1
00:00:00,000 --> 00:00:12,080

2
00:00:12,080 --> 00:00:13,799
>> JAMES cuff: Hi, nzuri
alasiri, kila mtu.

3
00:00:13,799 --> 00:00:14,715
Jina langu ni James cuff.

4
00:00:14,715 --> 00:00:18,970
Mimi nina Dean Naibu Utafiti
Kompyuta hapa Chuo Kikuu cha Harvard.

5
00:00:18,970 --> 00:00:24,540
Na leo mimi nina kwenda kuzungumza na wewe kuhusu
nini wadogo-out kompyuta ni muhimu.

6
00:00:24,540 --> 00:00:26,810
>> Kwa hiyo mimi nadhani, kwanza up, ambaye ni guy hii?

7
00:00:26,810 --> 00:00:27,750
Kwa nini niko hapa?

8
00:00:27,750 --> 00:00:29,200
Kwa nini mimi kuzungumza na wewe?

9
00:00:29,200 --> 00:00:33,730
Nina background katika kisayansi
kompyuta na kompyuta utafiti,

10
00:00:33,730 --> 00:00:38,530
kukaza mwendo nyuma United
Kingdom-- Wellcome Trust Sanger

11
00:00:38,530 --> 00:00:43,270
Taasisi ya genome-- binadamu na
basi hivi karibuni zaidi katika United States

12
00:00:43,270 --> 00:00:50,170
kufanya kazi katika Broad na wengine Tukufu
maeneo ya kujifunza, kama vile Harvard.

13
00:00:50,170 --> 00:00:53,930
>> Mimi nadhani nini kwamba kweli maana yake ni
kwamba mimi nina nafuu Masi bio

14
00:00:53,930 --> 00:00:55,740
Fizikia.

15
00:00:55,740 --> 00:01:01,250
Hivyo kile haki na I got kuwaambia
kuhusu kiwango-out kompyuta?

16
00:01:01,250 --> 00:01:03,570
Kuna hata hivyo ni.

17
00:01:03,570 --> 00:01:09,530
Miaka 18 au hivyo nimekuwa tu kuonekana zaidi
ongezeko makubwa katika kiwango utata

18
00:01:09,530 --> 00:01:13,570
na ufanisi wa jumla
ya kompyuta mifumo.

19
00:01:13,570 --> 00:01:18,890
>> Wakati mimi alikuwa akifanya PhD yangu katika Oxford, mimi
ilikuwa pretty msisimko na 200 megahertz

20
00:01:18,890 --> 00:01:23,830
Silicon Graphics mashine na 18
gigabytes ya kuhifadhi na CPU moja.

21
00:01:23,830 --> 00:01:24,910
Mara iliyopita.

22
00:01:24,910 --> 00:01:29,860
Kama kufunga mbele sasa, sisi ni inazunguka
zaidi ya 60,000 CPUs hapa katika Harvard.

23
00:01:29,860 --> 00:01:32,810
Mashirika mengine mengi
ni inazunguka wengi zaidi.

24
00:01:32,810 --> 00:01:37,740
>> takeaway muhimu kutokana na hili ni
kwamba wadogo ni sasa si tu kuepukika,

25
00:01:37,740 --> 00:01:41,910
ni kilichotokea na ni
kwenda kuendelea kutokea.

26
00:01:41,910 --> 00:01:44,760
Basi hebu, kwa sasa, aina
ya rewind na kuzungumza kwa haraka sana

27
00:01:44,760 --> 00:01:50,530
kuhusu sayansi, favorite yangu
somo, njia ya kisayansi.

28
00:01:50,530 --> 00:01:53,180
>> Kama wewe ni kuwa mwanasayansi,
una kufanya mambo muhimu chache.

29
00:01:53,180 --> 00:01:56,140
Kama huna kufanya mambo haya unaweza
si kufikiria mwenyewe mwanasayansi

30
00:01:56,140 --> 00:02:03,250
na utakuwa mapambano kuwa na uwezo wa
kuelewa eneo lako la nidhamu.

31
00:02:03,250 --> 00:02:07,290
>> Hivyo kwanza ya yote, ungekuwa kuunda
swali lako, wewe kuzalisha hypotheses,

32
00:02:07,290 --> 00:02:09,289
lakini muhimu zaidi, wewe
kutabiri results-- yako

33
00:02:09,289 --> 00:02:13,090
una nadhani kama
nini matokeo itakuwa.

34
00:02:13,090 --> 00:02:19,560
Na kisha hatimaye, unaweza mtihani yako
hypothesis na kuchambua matokeo yako.

35
00:02:19,560 --> 00:02:25,460
>> Hivyo njia hii ya kisayansi
muhimu sana katika kompyuta.

36
00:02:25,460 --> 00:02:28,450
Computing wa wote utabiri
na kuwa na uwezo wa kupima matokeo yako

37
00:02:28,450 --> 00:02:33,660
ni sehemu muhimu ya nini tunahitaji
kufanya katika njia ya kisayansi.

38
00:02:33,660 --> 00:02:37,310
Utabiri haya na testings
ni kweli cornerstones mbili

39
00:02:37,310 --> 00:02:42,350
ya mbinu za kisayansi, na kila
zinahitaji maendeleo makubwa zaidi

40
00:02:42,350 --> 00:02:45,240
katika computation kisasa.

41
00:02:45,240 --> 00:02:51,210
>> nguzo mbili za sayansi ni kwamba
ya nadharia na ile ya majaribio.

42
00:02:51,210 --> 00:02:54,300
Na hivi karibuni zaidi,
kompyuta ni mara nyingi zilizotajwa

43
00:02:54,300 --> 00:02:58,090
kama kuwa nguzo ya tatu ya sayansi.

44
00:02:58,090 --> 00:03:01,440
Hivyo kama wewe ni kuangalia wanafunzi hii,
una kabisa hakuna shinikizo.

45
00:03:01,440 --> 00:03:03,960

46
00:03:03,960 --> 00:03:08,720
Tatu nguzo ya science-- hakuna kubwa
kompyuta deal--, aina ya muhimu.

47
00:03:08,720 --> 00:03:14,000
Nimefurahi hii ni kompyuta sehemu
wa sayansi ya kompyuta bila shaka 50.

48
00:03:14,000 --> 00:03:16,220
>> Hivyo kutosha wa background.

49
00:03:16,220 --> 00:03:20,226
Nataka kukuambia mpango wa nini
tunakwenda kuzungumzia leo.

50
00:03:20,226 --> 00:03:22,870
Mimi nina kwenda juu ya baadhi ya historia.

51
00:03:22,870 --> 00:03:25,250
Mimi nina kwenda kueleza kwa nini sisi got hapa.

52
00:03:25,250 --> 00:03:27,750
Mimi nina kwenda kuzungumza kuhusu baadhi
historia ya kompyuta

53
00:03:27,750 --> 00:03:33,890
hapa katika Harvard, baadhi
shughuli mbalimbali ya vyombo vya habari ya kijamii,

54
00:03:33,890 --> 00:03:36,200
kijani things-- sana
passionate juu ya mambo yote

55
00:03:36,200 --> 00:03:43,640
green-- storage-- kompyuta storage--
jinsi machafuko huathiri wadogo-nje nje mifumo,

56
00:03:43,640 --> 00:03:45,640
na mifumo ya distributive hasa.

57
00:03:45,640 --> 00:03:48,473
>> Na kisha mimi nina kwenda kugusa juu ya baadhi
wa wadogo-out vifaa kwamba ni

58
00:03:48,473 --> 00:03:51,370
wanatakiwa kuwa na uwezo wa
kufanya kompyuta katika kiwango.

59
00:03:51,370 --> 00:03:55,830
Na kisha hatimaye, tunakwenda
wrap up na baadhi ya sayansi kutisha.

60
00:03:55,830 --> 00:04:00,894
>> Kwa hiyo, hebu chukua dakika ya
kuangalia historia yetu halisi.

61
00:04:00,894 --> 00:04:01,810
Computing ina tolewa.

62
00:04:01,810 --> 00:04:07,370
Hivyo tangu '60s, kila
mbali kwa njia ya leo,

63
00:04:07,370 --> 00:04:11,260
tumeona kimsingi mabadiliko ya
wigo kutoka kompyuta centralized

64
00:04:11,260 --> 00:04:14,679
kwa zifikishwe kompyuta, kwa
shirikishi na kisha kujitegemea

65
00:04:14,679 --> 00:04:15,970
kompyuta na haki tena.

66
00:04:15,970 --> 00:04:17,709
>> Na napenda kutolea maelezo kwamba kidogo.

67
00:04:17,709 --> 00:04:20,370
Wakati sisi kwanza kuanza mbali na
kompyuta, tulikuwa mainframes.

68
00:04:20,370 --> 00:04:22,824
Walikuwa inordinately
ghali vifaa.

69
00:04:22,824 --> 00:04:23,990
Kila kitu alikuwa na kuwa pamoja.

70
00:04:23,990 --> 00:04:25,556
kompyuta ilikuwa ngumu.

71
00:04:25,556 --> 00:04:29,060
Unaweza kuona, ni kujazwa vyumba na
kulikuwa na waendeshaji na kanda

72
00:04:29,060 --> 00:04:32,780
na kila aina ya Whirry,
clicky, vifaa spinny.

73
00:04:32,780 --> 00:04:39,930
>> Karibu '70s' 80s mapema, wewe kuanza
kuona matokeo ya mashine ya faksi.

74
00:04:39,930 --> 00:04:43,620
Hivyo wewe ni mapya ya kuona kompyuta
kuanza kuonekana nyuma katika maabara

75
00:04:43,620 --> 00:04:45,880
na kuwa karibu na wewe.

76
00:04:45,880 --> 00:04:49,800
kupanda binafsi
kompyuta, hakika

77
00:04:49,800 --> 00:04:57,460
katika '80s, sehemu ya kwanza ya
muongo mmoja, kweli iliyopita kompyuta.

78
00:04:57,460 --> 00:04:59,570
>> Na kuna kidokezo katika
cheo, kwa sababu

79
00:04:59,570 --> 00:05:04,080
ilikuwa inaitwa kompyuta binafsi,
ambayo maana yake ni mali ya wewe.

80
00:05:04,080 --> 00:05:07,630
Hivyo kama mageuzi ya
kompyuta kuendelea,

81
00:05:07,630 --> 00:05:10,530
watu waligundua kwamba yao binafsi
kompyuta ilikuwa si kweli kubwa ya kutosha

82
00:05:10,530 --> 00:05:15,020
kuwa na uwezo wa kufanya kitu chochote sifa ya mtu yoyote,
au sifa muhimu, katika sayansi.

83
00:05:15,020 --> 00:05:17,790
>> Na hivyo folks kuanza
kuendeleza mtandao kifaa

84
00:05:17,790 --> 00:05:21,920
madereva kuwa na uwezo wa kuungana PC
pamoja na kuwa na uwezo wa kujenga makundi.

85
00:05:21,920 --> 00:05:26,430
Na hivyo hii akazaa era
ya Beowulf nguzo.

86
00:05:26,430 --> 00:05:32,470
Linux ililipuka kama jibu kwa
wamiliki mfumo wa uendeshaji, gharama zote

87
00:05:32,470 --> 00:05:33,650
na utata.

