to the frame buffer
and do that simulation,

151
00:06:32,110 --> 00:06:34,410
or do the graphics rendering.

152
00:06:34,410 --> 00:06:37,130
And I think some mumbling
about a real cool demo

153
00:06:37,130 --> 00:06:38,830
that they've gotten
working overnight.

154
00:06:38,830 --> 00:06:41,580
So we'll try to set up
over in [? Stoddart, ?]

155
00:06:41,580 --> 00:06:44,430
where we have pizza, and you
guys can take a look at that.

156
00:06:44,430 --> 00:06:49,490
So with that, I'll hand out
the certificates and the awards

157
00:06:49,490 --> 00:06:50,890
offline.

158
00:06:50,890 --> 00:06:53,520
But you've at this
point been Cell-ified.

159
00:06:53,520 --> 00:06:57,250
And the cake is in the
back says, Congratulations

160
00:06:57,250 --> 00:06:58,660
on your Cell-ification.

161
00:06:58,660 --> 00:07:00,020
So--

162
00:07:00,020 --> 00:07:01,760
Resistance is futile.

163
00:07:01,760 --> 00:07:03,460
That's right.

164
00:07:03,460 --> 00:07:05,400
So thanks a lot for
taking the course.

165
00:07:05,400 --> 00:07:07,440
I think we've learned
a lot as instructors.

166
00:07:07,440 --> 00:07:11,300
I'm hoping to do this again
with Saman next year, next IAP,

167
00:07:11,300 --> 00:07:14,000
and maybe eventually turn it
into a regular semester course.

168
00:07:14,000 --> 00:07:16,620
I think there's a lot of
things we can do better,

169
00:07:16,620 --> 00:07:19,210
and hopefully you've told us
some ideas of how to do that.

170
00:07:19,210 --> 00:07:24,020
So with that, thanks, and start
enjoying cake in the back.