88
00:05:33,650 --> 00:05:36,530
>> Na kisha, hapa sisi ni leo,
ambapo, lakini tena, tuko

89
00:05:36,530 --> 00:05:40,610
wanakabiliwa na vyumba kamili ya kompyuta
vifaa na uwezo

90
00:05:40,610 --> 00:05:44,570
swipe moja ya kadi ya mikopo na kupata
upatikanaji wa vifaa hizi kompyuta,

91
00:05:44,570 --> 00:05:45,290
remotely.

92
00:05:45,290 --> 00:05:49,680
>> Na hivyo basi unaweza kuona, katika
suala la historia athari

93
00:05:49,680 --> 00:05:52,180
jinsi hatuwezi kompyuta
leo, ni dhahiri

94
00:05:52,180 --> 00:05:56,090
tolewa kutoka mashine
vyumba kamili ya kompyuta

95
00:05:56,090 --> 00:05:59,160
kupitia baadhi kompyuta binafsi
njia yote haki ya nyuma tena

96
00:05:59,160 --> 00:06:02,400
mashine ya vyumba kamili ya kompyuta.

97
00:06:02,400 --> 00:06:06,620
>> Hivyo hii ni nguzo yangu ya kwanza.

98
00:06:06,620 --> 00:06:10,170
Hivyo mwaka 2000, sisi kujengwa
mfumo wa kompyuta katika Ulaya

99
00:06:10,170 --> 00:06:13,900
kwa ufanisi kutolea maelezo
genome binadamu.

100
00:06:13,900 --> 00:06:16,521
Kuna mengi ya teknolojia
waliotajwa kwenye upande wa kulia

101
00:06:16,521 --> 00:06:18,520
pale kwamba, kwa bahati mbaya,
tena na sisi.

102
00:06:18,520 --> 00:06:23,460
Ni kupita mbali na
teknolojia kubwa angani.

103
00:06:23,460 --> 00:06:26,610
>> mashine yenyewe pengine
sawa wachache heshima Laptops

104
00:06:26,610 --> 00:06:29,020
leo, na kwamba tu aina ya inaonyesha wewe.

105
00:06:29,020 --> 00:06:36,260
Hata hivyo, hatukuwa makini kutolea maelezo
genome binadamu na wote hifadhi ni

106
00:06:36,260 --> 00:06:43,190
kwa karatasi hii hasa katika
Nature na matatizo ya data

107
00:06:43,190 --> 00:06:45,380
kuwa umma au binafsi.

108
00:06:45,380 --> 00:06:48,610
>> Hivyo hii ni ajabu, haki?

109
00:06:48,610 --> 00:06:50,280
Hivyo sisi tumepewa genome binadamu.

110
00:06:50,280 --> 00:06:51,510
Tumefanya kompyuta.

111
00:06:51,510 --> 00:06:53,400
Mimi nina hisia sana radhi mwenyewe.

112
00:06:53,400 --> 00:06:59,090
Mimi akavingirisha hadi Harvard mwaka 2006,
hisia mengi chini radhi na mimi mwenyewe.

113
00:06:59,090 --> 00:07:00,210
>> Hii ni nini mimi kurithi.

114
00:07:00,210 --> 00:07:03,575
Hii ni idara
mail na faili server.

115
00:07:03,575 --> 00:07:05,450
Unaweza kuona hapa kuna
kidogo ya mkanda

116
00:07:05,450 --> 00:07:07,710
hiyo kutumika kushikilia mfumo wa pamoja.

117
00:07:07,710 --> 00:07:09,890
Hii ni leseni yetu na magazeti server.

118
00:07:09,890 --> 00:07:13,990
Mimi nina uhakika pretty kuna labda nywila
juu ya baadhi ya haya Post-ni Maelezo.

119
00:07:13,990 --> 00:07:16,560

120
00:07:16,560 --> 00:07:17,360
>> Si kutisha.

121
00:07:17,360 --> 00:07:18,530
Pretty mbali na kutisha.

122
00:07:18,530 --> 00:07:22,060
Na hivyo, mimi kutambua hili chati kidogo
kwamba mimi ilionyesha tangu mwanzo

123
00:07:22,060 --> 00:07:25,350
kwa kushirikiana na umiliki
nyuma kugawana,

124
00:07:25,350 --> 00:07:27,930
kwamba sisi zinahitajika kubadili mchezo.

125
00:07:27,930 --> 00:07:31,330
Na hivyo sisi iliyopita mchezo
kwa kutoa motisha.

126
00:07:31,330 --> 00:07:34,250
Na hivyo binadamu, kama hii
kidogo Wikipedia article

127
00:07:34,250 --> 00:07:35,990
anasema hapa, viumbe yetu makusudi.

128
00:07:35,990 --> 00:07:39,250
Na utafiti wa motisha
miundo ni muhimu kwa utafiti

129
00:07:39,250 --> 00:07:41,100
ya shughuli za kiuchumi.

130
00:07:41,100 --> 00:07:44,580
>> Hivyo sisi ilianza motisha
Kitivo yetu na watafiti wetu.

131
00:07:44,580 --> 00:07:47,720
Na hivyo sisi incentivized yao na
kweli kubwa mfumo wa kompyuta.

132
00:07:47,720 --> 00:07:52,720
Hivyo mwaka 2008, sisi kujengwa 4096
processor machine-- 10 racks,

133
00:07:52,720 --> 00:07:54,470
wanandoa kilowatts mia ya madaraka.

134
00:07:54,470 --> 00:07:56,178
>> Jambo ambalo nadhani ni
kuvutia ni haina

135
00:07:56,178 --> 00:07:58,300
jambo ambapo wewe ni katika mzunguko.

136
00:07:58,300 --> 00:08:03,510
Hii kiasi kama hicho cha nguvu na
compute, nguvu ni mara kwa mara.

137
00:08:03,510 --> 00:08:06,270
Ilikuwa kilowatts 200 wakati sisi
na kujenga mifumo katika Ulaya.

138
00:08:06,270 --> 00:08:09,770
Ni kilowatts mia mbili
mwaka 2008, na kwamba

139
00:08:09,770 --> 00:08:15,820
inaonekana kuwa [? quanter?] ya ndogo
chuo kikuu makao mifumo ya kompyuta.

140
00:08:15,820 --> 00:08:20,540
>> Hivyo Harvard today-- kufunga mbele, mimi nina hakuna
tena kusikitisha panda, kabisa na furaha panda.

141
00:08:20,540 --> 00:08:25,860
Sisi 60-isiyo ya kawaida elfu mzigo uwiano
CPUs, na kupanda yao kwa kiwango kikubwa.

142
00:08:25,860 --> 00:08:28,780
Tuna 15 petabytes ya
kuhifadhi, pia kupanda.

143
00:08:28,780 --> 00:08:30,720
Tena, hii kilowatt 200
increment, tunaonekana

144
00:08:30,720 --> 00:08:33,000
kuwa na kuongeza kuwa
kila sita au hivyo miezi.

145
00:08:33,000 --> 00:08:35,480
Kura na kura ya mashine virtual.

146
00:08:35,480 --> 00:08:37,620
Na muhimu zaidi,
megawati zipatazo 1.8

147
00:08:37,620 --> 00:08:39,669
ya kompyuta vifaa vya utafiti.

148
00:08:39,669 --> 00:08:41,820
>> Na mimi nina kwenda kuja
nyuma hii baadaye,

149
00:08:41,820 --> 00:08:46,913
kwa nini mimi sasa hakuna tena lazima
kuhesabu ni kiasi gani CPU tuna,

150
00:08:46,913 --> 00:08:48,980
lakini jinsi kubwa ni muswada wa umeme.

151
00:08:48,980 --> 00:08:52,690
20 wengine ili kujitolea
utafiti kompyuta wafanyakazi.

152
00:08:52,690 --> 00:08:57,250
Na muhimu zaidi, tuko
mapya na kukua GPGPUs yetu.

153
00:08:57,250 --> 00:09:05,030
Mimi nilikuwa kujikongoja kiasi gani ya hii
ni kuwa aliongeza kwa msingi wa siku hadi siku.

154
00:09:05,030 --> 00:09:07,310
Hivyo, historia somo juu, haki?

155
00:09:07,310 --> 00:09:11,280
>> Hivyo ni jinsi gani sisi kupata huko kutoka hapa?

156
00:09:11,280 --> 00:09:14,560
Hebu tuangalie baadhi ya kisasa
wadogo-nje compute mifano.

157
00:09:14,560 --> 00:09:18,290

158
00:09:18,290 --> 00:09:23,230
Mimi nina kidogo obsessed na
ukubwa na ukubwa wa kijamii vyombo vya habari.

159
00:09:23,230 --> 00:09:30,850
Kuna idadi ya sana
mafanikio makubwa kompyuta wadogo

160
00:09:30,850 --> 00:09:34,820
mashirika sasa juu ya sayari,
kutoa msaada na huduma

161
00:09:34,820 --> 00:09:36,810
kwetu sote.

162
00:09:36,810 --> 00:09:39,340
Hivyo hiyo ni disclaimer.

163
00:09:39,340 --> 00:09:42,990
>> Na mimi nataka kuanza na
idadi ya ounces katika Instagram.

164
00:09:42,990 --> 00:09:48,336
Ni kweli si
kusababisha-katika utani, ni

165
00:09:48,336 --> 00:09:50,460
kwamba hata funny,
kweli, kuja kufikiria hivyo.

166
00:09:50,460 --> 00:09:52,751
Lakini anyway, tunakwenda
kuangalia ounces katika Instagram.

167
00:09:52,751 --> 00:09:55,260
Na sisi ni kwenda kuanza
na "nyuki yangu na maua."

168
00:09:55,260 --> 00:09:57,600
Nilikuwa katika [inaudible] Kijiji
na mimi alichukua picha kidogo

169
00:09:57,600 --> 00:10:00,460
ya nyuki ameketi juu ya maua.

170
00:10:00,460 --> 00:10:03,270
Na kisha mimi kuanza kufikiri juu ya
nini hii kwa kweli maana.

171
00:10:03,270 --> 00:10:07,013
Na mimi alichukua picha hii mbali ya simu yangu
na kuhesabiwa jinsi ka wengi ni ndani yake,

172
00:10:07,013 --> 00:10:09,070
na ni kuhusu 256 kilobytes.

173
00:10:09,070 --> 00:10:13,550
Ambayo wakati mimi kuanza, ingekuwa kimsingi
kujaza 5 na 1/4 inch floppy.

174
00:10:13,550 --> 00:10:15,340
Na kuanza kufikiri, vizuri, hiyo ni cool.

175
00:10:15,340 --> 00:10:18,630
>> Na mimi kuanza kuangalia na kufanya
baadhi ya utafiti kwenye mtandao.

176
00:10:18,630 --> 00:10:22,490
Na nimeona kwamba Instagram
ina Maus milioni 200.

177
00:10:22,490 --> 00:10:25,105
Mimi nilikuwa si kweli kwamba
uhakika nini Mau mara.

178
00:10:25,105 --> 00:10:28,960
Na Mau, hapa chini, ni
kila mwezi kazi user.

179
00:10:28,960 --> 00:10:34,270
>> Hivyo, MAUs-- milioni 200 pretty cool.

180
00:10:34,270 --> 00:10:38,190
Bilioni 20 photographs-- hivyo
mengi kabisa ya picha.

181
00:10:38,190 --> 00:10:42,300
Milioni 60 picha mpya
kila siku

182
00:10:42,300 --> 00:10:46,990
kuja nje saa kuhusu 0.002 GIG kwa photo.

183
00:10:46,990 --> 00:10:51,290
Hiyo ni petabytes tano
ya disk haki tu huko.

184
00:10:51,290 --> 00:10:55,480
Na kwamba ni kweli si sehemu ya kati
ya nini tunakwenda kuzungumzia.

185
00:10:55,480 --> 00:10:57,830
Hiyo ni viazi ndogo.

186
00:10:57,830 --> 00:11:00,710
Au kama sisi kusema katika England, Spuds vidogo.

187
00:11:00,710 --> 00:11:05,050
>> Hivyo hebu tuangalie tembo halisi
katika room-- nyuso kipekee.

188
00:11:05,050 --> 00:11:09,170
Tena, hebu kupima katika
Quanta hii mpya kuwaita Mau.

189
00:11:09,170 --> 00:11:13,260
Facebook yenyewe ina bilioni 1.3 Maus.

190
00:11:13,260 --> 00:11:17,510
Whatsapp, ambayo nilikuwa hata
habari za hadi hivi karibuni, ni

191
00:11:17,510 --> 00:11:23,260
baadhi ya huduma aina ya ujumbe,
ni milioni 500 Maus.

192
00:11:23,260 --> 00:11:26,620
Instagram, ambayo sisi tu
kuongelea, milioni 200 Maus.

193
00:11:26,620 --> 00:11:29,370
Na Mtume, ambayo ni
huduma nyingine ujumbe,

194
00:11:29,370 --> 00:11:31,120
Pia Maus milioni 200.

195
00:11:31,120 --> 00:11:35,920
>> Hivyo jumla kuwa juu, ni kuhusu
Bilioni 2.2 jumla watumiaji.

196
00:11:35,920 --> 00:11:39,880
Ni wazi kuna baadhi ya mwingiliano, lakini hiyo ni
sawa na ya tatu ya dunia.

197
00:11:39,880 --> 00:11:44,270
Na wao kutuma kitu katika
kanda ya ujumbe bilioni 12 kwa siku.

198
00:11:44,270 --> 00:11:46,680
Na tena, kuna tu 7
watu bilioni katika dunia.

199
00:11:46,680 --> 00:11:48,550
Si kila mtu ana smartphone.

200
00:11:48,550 --> 00:11:53,960
Hivyo hii ni namba mwendawazimu.

201
00:11:53,960 --> 00:12:02,050
>> Na mimi nina kwenda kusema kwamba si
hata juu ya kuhifadhi au compute.

202
00:12:02,050 --> 00:12:05,610
Na kunukuu wimbo,
ni wote kuhusu kwamba graph.

203
00:12:05,610 --> 00:12:09,045
Hapa ni Meghan wetu lovely Trainor chini
hapa, kuimba kuhusu bass wote.

204
00:12:09,045 --> 00:12:12,570
Kumbuka, yeye pia ana kabisa
kidogo ya bass herself-- 207,

205
00:12:12,570 --> 00:12:16,460
vizuri watu 218,000,000 tumeona
hii mwanamke vijana kuimba wimbo wake.

206
00:12:16,460 --> 00:12:19,910
>> Hivyo hoja yangu ni
ni wote kuhusu graph.

207
00:12:19,910 --> 00:12:23,480
Hivyo sisi alichukua baadhi wazi chanzo programu
na kuanza kuangalia graph.

208
00:12:23,480 --> 00:12:27,740
Na hii ni LinkedIn, hivyo hii
ni Facebook kwa watu wa zamani.

209
00:12:27,740 --> 00:12:29,910
Na hivyo, hii ni yangu LinkedIn graph.

210
00:12:29,910 --> 00:12:34,080
Nina nodes 1,200 au hivyo,
kinachojulikana "Friends."

211
00:12:34,080 --> 00:12:36,360
Na hapa ni mimi saa ya juu.

212
00:12:36,360 --> 00:12:38,140
Na hapa ni wote wa kuunganishwa.

213
00:12:38,140 --> 00:12:40,570
>> Sasa, kufikiri nyuma hadithi Instagram.

214
00:12:40,570 --> 00:12:42,815
Kila mmoja wa haya ni
si tu picha, ni

215
00:12:42,815 --> 00:12:46,860
ina plethora mzima wa uhusiano
kati ya mtu binafsi hasa hili

216
00:12:46,860 --> 00:12:48,220
na wengine wengi.

217
00:12:48,220 --> 00:12:52,190
Hii ni kipande kuu ni ama
mdudu katika graph kuchora algorithm,

218
00:12:52,190 --> 00:12:55,982
au hii labda David
Malan, mimi nina uhakika bado.

219
00:12:55,982 --> 00:12:57,690
Hivyo unaweza redraw
Diagram katika kila aina

220
00:12:57,690 --> 00:13:02,510
ya ways-- gephi.gihub.io ni mahali ambapo
unaweza kuvuta kwamba programu kutoka.

221
00:13:02,510 --> 00:13:05,410
Ni kweli baridi kwa kuwa
uwezo wa kuandaa jamii.

222
00:13:05,410 --> 00:13:08,640
Unaweza kuona hapa, hii ni Harvard na
maeneo mengine mbalimbali kwamba nimepata kazi,

223
00:13:08,640 --> 00:13:12,160
kwa sababu hii ni yangu data yanayohusiana na kazi.

224
00:13:12,160 --> 00:13:15,080
>> Hivyo tu kufikiri kuhusu utata
ya graph na wote wa data

225
00:13:15,080 --> 00:13:17,070
kwamba wewe vuta pamoja na.

226
00:13:17,070 --> 00:13:20,870
Hivyo wakati huo huo, nyuma katika FriendFace, haki?

227
00:13:20,870 --> 00:13:24,360
Tuliangalia data Instagram kwamba
na utaratibu wa petabytes tano.

228
00:13:24,360 --> 00:13:25,300
Hakuna mpango mkubwa.

229
00:13:25,300 --> 00:13:28,830
Bado mengi kabisa ya data, lakini hakuna kubwa
kuhusika katika mpango mkubwa wa mambo.

230
00:13:28,830 --> 00:13:33,850
>> Kutoka makala hii kwenye mtandao wa zamani,
"Kuongeza Facebook data ghala

231
00:13:33,850 --> 00:13:36,250
300 petabytes. "

232
00:13:36,250 --> 00:13:38,110
Hiyo ni tofauti nzima
Changer mchezo sasa,

233
00:13:38,110 --> 00:13:40,234
wakati wewe ni mapya ya
kufikiria data na graph

234
00:13:40,234 --> 00:13:41,690
na nini kuleta pamoja na.

235
00:13:41,690 --> 00:13:47,480
Na data zao juu ni kuongezeka ya
utaratibu wa 600 terrabytes siku.

236
00:13:47,480 --> 00:13:52,980
>> Sasa, unajua, vizuri, then--
I mean, terrabytes 600 kwa siku,

237
00:13:52,980 --> 00:13:55,670
300 petabytes-- wao uko
pia sasa kuanza

238
00:13:55,670 --> 00:13:58,550
kupata wasiwasi sana kuhusu
jinsi ya kuweka mambo haya

239
00:13:58,550 --> 00:14:01,160
na kuhakikisha data hii anakaa kote.

240
00:14:01,160 --> 00:14:04,630
Na hii muungwana hapa,
Jay Parikh, ni kutafuta

241
00:14:04,630 --> 00:14:08,250
jinsi ya kuhifadhi exabyte ya data.

242
00:14:08,250 --> 00:14:10,180
>> Tu kwa ajili ya wale wa wewe
ambao ni kuangalia pamoja

243
00:14:10,180 --> 00:14:13,940
nyumbani, exabyte-- 10 hadi 18.

244
00:14:13,940 --> 00:14:18,210
Ni got Wikipedia yake mwenyewe
ukurasa, ni kwamba kubwa ya idadi.

245
00:14:18,210 --> 00:14:23,120
Hiyo ni ukubwa na ukubwa wa nini tuko
kuangalia, kuwa na uwezo wa kuhifadhi data.

246
00:14:23,120 --> 00:14:27,090
Na hawa guys si mucking karibu,
re kuhifadhi kwamba kiasi cha data.

247
00:14:27,090 --> 00:14:29,550
Hivyo moja ya dalili kwamba
re kuangalia hapa

248
00:14:29,550 --> 00:14:32,185
ni data vituo kwa ajili ya
kinachojulikana kuhifadhi baridi.

249
00:14:32,185 --> 00:14:35,020

250
00:14:35,020 --> 00:14:36,470
>> Ambayo huleta mimi kuwa ya kijani.

251
00:14:36,470 --> 00:14:38,340
Na hapa ni Kermit.

252
00:14:38,340 --> 00:14:43,050
Yeye na mimi agree-- ni sana
vigumu kuwa kijani,

253
00:14:43,050 --> 00:14:44,920
lakini sisi kutoa ni kujaribu bora yetu.

254
00:14:44,920 --> 00:14:47,430
Kermit hawawezi kusaidia, ana
kuwa kijani wakati wote,

255
00:14:47,430 --> 00:14:49,945
hawezi kuchukua kijani-ness yake mbali wakati wote.

256
00:14:49,945 --> 00:14:55,410
>> Hivyo, kuwa concepts-- a
chache aina ya dhana ya msingi

257
00:14:55,410 --> 00:14:59,510
ya greenness, wakati
inahusiana na kompyuta.

258
00:14:59,510 --> 00:15:05,510
moja kwamba ni muhimu zaidi
ni longevity ya bidhaa.

259
00:15:05,510 --> 00:15:09,405
Kama bidhaa yako ina maisha short,
huwezi, kwa ufafanuzi, kuwa kijani.

260
00:15:09,405 --> 00:15:13,280
nishati kuchukuliwa na utengenezaji
disk drive, Motherboard, kompyuta

261
00:15:13,280 --> 00:15:17,890
mfumo, kibao, chochote inaweza
kuwa, longevity ya mifumo yako

262
00:15:17,890 --> 00:15:21,700
ni sehemu muhimu ya jinsi ya kijani unaweza kuwa.

263
00:15:21,700 --> 00:15:27,960
>> sehemu muhimu, kama nyote
ni kujenga programu algorithms--

264
00:15:27,960 --> 00:15:30,455
algorithm ya ubaguzi
neno kwa ajili ya programu, haki?

265
00:15:30,455 --> 00:15:34,000
Hivyo, algorithm yako kubuni ni
muhimu kabisa katika suala

266
00:15:34,000 --> 00:15:43,080
ya jinsi wewe ni kwenda kuwa na uwezo wa kufanya
computations haraka na sahihi kwa kutumia

267
00:15:43,080 --> 00:15:44,710
kiasi angalau ya nishati iwezekanavyo.

268
00:15:44,710 --> 00:15:47,280
Na mimi itabidi kupata hii katika kidogo.

269
00:15:47,280 --> 00:15:51,270
>> Kituo cha data design-- ve kuonekana
kwamba tayari tuna maelfu

270
00:15:51,270 --> 00:15:54,870
juu ya maelfu ya mashine, ameketi
kimya kimya katika ndogo, pembe giza

271
00:15:54,870 --> 00:15:57,760
wa dunia, kompyuta.

272
00:15:57,760 --> 00:16:01,670
Rasilimali allocation-- jinsi ya kupata
kwa compute, na kuhifadhi,

273
00:16:01,670 --> 00:16:03,840
njia ya mtandao.

274
00:16:03,840 --> 00:16:08,530
Mifumo ya uendeshaji ni sehemu muhimu ya
hii, na mengi ya virtualization

275
00:16:08,530 --> 00:16:12,080
kuwa na uwezo wa pakiti zaidi na
compute zaidi katika nafasi ndogo.

276
00:16:12,080 --> 00:16:15,530
>> Mimi nitakupa mfano mdogo
kutoka kompyuta utafiti.

277
00:16:15,530 --> 00:16:18,220
Sisi zinahitajika zaidi Ping, zaidi
nguvu, na bomba zaidi.

278
00:16:18,220 --> 00:16:21,030
Sisi zinahitajika kubwa zaidi,
bora, kasi kompyuta,

279
00:16:21,030 --> 00:16:23,390
na zinahitajika kutumia chini juisi.

280
00:16:23,390 --> 00:16:26,856
Na tunaweza kufanya kazi nje ya jinsi ya kufanya hii.

281
00:16:26,856 --> 00:16:29,980
Sijui kama hashtag gowest kama
pengine zimetumika kwa Kardashian,

282
00:16:29,980 --> 00:16:32,560
lakini anyway, gowest.

283
00:16:32,560 --> 00:16:33,220
Na sisi alivyofanya.

284
00:16:33,220 --> 00:16:36,610
>> Sisi ilichukua kazi yetu
na sisi wakiongozwa nje

285
00:16:36,610 --> 00:16:39,660
Magharibi Massachusetts
katika mji mdogo kinu

286
00:16:39,660 --> 00:16:45,000
kuitwa Holyoke, kaskazini tu
ya Chikopee na Springfield.

287
00:16:45,000 --> 00:16:49,280
Sisi alifanya hivyo kwa wanandoa ya sababu.

288
00:16:49,280 --> 00:16:55,150
moja kuu ni kwamba sisi
alikuwa bwawa kubwa sana.

289
00:16:55,150 --> 00:17:00,080
Na hii bwawa kubwa sana ni uwezo wa
kuweka nje plus megawati 30 ya nishati,

290
00:17:00,080 --> 00:17:02,980
na ilikuwa ambacho hakitumiki wakati huo.

291
00:17:02,980 --> 00:17:06,170
>> Muhimu zaidi, sisi pia alikuwa
mtandao ngumu sana

292
00:17:06,170 --> 00:17:07,254
kwamba alikuwa tayari katika mahali.

293
00:17:07,254 --> 00:17:09,711
Kama ukiangalia ambapo mtandao
huenda katika Umoja wa Mataifa,

294
00:17:09,711 --> 00:17:11,230
kinachofuata nyimbo zote treni.

295
00:17:11,230 --> 00:17:14,290
Hii kipande fulani ya mtandao ilikuwa
inayomilikiwa na wenzetu na marafiki

296
00:17:14,290 --> 00:17:16,480
katika Massachusetts
Taasisi ya Teknolojia,

297
00:17:16,480 --> 00:17:19,720
na ilikuwa kimsingi kujengwa
njia yote kufanyika kwa Route 90.

298
00:17:19,720 --> 00:17:24,760
>> Hivyo tulikuwa kubwa mto tick, Route 90
Jibu, tulikuwa short njia ya umbali wa kilomita 100,

299
00:17:24,760 --> 00:17:26,960
na njia ndefu ya juu maili 1,000.

300
00:17:26,960 --> 00:17:29,890
Hatukuwa kufanya kubwa sana
mtandao ganga, kama unaweza kuona hapa,

301
00:17:29,890 --> 00:17:32,990
kimsingi kuweka kiungo katika,
kuwa na uwezo wa kuungana na Holyoke,

302
00:17:32,990 --> 00:17:36,390
lakini tulikuwa wote wa zinazohitajika
infrastructure-- Ping, nguvu, bomba.

303
00:17:36,390 --> 00:17:37,280
Maisha ilikuwa nzuri.

304
00:17:37,280 --> 00:17:38,980
Na tena, bwawa kubwa.

305
00:17:38,980 --> 00:17:42,120
>> Hivyo sisi kujengwa kimsingi Massachusetts
Green High Performance Computing

306
00:17:42,120 --> 00:17:42,850
Center.

307
00:17:42,850 --> 00:17:46,580
Hii ilikuwa kazi ya upendo kwa njia ya tano
universities-- MIT, Harvard, UMass,

308
00:17:46,580 --> 00:17:47,870
Kaskazini mashariki, na BU.

309
00:17:47,870 --> 00:17:49,554
Tano megawati siku moja kushikamana mzigo.

310
00:17:49,554 --> 00:17:51,845
Sisi alifanya kila aina ya ujanja
na airside economizers

311
00:17:51,845 --> 00:17:53,585
kuweka mambo ya kijani.

312
00:17:53,585 --> 00:18:03,330
Na sisi kujengwa nje racks 640-isiyo ya kawaida,
kujitolea kwa ajili ya kompyuta utafiti.

313
00:18:03,330 --> 00:18:08,770
>> Ni alikuwa na umri wa BROWNFIELD tovuti, hivyo sisi
alikuwa na baadhi ya reclamation na baadhi tidy-up

314
00:18:08,770 --> 00:18:10,500
na baadhi safi-up ya tovuti.

315
00:18:10,500 --> 00:18:13,590
Na kisha sisi kuanza
kujenga kituo

316
00:18:13,590 --> 00:18:19,710
na, boom-- lovely kituo na
uwezo wa kukimbia sandbox kompyuta,

317
00:18:19,710 --> 00:18:24,430
kuwa na mikutano na semina,
na pia mkubwa katikati data sakafu.

318
00:18:24,430 --> 00:18:26,007
>> Hapa ni yangu binafsi nzuri.

319
00:18:26,007 --> 00:18:27,590
Mimi nina wazi amevaa koti huo.

320
00:18:27,590 --> 00:18:29,423
Mimi labda moja tu
koti, lakini kuna me

321
00:18:29,423 --> 00:18:34,030
na John Goodhue-- yeye
mkurugenzi mtendaji wa Center--

322
00:18:34,030 --> 00:18:36,740
amesimama katika chumba mashine
sakafu, ambayo, kama unaweza kuona,

323
00:18:36,740 --> 00:18:40,560
ni pretty makubwa, na ni
inakwenda nyuma kwa muda mrefu, kwa muda mrefu njia.

324
00:18:40,560 --> 00:18:44,830
>> Mimi mara nyingi kucheza michezo kuendesha
kutoka Boston kufanyika kwa Holyoke,

325
00:18:44,830 --> 00:18:47,260
kujifanya kuwa mimi nina TCP / IP pakiti.

326
00:18:47,260 --> 00:18:54,290
Na mimi wasiwasi kuhusu latency yangu
kuendesha gari karibu katika gari langu.

327
00:18:54,290 --> 00:18:56,690
Hivyo hiyo ni kipande kijani.

328
00:18:56,690 --> 00:19:00,070
Basi hebu tu kuchukua dakika
na kufikiri kuhusu mwingi.

329
00:19:00,070 --> 00:19:04,060
Hivyo sisi ni kujaribu kwa makini sana
kujenga vituo vya data kwa ufanisi,

330
00:19:04,060 --> 00:19:08,770
kompyuta kwa ufanisi, kufanya vizuri
uteuzi kwa ajili ya vifaa kompyuta

331
00:19:08,770 --> 00:19:12,060
na kutoa, zaidi
muhimu zaidi, maombi yetu,

332
00:19:12,060 --> 00:19:17,860
kuwa ni huduma ya ujumbe
au maombi ya kisayansi.

333
00:19:17,860 --> 00:19:19,110
>> Hivyo hapa ni mwingi.

334
00:19:19,110 --> 00:19:22,762
Safu hiyo ya kimwili, kila
njia ya juu kwa njia ya application--

335
00:19:22,762 --> 00:19:25,220
matumaini kwamba hii ni kwenda
kuwa sehemu nzuri ya kozi yako.

336
00:19:25,220 --> 00:19:31,450
OSI saba safu mfano ni kimsingi,
utakuwa kuishi, kula, na kupumua

337
00:19:31,450 --> 00:19:35,270
hii katika kazi yako kompyuta.

338
00:19:35,270 --> 00:19:37,800
Hii dhana nzima ya kimwili
waya infrastructure--,

339
00:19:37,800 --> 00:19:40,080
cables, vituo vya data, viungo.

340
00:19:40,080 --> 00:19:42,190
Na hii ni kuelezea tu mtandao.

341
00:19:42,190 --> 00:19:44,780
>> Up hapa ni, vizuri, ni wazi,
hii ni slide zamani,

342
00:19:44,780 --> 00:19:49,342
sababu hii lazima kusema HTTP, kwa sababu
hakuna anayejali kuhusu barua pepe rahisi

343
00:19:49,342 --> 00:19:50,550
itifaki ya usafiri, tena.

344
00:19:50,550 --> 00:19:53,960
Ni yote yanayotokea katika nafasi HTTP.

345
00:19:53,960 --> 00:19:55,850
Hivyo hiyo ni ngazi moja ya stack.

346
00:19:55,850 --> 00:19:59,460
>> Hapa ni mwingine seti ya mwingi, ambapo
kuwa server, jeshi, hypervisor,

347
00:19:59,460 --> 00:20:02,470
mgeni, binary maktaba,
na kisha maombi yako.

348
00:20:02,470 --> 00:20:06,070
Au, katika kesi hii, kifaa
dereva, Linux kernel, asili c,

349
00:20:06,070 --> 00:20:08,080
Java virtual mashine,
Java API, basi Java

350
00:20:08,080 --> 00:20:11,220
maombi, na kadhalika na kadhalika.

351
00:20:11,220 --> 00:20:14,090
Hii ni maelezo
ya mashine virtual.

352
00:20:14,090 --> 00:20:15,450
>> Mwingi Mtakatifu, Batman!

353
00:20:15,450 --> 00:20:18,260
Fikiria hili katika
suala la compute kiasi gani

354
00:20:18,260 --> 00:20:20,850
unahitaji kupata kutoka
nini kinatokea hapa,

355
00:20:20,850 --> 00:20:23,110
njia yote hadi juu
ya stack hii, na kisha

356
00:20:23,110 --> 00:20:26,840
kuwa na uwezo wa kufanya halisi yako
utoaji wa maombi.

357
00:20:26,840 --> 00:20:29,130
>> Na kama wewe aina ya
rewind na kuanza kufikiri

358
00:20:29,130 --> 00:20:33,450
kuhusu nini inachukua kutoa
floating uhakika operesheni,

359
00:20:33,450 --> 00:20:37,650
floating uhakika kazi yako ni kiasi
ya soketi, idadi ya cores

360
00:20:37,650 --> 00:20:44,490
katika tundu, saa, ambayo ni
jinsi ya kufunga unaweza saa turnover--

361
00:20:44,490 --> 00:20:47,490
gigahertz nne, mbili
gigahertz-- na kisha simu

362
00:20:47,490 --> 00:20:50,890
wa shughuli unaweza
kufanya katika Hertz aliyopewa.

363
00:20:50,890 --> 00:20:54,350
>> Hivyo wale microprocessors leo
kufanya kati ya nne na 6 flops

364
00:20:54,350 --> 00:20:55,400
kwa saa mzunguko.

365
00:20:55,400 --> 00:20:59,810
Na hivyo moja-msingi 2.5 GIG saa
ina utendaji kinadharia

366
00:20:59,810 --> 00:21:03,490
ya juu flop mega, kutoa au kuchukua.

367
00:21:03,490 --> 00:21:05,940
>> Lakini, kama kwa kila kitu,
tuna uchaguzi.

368
00:21:05,940 --> 00:21:12,280
Hivyo na Intel Core 2, NEHALEM
Sandy Bridge, Haswell, AMD,

369
00:21:12,280 --> 00:21:13,920
kuchukua choices-- yako Intel Atom.

370
00:21:13,920 --> 00:21:17,670
Wote wa architectures haya processor
wote wana njia tofauti kidogo

371
00:21:17,670 --> 00:21:19,650
ya kuweza kuongeza
namba mbili pamoja,

372
00:21:19,650 --> 00:21:23,520
ambayo kimsingi ni
lengo lao katika maisha.

373
00:21:23,520 --> 00:21:24,535
Lazima kuwa mgumu.

374
00:21:24,535 --> 00:21:27,100
Kuna mamilioni ya watu wamekaa
katika vituo vya data, sasa ingawa.

375
00:21:27,100 --> 00:21:30,410
>> Sor, flops kwa watt--
hili ni jambo kubwa.

376
00:21:30,410 --> 00:21:37,780
Hivyo kama mimi wanataka kupata zaidi ya hii
kupata njia stack hii, kasi,

377
00:21:37,780 --> 00:21:41,800
Mimi nimepata kazi na jinsi wengi
yaliyo shughuli hatua ya pili,

378
00:21:41,800 --> 00:21:43,770
Siwezi kufanya, na kisha kuwapa watt.

379
00:21:43,770 --> 00:21:46,160
Na kwa bahati nzuri, folks
kuwa na mawazo kuhusu hili.

380
00:21:46,160 --> 00:21:49,140
>> Hivyo kuna kubwa
kugombea kila mwaka kuona

381
00:21:49,140 --> 00:21:52,310
ambao wanaweza kujenga kompyuta kasi
ambayo inaweza diagonalize tumbo.

382
00:21:52,310 --> 00:21:53,980
Ni wito Juu 500.

383
00:21:53,980 --> 00:21:56,420
Wao kuchukua juu kutoka
bora kompyuta 500

384
00:21:56,420 --> 00:21:58,610
juu ya dunia ambayo yanaweza
diagonalize matrices.

385
00:21:58,610 --> 00:22:00,760
Na kupata baadhi ya matokeo ya ajabu.

386
00:22:00,760 --> 00:22:04,660
>> mengi ya mashine hizo ni
kati ya megawati 10 na 20.

387
00:22:04,660 --> 00:22:09,380
Wanaweza diagonalize matrices
inordinately haraka.

388
00:22:09,380 --> 00:22:13,550
Hawana lazima diagonalized
yao kama ufanisi kwa watt,

389
00:22:13,550 --> 00:22:18,060
hivyo kulikuwa kushinikiza hii kubwa kuangalia
nini kijani 500 orodha bila kuangalia kama.

390
00:22:18,060 --> 00:22:20,360
Na hapa ni orodha kutoka Juni.

391
00:22:20,360 --> 00:22:22,410
Ni lazima kuwe na mwezi mmoja muda mfupi sana.

392
00:22:22,410 --> 00:22:26,590
>> Na wito out-- mimi itabidi kuchukua
juu ya orodha hii tu.

393
00:22:26,590 --> 00:22:32,187
Kuna mawili maalum machines-- moja
kutoka Taasisi ya Teknolojia Tokyo

394
00:22:32,187 --> 00:22:34,520
na mmoja kutoka Cambridge
University nchini Uingereza.

395
00:22:34,520 --> 00:22:37,700
Na hawa wana pretty kubwa
mega flops kwa uwiano wa watt.

396
00:22:37,700 --> 00:22:42,620
Hii ni moja ya 4389, na
moja ijayo chini ni 3,631.

397
00:22:42,620 --> 00:22:47,660
>> Mimi itabidi kueleza tofauti kati ya
hizi mbili, katika slide ijayo.

398
00:22:47,660 --> 00:22:51,320
Lakini haya ni hizi ni
kiasi ukubwa nguzo mtihani.

399
00:22:51,320 --> 00:22:54,732
Hizi ni 34 tu
kilowatts au 52 kilowatts.

400
00:22:54,732 --> 00:22:56,940
Kuna baadhi ya wale kubwa
here-- moja hasa hili

401
00:22:56,940 --> 00:22:58,860
katika Swiss National
Superdatorer Centre.

402
00:22:58,860 --> 00:23:00,693
ujumbe kuchukua nyumbani
kwa hili ni kwamba tuko

403
00:23:00,693 --> 00:23:04,270
kujaribu kupata kompyuta
kwamba wanaweza kufanya kazi kwa ufanisi.

404
00:23:04,270 --> 00:23:09,860
>> Na hivyo, hebu tuangalie juu hii
moja, cutely kuitwa, KFC.

405
00:23:09,860 --> 00:23:12,960
Na kidogo ya matangazo hapa.

406
00:23:12,960 --> 00:23:15,730
Hii hasa kampuni ya chakula
ina chochote cha kufanya na hili.

407
00:23:15,730 --> 00:23:18,240
Ni ukweli kwamba
mfumo huu hasa

408
00:23:18,240 --> 00:23:23,830
ni kulowekwa katika sana
wajanja mafuta makao kiwanja.

409
00:23:23,830 --> 00:23:27,590
Na hivyo got yao
Fryer kuku moniker

410
00:23:27,590 --> 00:23:30,040
wakati wao wa kwanza kuanza
kujenga aina hii ya mifumo.

411
00:23:30,040 --> 00:23:32,740
>> Lakini kimsingi nini wameweza kuchukuliwa
hapa ni idadi ya vile,

412
00:23:32,740 --> 00:23:37,560
kuziweka katika hii
mafuta ya madini ya kisasa,

413
00:23:37,560 --> 00:23:40,979
na kisha kazi nje ya jinsi ya kupata yote
mitandao ndani na nje ya yake.

414
00:23:40,979 --> 00:23:42,895
Kisha, si tu kwamba,
wameweza kuiweka nje ili

415
00:23:42,895 --> 00:23:46,095
kwamba inaweza kunyonya nje ya hewa baridi.

416
00:23:46,095 --> 00:23:47,520
Ilikuwa pretty kuvutia.

417
00:23:47,520 --> 00:23:49,630
Hivyo una kufanya yote
ya shenanigans hii

418
00:23:49,630 --> 00:23:53,280
kuwa na uwezo wa kupata kiasi hiki cha
compute mikononi kwa wattage ndogo.

419
00:23:53,280 --> 00:23:57,360
>> Na unaweza kuona hii ni sura
ya ambapo mambo ni viongozi.

420
00:23:57,360 --> 00:24:01,240
Changamoto ni kwamba hewa ya mara kwa mara
baridi ni uchumi wa wadogo

421
00:24:01,240 --> 00:24:08,459
na ni kuendesha gari mengi ya
maendeleo ya kompyuta wote mara kwa mara,

422
00:24:08,459 --> 00:24:09,750
na juu ya utendaji kompyuta.

423
00:24:09,750 --> 00:24:11,080
Kwa hiyo, hili ni pretty usumbufu.

424
00:24:11,080 --> 00:24:13,280
Nadhani hii ni ya kuvutia.

425
00:24:13,280 --> 00:24:15,530
Ni kidogo messy wakati
kujaribu wabadilishane anatoa disk,

426
00:24:15,530 --> 00:24:18,090
lakini ni wazo kweli cool.

427
00:24:18,090 --> 00:24:22,200
>> Hivyo si tu kwamba, kuna
rundo zima la kazi

428
00:24:22,200 --> 00:24:25,450
inajengwa kuzunguka kile sisi ni
wito Project Open Compute.

429
00:24:25,450 --> 00:24:29,400
Na hivyo, zaidi juu ya kwamba
kidogo baadaye.

430
00:24:29,400 --> 00:24:32,740
Lakini sekta ya kuanza kwa
kutambua kwamba flops kwa watt

431
00:24:32,740 --> 00:24:33,670
ni kuwa muhimu.

432
00:24:33,670 --> 00:24:39,256
Na wewe, kama folks hapa, kama
kubuni algorithms yako

433
00:24:39,256 --> 00:24:41,130
na kubuni yako
kanuni, unapaswa kuwa na ufahamu

434
00:24:41,130 --> 00:24:43,620
kwamba kanuni yako unaweza
kuwa yangekuwa na madhara.

435
00:24:43,620 --> 00:24:48,380
>> Wakati Mark alikuwa amekaa hapa katika yake
Dorm chumba kuandika Facebook 1.0,

436
00:24:48,380 --> 00:24:51,050
Mimi nina uhakika pretty yeye alikuwa na mtazamo
kuwa ni kwenda kuwa kubwa.

437
00:24:51,050 --> 00:24:54,945
Lakini jinsi kubwa itakuwa ni juu ya
mazingira ni dealio kubwa.

438
00:24:54,945 --> 00:24:58,340
Na hivyo wote wa ya'll inaweza
kuja na algorithms

439
00:24:58,340 --> 00:25:01,370
ambayo inaweza kuwa changamoto ya
kitu kwa folks kama mimi,

440
00:25:01,370 --> 00:25:02,700
kujaribu kukimbia mifumo.

441
00:25:02,700 --> 00:25:07,360
>> Basi hebu tu kufikiri kuhusu
ulimwengu wa kweli nguvu mipaka.

442
00:25:07,360 --> 00:25:09,930
Hii karatasi na Landauer--
si jambo jipya.

443
00:25:09,930 --> 00:25:12,480
1961 hii ilichapishwa
katika IBM Journal.

444
00:25:12,480 --> 00:25:15,590
Hii ni kisheria
"Irreversibility na joto

445
00:25:15,590 --> 00:25:17,630
Generation katika Mchakato Computing. "

446
00:25:17,630 --> 00:25:22,050
Na hivyo alisema kuwa
mashine inevitably

447
00:25:22,050 --> 00:25:25,070
kutekeleza majukumu vifaa kwamba
hawana single-yenye thamani ya inverse.

448
00:25:25,070 --> 00:25:29,130
>> Hivyo kwamba sehemu nzima ya
hii ni kwamba nyuma katika '60s,

449
00:25:29,130 --> 00:25:31,890
folks alijua kwamba hii ilikuwa
kwenda kuwa tatizo.

450
00:25:31,890 --> 00:25:37,080
Na hivyo sheria ya mipaka alisema 25
digrii C, aina ya kisheria chumba

451
00:25:37,080 --> 00:25:41,120
joto, kikomo
inawakilisha 0.1 elektroni volts.

452
00:25:41,120 --> 00:25:44,920
Lakini kinadharia, hii ni
kumbukumbu nadharia, kompyuta,

453
00:25:44,920 --> 00:25:51,410
kazi katika kikomo hii inaweza kuwa
iliyopita katika bits bilioni moja ya pili.

454
00:25:51,410 --> 00:25:54,620
>> Sijui kuhusu wewe, lakini si
kuja hela nyingi bilioni moja bits

455
00:25:54,620 --> 00:25:57,190
pili kiwango cha data kubadilishana.

456
00:25:57,190 --> 00:26:01,360
hoja huko ni kwamba tu
2.8 trilioni watt ya nguvu

457
00:26:01,360 --> 00:26:03,180
tunapaswa milele kupanua.

458
00:26:03,180 --> 00:26:08,160
>> Haki, ulimwengu halisi wote example--
hii ni muswada wangu umeme.

459
00:26:08,160 --> 00:26:10,347
Mimi nina asilimia 65% ya
kituo cha data kwamba lovely

460
00:26:10,347 --> 00:26:11,930
Mimi ilionyesha wewe, katika wakati fulani.

461
00:26:11,930 --> 00:26:15,520
Hii ni nyuma katika Juni mwaka jana.

462
00:26:15,520 --> 00:26:21,300
Nimekuwa kuchukuliwa toleo la wazee ili tuweze
unaweza na aina ya kuficha jina kidogo.

463
00:26:21,300 --> 00:26:25,470
Mimi nilikuwa kutumia $ 45,000
mwezi kwa ajili ya nishati huko.

464
00:26:25,470 --> 00:26:34,990
>> Hivyo sababu kuwa kuna ni kwamba sisi
kuwa michakato ya zaidi ya 50,000 katika chumba hicho.

465
00:26:34,990 --> 00:26:38,110
Hivyo unaweza kufikiria yako mwenyewe
muswada makazi umeme

466
00:26:38,110 --> 00:26:39,540
kuwa kwamba juu?

467
00:26:39,540 --> 00:26:46,180
Lakini ilikuwa ni kwa ajili 199000000
watt masaa zaidi ya mwezi.

468
00:26:46,180 --> 00:26:51,670
>> Hivyo swali mimi pose ni, unaweza
kufikiria Mheshimiwa Zuckerberg ya umeme muswada huo?

469
00:26:51,670 --> 00:26:54,730
Mine ni pretty kubwa, na mimi mapambano.

470
00:26:54,730 --> 00:26:56,600
Na mimi nina peke yake katika hili ni.

471
00:26:56,600 --> 00:26:59,450
Kuna mengi ya watu
na kubwa vituo data.

472
00:26:59,450 --> 00:27:04,800
Na hivyo, mimi nadhani, disclosure-- full
yangu Facebook rafiki kidogo isiyo ya kawaida.

473
00:27:04,800 --> 00:27:07,900
>> Hivyo yangu Facebook rafiki ni
Kituo cha Prineville data,

474
00:27:07,900 --> 00:27:14,030
ambayo ni moja ya Facebook kubwa,
newest, chini ya kituo cha data nishati.

475
00:27:14,030 --> 00:27:19,360
Nao baada ya kwangu, mambo kama
nguvu matumizi ufanisi,

476
00:27:19,360 --> 00:27:24,020
kama katika jinsi ufanisi ni data
kituo cha dhidi ya kiasi gani nishati uko

477
00:27:24,020 --> 00:27:26,370
kuweka ndani yake, ni kiasi gani
maji ni wao kutumia, nini

478
00:27:26,370 --> 00:27:27,810
unyevu na joto.

479
00:27:27,810 --> 00:27:29,980
>> Na wao wana haya
lovely, lovely viwanja.

480
00:27:29,980 --> 00:27:32,600
Nadhani hii ni
kushangaza ukurasa wa Facebook,

481
00:27:32,600 --> 00:27:35,400
lakini mimi nadhani nina kidogo weird.

482
00:27:35,400 --> 00:27:39,930
>> Hivyo moja nguvu zaidi kitu,
utafiti kompyuta kwamba mimi kufanya

483
00:27:39,930 --> 00:27:44,060
ni tofauti kwa
nini Facebook na Yahoo na Google

484
00:27:44,060 --> 00:27:50,020
na wengine juu ya mahitaji, kikamilifu,
huduma daima inapatikana.

485
00:27:50,020 --> 00:27:53,530
Na hivyo mimi kuwa na faida kwamba wakati
ISO New England-- na ISO New England

486
00:27:53,530 --> 00:27:58,910
husaidia kuweka nishati
viwango kwa ajili ya kanda.

487
00:27:58,910 --> 00:28:01,110
>> Na inasema ni kupanua
ombi kwa watumiaji

488
00:28:01,110 --> 00:28:05,870
kwa hiari kuhifadhi nishati,
kwa sababu ya joto ya juu na unyevunyevu.

489
00:28:05,870 --> 00:28:08,680
Na hii ilikuwa nyuma juu ya 18 ya mwezi Julai.

490
00:28:08,680 --> 00:28:12,600
Na hivyo mimi furaha Tweet nyuma, Hey,
ISO New England, Green Harvard.

491
00:28:12,600 --> 00:28:14,880
Sisi ni kufanya sehemu yetu juu ya
hapa katika kompyuta utafiti.

492
00:28:14,880 --> 00:28:16,760
Na hii ni kwa sababu sisi ni kufanya sayansi.

493
00:28:16,760 --> 00:28:20,380
>> Na kama vile watu wanasema sayansi
kamwe kulala, sayansi wanaweza kusubiri.

494
00:28:20,380 --> 00:28:25,030
Hivyo sisi ni uwezo wa quiesce wetu
mifumo, kuchukua faida ya viwango vya daraja

495
00:28:25,030 --> 00:28:30,550
nishati yetu ya muswada huo, na
kusaidia nzima New England

496
00:28:30,550 --> 00:28:35,910
kanda kwa kumwaga
megawati nyingi za mzigo.

497
00:28:35,910 --> 00:28:40,020
Hivyo kwamba ni jambo kipekee kwamba
hutofautiana kuhusu kompyuta data kisayansi

498
00:28:40,020 --> 00:28:48,890
vituo na wale walio
katika uzalishaji kamili 24/7.

499
00:28:48,890 --> 00:28:51,670
>> Basi hebu tu kuchukua gia nyingine hapa.

500
00:28:51,670 --> 00:28:55,170
Hivyo, nataka kujadili
machafuko kidogo.

501
00:28:55,170 --> 00:28:59,900
Na mimi nataka kuiweka katika
mwamvuli wa hifadhi.

502
00:28:59,900 --> 00:29:03,150
Hivyo kwa wale aina hiyo
ya walikuwa wanajitahidi

503
00:29:03,150 --> 00:29:08,680
kupata kichwa yao kuzunguka kile petabytes
ya kuhifadhi kuangalia kama, hii mfano.

504
00:29:08,680 --> 00:29:11,660
Na hii ni aina ya mambo
Mimi kukabiliana na wakati wote.

505
00:29:11,660 --> 00:29:15,550
>> Kila mmoja wa haya fellas kidogo
ni nne terabyte ngumu kuendesha,

506
00:29:15,550 --> 00:29:17,420
hivyo unaweza aina ya kuhesabu yao juu.

507
00:29:17,420 --> 00:29:21,370
Tumekuwa tukipata sasa kati ya
moja kwa 1 na 1/2 petabytes

508
00:29:21,370 --> 00:29:22,970
katika kiwango rack sekta hiyo.

509
00:29:22,970 --> 00:29:26,430
Na tuna vyumba na vyumba, kama wewe
aliona katika picha hiyo mapema na John

510
00:29:26,430 --> 00:29:31,230
na mimi, kamili ya racks haya ya vifaa.

511
00:29:31,230 --> 00:29:40,400
Hivyo ni kuwa sana, ni rahisi sana
kujenga arrays mkubwa kuhifadhi

512
00:29:40,400 --> 00:29:44,140
>> Ni zaidi rahisi ndani ya Unix kwa
aina ya kuhesabu hadi jinsi mambo ni kwenda.

513
00:29:44,140 --> 00:29:48,270
Hivyo hii ni kuhesabu ngapi
Mau pointi kuwa mimi got huko.

514
00:29:48,270 --> 00:29:50,880
Hivyo hiyo ni 423 intercept pointi.

515
00:29:50,880 --> 00:29:55,660
Na kisha kama mimi kukimbia baadhi awk sketchy, mimi
Unaweza kuongeza hadi, katika mfumo huu hasa,

516
00:29:55,660 --> 00:29:59,080
kulikuwa 7.3 petabytes
ya kuhifadhi inapatikana.

517
00:29:59,080 --> 00:30:01,350
>> Hivyo hiyo ni mengi ya mambo.

518
00:30:01,350 --> 00:30:03,030
Na hifadhi ni ngumu kweli kweli.

519
00:30:03,030 --> 00:30:06,850
Na hata hivyo, kwa sababu fulani,
hii ni sekta ya mwenendo.

520
00:30:06,850 --> 00:30:11,500
Wakati mimi kuzungumza na watafiti wetu
na kitivo yetu na kusema,

521
00:30:11,500 --> 00:30:14,180
hey, siwezi kukimbia kuhifadhi kwa ajili yenu.

522
00:30:14,180 --> 00:30:17,690
Kwa bahati mbaya, nina
kuokoa gharama za kuhifadhi.

523
00:30:17,690 --> 00:30:19,430
Mimi kupata biashara hii.

524
00:30:19,430 --> 00:30:23,300
Na watu rejea Newegg
au wao rejea Staples

525
00:30:23,300 --> 00:30:27,040
au ni kiasi gani wanaweza kununua
single terabyte disk gari kwa ajili ya.

526
00:30:27,040 --> 00:30:29,390
>> Hivyo hii, utasikia kumbuka
hapa, kwamba kuna fununu.

527
00:30:29,390 --> 00:30:31,310
Kuna moja disk drive hapa.

528
00:30:31,310 --> 00:30:33,290
Na kama sisi kwenda nyuma, nina mengi.

529
00:30:33,290 --> 00:30:36,130
Si tu kufanya mimi wengi, nina
interconnects kisasa

530
00:30:36,130 --> 00:30:38,750
kuwa na uwezo wa kushona
mambo haya kwa pamoja.

531
00:30:38,750 --> 00:30:44,080
Hivyo hatari zinazohusiana na hizi kubwa
arrays kuhifadhi si insignificant.

532
00:30:44,080 --> 00:30:46,370
>> Kwa kweli, sisi alichukua
biashara na sisi aliandika

533
00:30:46,370 --> 00:30:51,670
hadithi kidogo juu vizuri maana,
kali-mannered mkurugenzi wa utafiti

534
00:30:51,670 --> 00:30:54,640
computing-- hutokea kwa kuwa
ajabu Kiingereza accent-- kujaribu

535
00:30:54,640 --> 00:30:59,930
kuelezea kwa mtafiti nini
hakuna folder underscore Backup kweli

536
00:30:59,930 --> 00:31:01,070
maana.

537
00:31:01,070 --> 00:31:05,690
Ilikuwa muda mrefu kabisa, hadithi kidogo,
nzuri ya dakika nne ya ugunduzi.

538
00:31:05,690 --> 00:31:09,380
>> Na kumbuka, nina kubwa
mengi chini nafasi zaidi mwanamke

539
00:31:09,380 --> 00:31:11,800
kwamba kuimba juu ya yote bass.

540
00:31:11,800 --> 00:31:13,910
Tuko akaunti chache kabisa chini.

541
00:31:13,910 --> 00:31:16,160
Lakini anyway, hii ni
Jambo muhimu kufikiria,

542
00:31:16,160 --> 00:31:18,532
katika suala la nini inaweza kwenda vibaya.

543
00:31:18,532 --> 00:31:20,990
Hivyo kama mimi kupata disk kuendesha gari, na
Mimi kutupa katika mashine Unix,

544
00:31:20,990 --> 00:31:24,300
na mimi kuanza kuandika mambo yake,
kuna sumaku, kuna gari kichwa,

545
00:31:24,300 --> 00:31:30,150
kuna zenye, moja au sifuri
kuwa kuandikwa juu ya kifaa hicho.

546
00:31:30,150 --> 00:31:32,180
>> Motors-- spinny, twirly
mambo daima kuvunja.

547
00:31:32,180 --> 00:31:33,490
Fikiria juu ya mambo ambayo kuvunja.

548
00:31:33,490 --> 00:31:35,170
Ni daima imekuwa spinny, twirly mambo.

549
00:31:35,170 --> 00:31:38,560
Printers, anatoa disk,
magari, nk

550
00:31:38,560 --> 00:31:40,590
Kitu kinachotembea ni uwezekano wa kuvunja.

551
00:31:40,590 --> 00:31:42,575
>> Hivyo haja motors, wewe
haja firmware kuendesha gari,

552
00:31:42,575 --> 00:31:47,110
unahitaji SAS / SATA controllers, waya,
firmware juu ya controllers SAS / SATA,

553
00:31:47,110 --> 00:31:48,530
kiwango cha chini vitalu.

554
00:31:48,530 --> 00:31:54,580
Pick hifadhi yako mtawala faili
mfumo wa kanuni, kwa namna yoyote moja inaweza kuwa,

555
00:31:54,580 --> 00:31:56,780
jinsi kushona mambo pamoja.

556
00:31:56,780 --> 00:32:00,956
Na virtual kumbukumbu meneja wako
kurasa, DRAM kuchota na maduka.

557
00:32:00,956 --> 00:32:02,705
Basi, wewe kupata mwingine
stack, ambayo ni aina

558
00:32:02,705 --> 00:32:05,440
ya chini orodha hii
moja, algorithms, watumiaji.

559
00:32:05,440 --> 00:32:09,050
>> Na kama wewe kuzidisha hii
up, sijui ngapi,

560
00:32:09,050 --> 00:32:11,640
kuna mengi ya maeneo
ambapo mambo yanaweza kwenda sideways.

561
00:32:11,640 --> 00:32:14,430
I mean, hiyo ni mfano kuhusu math.

562
00:32:14,430 --> 00:32:18,070
Lakini ni aina ya furaha kufikiria
njia ngapi mambo inaweza kwenda vibaya,

563
00:32:18,070 --> 00:32:21,650
tu kwa ajili ya disk drive.

564
00:32:21,650 --> 00:32:25,440
Tuko tayari katika 300 petabytes, hivyo
kufikiria idadi ya anatoa disk

565
00:32:25,440 --> 00:32:27,741
unahitaji katika 300 petabytes
ambayo yanaweza kwenda vibaya.

566
00:32:27,741 --> 00:32:28,240

567
00:32:28,240 --> 00:32:30,390
Si tu that-- hivyo hiyo ni kuhifadhi.

568
00:32:30,390 --> 00:32:34,220
Na kwamba inataja
mtu Ningependa kuona

569
00:32:34,220 --> 00:32:38,780
kuingia hatua kushoto, ambayo
ni machafuko Monkey.

570
00:32:38,780 --> 00:32:43,140
Hivyo katika hatua fulani, anapata hata
kubwa kuliko tu disk drive

571
00:32:43,140 --> 00:32:43,920
tatizo.

572
00:32:43,920 --> 00:32:50,610
>> Na hivyo, haya wanawake faini na muungwana
kwamba kukimbia Streaming video huduma

573
00:32:50,610 --> 00:32:55,430
waligundua kwamba kompyuta zao walikuwa
pia kubwa na pia ngumu sana

574
00:32:55,430 --> 00:33:00,010
na pia kutoa huduma
kwa kutisha mengi ya watu.

575
00:33:00,010 --> 00:33:05,180
Wao nimepata milioni 37 members-- na
slide huu labda mwaka mmoja au old--

576
00:33:05,180 --> 00:33:07,350
maelfu ya vifaa.

577
00:33:07,350 --> 00:33:10,810
Kuna mabilioni ya masaa ya video.

578
00:33:10,810 --> 00:33:13,600
Wao kuingia mabilioni ya matukio ya siku.

579
00:33:13,600 --> 00:33:17,330
>> Na unaweza kuona, watu wengi kuangalia
telly baadaye jioni,

580
00:33:17,330 --> 00:33:19,429
na mbali outweighs kila kitu.

581
00:33:19,429 --> 00:33:21,220
Na hivyo, walitaka
kuwa na uwezo wa kuhakikisha

582
00:33:21,220 --> 00:33:24,854
kwamba huduma ilikuwa juu na
kuaminika na kufanya kazi kwa ajili yao.

583
00:33:24,854 --> 00:33:27,020
Basi wakapanda na hii
kitu kinachoitwa machafuko Monkey.

584
00:33:27,020 --> 00:33:29,000
Ni kipande cha programu
ambayo, wakati unafikiri

585
00:33:29,000 --> 00:33:34,190
juu ya kuzungumza kuhusu jina
ya mada hii nzima,

586
00:33:34,190 --> 00:33:36,530
wadogo-nje ina maana
unapaswa mtihani mambo haya.

587
00:33:36,530 --> 00:33:38,585
Siyo nzuri tu kuwa
milioni mashine.

588
00:33:38,585 --> 00:33:40,460
Kwa hiyo, jambo zuri kuhusu
hii ni, machafuko Monkey

589
00:33:40,460 --> 00:33:43,090
ni huduma ambayo
kubainisha makundi ya mifumo ya

590
00:33:43,090 --> 00:33:47,220
na nasibu vipoe moja
mifumo ya katika kundi.

591
00:33:47,220 --> 00:33:48,429
Kutisha.

592
00:33:48,429 --> 00:33:50,220
Hivyo mimi sijui kuhusu
wewe, lakini kama mimi wameweza milele

593
00:33:50,220 --> 00:33:52,990
kujengwa mfumo kwamba hutegemea nyingine
mifumo kuzungumza na kila mmoja,

594
00:33:52,990 --> 00:33:55,865
wewe kuchukua mmoja wao nje,
uwezekano wa jambo nzima kazi,

595
00:33:55,865 --> 00:33:57,130
utapungua kwa kasi.

596
00:33:57,130 --> 00:34:00,475
>> Na hivyo kipande hii ya programu anaendesha
karibu miundombinu Netflix.

597
00:34:00,475 --> 00:34:03,100
Kwa bahati nzuri, anasema ni anaendesha tu katika
masaa ya biashara kwa nia ya

598
00:34:03,100 --> 00:34:05,810
kwamba wahandisi itakuwa
macho na uwezo wa kujibu.

599
00:34:05,810 --> 00:34:08,020
Basi hizi ni aina
ya mambo ya sisi ni sasa

600
00:34:08,020 --> 00:34:13,360
baada ya kufanya kwa perturb kompyuta yetu
mazingira, kuanzisha machafuko

601
00:34:13,360 --> 00:34:15,739
na kuanzisha utata.

602
00:34:15,739 --> 00:34:19,139
>> Hivyo ni nani, katika haki zao
akili, kwa hiari kuchagua

603
00:34:19,139 --> 00:34:22,540
kufanya kazi na machafuko Monkey?

604
00:34:22,540 --> 00:34:24,150
Hutegemea, yeye inaonekana kuwa akizungumzia yangu.

605
00:34:24,150 --> 00:34:28,719
Naam, mimi nadhani should-- cute.

606
00:34:28,719 --> 00:34:32,909
Lakini tatizo ni wewe
hawapati uchaguzi.

607
00:34:32,909 --> 00:34:37,440
Machafuko Monkey, kama wewe
unaweza kuona, akiamua wewe.

608
00:34:37,440 --> 00:34:42,650
>> Na hili ni tatizo kwa kompyuta
katika kiwango ni kwamba huwezi kuepuka hii.

609
00:34:42,650 --> 00:34:49,989
Ni inevitability ya utata
na ya wadogo na wa mageuzi yetu,

610
00:34:49,989 --> 00:34:53,280
katika baadhi ya njia, ya kompyuta utaalamu.

611
00:34:53,280 --> 00:34:55,510
Na kumbuka, hii ni
jambo moja kukumbuka,

612
00:34:55,510 --> 00:35:00,030
Machafuko Monkeys upendo
snowflakes-- upendo snowflakes.

613
00:35:00,030 --> 00:35:03,470
snowflake-- tumekuwa alielezea
Machafuko Monkey-- lakini snowflake

614
00:35:03,470 --> 00:35:09,630
ni server kuwa ni ya kipekee na
maalum na maridadi na mtu binafsi

615
00:35:09,630 --> 00:35:11,770
na kamwe kuwa tena.

616
00:35:11,770 --> 00:35:14,790
>> Mara kwa mara kupata snowflake
huduma katika mazingira yetu.

617
00:35:14,790 --> 00:35:16,700
Na sisi daima kujaribu na
kuyeyuka snowflake huduma.

618
00:35:16,700 --> 00:35:18,880
Lakini kama wewe kupata server
katika mazingira yako

619
00:35:18,880 --> 00:35:23,240
ambayo ni muhimu kwa longevity
wa asasi yako, nayo ikayeyuka,

620
00:35:23,240 --> 00:35:25,300
huwezi kuiweka nyuma pamoja tena.

621
00:35:25,300 --> 00:35:28,071
Hivyo kazi machafuko Monkey ilikuwa
kwenda na kusitisha matukio.

622
00:35:28,071 --> 00:35:30,820
Kama machafuko Monkey melts
snowflake, wewe juu, wewe ni kosa.

623
00:35:30,820 --> 00:35:34,390

624
00:35:34,390 --> 00:35:37,950
Nataka kuzungumza kuhusu
baadhi ya vifaa kwamba sisi ni

625
00:35:37,950 --> 00:35:40,415
kuona katika suala la aina ya
wadogo-shughuli pia.

626
00:35:40,415 --> 00:35:43,810
Na baadhi ya mambo ya kipekee kwamba ni katika
na kuzunguka shughuli sayansi.

627
00:35:43,810 --> 00:35:46,990
Sisi sasa ni mapya ya kuona, kumbuka
kitengo hii ya suala hilo, rack hii?

628
00:35:46,990 --> 00:35:51,780
Hivyo hii ni rack ya GPGPUs-- hivyo kwa ujumla
lengo graphics vitengo vya usindikaji.

629
00:35:51,780 --> 00:35:55,790
>> Sisi hizi ziko katika takwimu zetu
kituo hicho, 100 au hivyo maili.

630
00:35:55,790 --> 00:35:59,780
Rack hii hasa
ni kuhusu 96 tera flops

631
00:35:59,780 --> 00:36:04,090
ya single-usahihi math na uwezo
kutoa nje nyuma yake.

632
00:36:04,090 --> 00:36:10,530
Na tuna ili 130-isiyo ya kawaida
kadi katika mfano

633
00:36:10,530 --> 00:36:16,620
kwamba we-- nyingi
racks ya mfano huu.

634
00:36:16,620 --> 00:36:22,730
>> Hivyo hii ni ya kuvutia kwa maana kwamba
madhumuni ya jumla graphics michakato

635
00:36:22,730 --> 00:36:27,880
ni uwezo wa kufanya hisabati incredibly
haraka kwa kiasi cha chini sana ya nishati.

636
00:36:27,880 --> 00:36:32,060
Hivyo kuna uptick kubwa katika
maeneo kisayansi kompyuta,

637
00:36:32,060 --> 00:36:36,400
kuangalia graphics
vitengo vya usindikaji katika njia kubwa.

638
00:36:36,400 --> 00:36:41,990
>> Hivyo mimi mbio baadhi Mcollective
kupitia bandia miundombinu yetu

639
00:36:41,990 --> 00:36:45,330
jana, msisimko sana kuhusu hili.

640
00:36:45,330 --> 00:36:48,260
muda mfupi tu wa petaflop
ya single-usahihi.

641
00:36:48,260 --> 00:36:52,440
Tu kuwa wazi hapa, hii
multiplier kidogo ni 3.95.

642
00:36:52,440 --> 00:36:54,820
Double-usahihi math
itakuwa juu ya 1.2,

643
00:36:54,820 --> 00:36:57,010
lakini yangu Twitter kulisha
inaonekana njia bora kama mimi

644
00:36:57,010 --> 00:37:02,670
alisema tulikuwa karibu petaflop
ya single-usahihi GPGPUs.

645
00:37:02,670 --> 00:37:04,220
>> Lakini ni kupata huko.

646
00:37:04,220 --> 00:37:06,280
Ni kupata kuwa
sana, kuvutia sana.

647
00:37:06,280 --> 00:37:08,550
Na kwa nini tunafanya hivi?

648
00:37:08,550 --> 00:37:11,570
Kwa sababu kemia quantum,
miongoni mwa mambo mengine,

649
00:37:11,570 --> 00:37:15,300
lakini sisi ni kuanzia kubuni
baadhi photovoltaics mpya.

650
00:37:15,300 --> 00:37:20,210
>> Na hivyo Alan Aspuru-Guzik, ambaye ni
profesa katika chemistry-- mpenzi wangu

651
00:37:20,210 --> 00:37:22,390
katika crime-- kwa miaka michache iliyopita.

652
00:37:22,390 --> 00:37:25,660
Tumekuwa kusukuma
bahasha ya kompyuta.

653
00:37:25,660 --> 00:37:30,250
Na GPGPU ni bora
teknolojia ya kuwa na uwezo wa kufanya

654
00:37:30,250 --> 00:37:34,760
mengi ya kutisha ya ngumu
math, kwa haraka sana,.

655
00:37:34,760 --> 00:37:36,750
>> Hivyo, pamoja na wadogo, huja na changamoto mpya.

656
00:37:36,750 --> 00:37:41,070
Hivyo scale-- mkubwa wewe kuwa
makini jinsi waya mambo haya.

657
00:37:41,070 --> 00:37:45,300
Na tuna ngazi fulani ya
obsessive compulsive disorder.

658
00:37:45,300 --> 00:37:49,530
Picha hizi pengine
gari mengi ya watu karanga.

659
00:37:49,530 --> 00:37:53,390
Na makabati ambayo si
wired vizuri hasa

660
00:37:53,390 --> 00:37:56,050
gari mtandao wetu na
vifaa wahandisi karanga.

661
00:37:56,050 --> 00:37:58,620
Plus pia kuna airflow
masuala ambayo una vyenye.

662
00:37:58,620 --> 00:38:01,430
>> Hivyo haya ni mambo ambayo mimi
kamwe kuwa na mawazo ya.

663
00:38:01,430 --> 00:38:03,480
Na wadogo, huja zaidi utata.

664
00:38:03,480 --> 00:38:05,869
Hii ni aina mpya ya mfumo wa faili.

665
00:38:05,869 --> 00:38:06,410
Ni kutisha.

666
00:38:06,410 --> 00:38:07,660
Ni petabyte.

667
00:38:07,660 --> 00:38:09,905
Ni inaweza kuhifadhi files bilioni 1.1.

668
00:38:09,905 --> 00:38:15,940
Ni unaweza kusoma na kuandika kwa 13 gigabytes
na 20 gigabytes gigabytes second--

669
00:38:15,940 --> 00:38:17,150
pili.

670
00:38:17,150 --> 00:38:20,900
Hivyo inaweza kupakua terabytes
katika wakati hakuna wakati wote.

671
00:38:20,900 --> 00:38:22,070
>> Na ni yenye inapatikana.

672
00:38:22,070 --> 00:38:26,989
Na ni got lookup kushangaza
rates-- 220,000 lookups pili.

673
00:38:26,989 --> 00:38:29,780
Na kuna watu wengi tofauti tofauti
kujenga aina hii ya mifumo.

674
00:38:29,780 --> 00:38:32,830
Na unaweza kuona hapa graphically.

675
00:38:32,830 --> 00:38:35,800
Hii ni moja ya mifumo ya faili wetu
hiyo ni chini ya mzigo, kabisa

676
00:38:35,800 --> 00:38:41,250
furaha kusoma katika muda mfupi tu
ya 22 gigabytes ya pili.

677
00:38:41,250 --> 00:38:42,790
Hivyo hiyo ni cool-- hivyo utata.

678
00:38:42,790 --> 00:38:47,230
>> Hivyo, pamoja na utata na wadogo,
huja zaidi utata, haki?

679
00:38:47,230 --> 00:38:51,830
Hii ni moja ya wengi wetu,
michoro wengi mtandao,

680
00:38:51,830 --> 00:38:54,970
ambapo una mengi tofauti
chassier wote kusaidia up

681
00:38:54,970 --> 00:38:57,730
katika msingi kuu kubadili,
kushikamana na kuhifadhi,

682
00:38:57,730 --> 00:39:00,731
kuunganisha na chini interconnects latency.

683
00:39:00,731 --> 00:39:03,605
Na kisha wote wa upande huu wa
nyumba, ni tu yote ya usimamizi wa

684
00:39:03,605 --> 00:39:09,740
kwamba unahitaji kuwa na uwezo wa kushughulikia
mifumo kutoka eneo kijijini.

685
00:39:09,740 --> 00:39:12,070
Hivyo wadogo ina mengi ya
utata na hayo.

686
00:39:12,070 --> 00:39:14,910

687
00:39:14,910 --> 00:39:17,785
>> Kubadilisha gia tena, hebu kwenda nyuma
na kuwa na doa kidogo ya sayansi.

688
00:39:17,785 --> 00:39:21,450
Hivyo, kumbuka, utafiti
kompyuta na hii shim-- kidogo

689
00:39:21,450 --> 00:39:25,310
kidogo pink shim kati ya
Kitivo na yote ya algorithms yao

690
00:39:25,310 --> 00:39:30,650
na wote wa sayansi baridi na wote wa
nguvu na baridi na data kituo hiki

691
00:39:30,650 --> 00:39:35,330
sakafu na mitandao na kompyuta kubwa
na madawati huduma na madawati msaada

692
00:39:35,330 --> 00:39:39,330
na hivyo sana aliye mzaa na hivyo, tuko tu
hii shim kidogo kati yao.

693
00:39:39,330 --> 00:39:42,820
>> Nini tumekuwa walianza
kuona ni kuwa dunia ya

694
00:39:42,820 --> 00:39:45,730
na uwezo wa kujenga
kubwa vituo hivi vya data

695
00:39:45,730 --> 00:39:48,020
na kuwa na uwezo wa kujenga
kubwa kompyuta hizi.

696
00:39:48,020 --> 00:39:49,420
Tumekuwa wamezipata nzuri saa yake.

697
00:39:49,420 --> 00:39:53,600
Nini sisi ni nzuri sana katika si ni hii
shim kidogo kati ya utafiti

698
00:39:53,600 --> 00:39:56,670
na chuma wazi na teknolojia.

699
00:39:56,670 --> 00:39:58,600
Na ni ngumu.

700
00:39:58,600 --> 00:40:03,330
>> Na hivyo tumekuwa na uwezo wa kuajiri
folks kwamba kuishi katika dunia hii.

701
00:40:03,330 --> 00:40:07,590
Na hivi karibuni zaidi, sisi alizungumza na
Sayansi ya Taifa Foundation, akasema,

702
00:40:07,590 --> 00:40:11,440
mambo haya wadogo-nje ni kubwa,
lakini hatuwezi kupata wanasayansi wetu

703
00:40:11,440 --> 00:40:13,690
juu ya mashine hizo kubwa ngumu.

704
00:40:13,690 --> 00:40:16,040
Na hivyo, kumekuwa na
idadi ya programu mbalimbali

705
00:40:16,040 --> 00:40:20,100
ambapo sisi kweli walikuwa wengi
wasiwasi kuhusu kujaribu

706
00:40:20,100 --> 00:40:22,800
kuona kama tunaweza kubadilisha
miundombinu ya chuo.

707
00:40:22,800 --> 00:40:25,850
>> Kuna mengi ya programu
karibu vituo vya kitaifa.

708
00:40:25,850 --> 00:40:28,300
Na hivyo, sisi wenyewe, yetu
marafiki katika Clemson,

709
00:40:28,300 --> 00:40:32,620
Chuo Kikuu cha Wisconsin Madison,
Southern California, Utah, na Hawaii

710
00:40:32,620 --> 00:40:35,780
aina ya got pamoja na
kuangalia tatizo hili.

711
00:40:35,780 --> 00:40:39,340
Na hii graph kidogo hapa
ni mkia mrefu wa sayansi.

712
00:40:39,340 --> 00:40:41,602
>> Hivyo hii is-- hana
jambo nini juu ya mhimili huu,

713
00:40:41,602 --> 00:40:45,485
lakini mhimili hii ni kweli simu
ya ajira kwenda kwa nguzo.

714
00:40:45,485 --> 00:40:48,940
Hivyo kuna 350,000 juu ya
muda wowote katika kipindi hicho.

715
00:40:48,940 --> 00:40:51,730
Hizi ni watuhumiwa yetu ya kawaida
pamoja chini hapa.

716
00:40:51,730 --> 00:40:55,992
Kwa kweli, kuna Alan Aspuru-Guzik,
ambao tulikuwa tu kuzungumza about-- tani

717
00:40:55,992 --> 00:40:58,700
na tani ya compute, kwa kweli
ufanisi, anajua anachokifanya.

718
00:40:58,700 --> 00:41:02,840
>> Hapa ni maabara mwingine kwamba mimi itabidi kuzungumza
kuhusu katika moment-- maabara John Kovac ya.

719
00:41:02,840 --> 00:41:03,610
Wao nimepata yake.

720
00:41:03,610 --> 00:41:04,210
Wao ni nzuri.

721
00:41:04,210 --> 00:41:04,830
Wao ni furaha.

722
00:41:04,830 --> 00:41:05,960
Wao ni kompyuta.

723
00:41:05,960 --> 00:41:07,664
Mkuu sayansi ni kupata kufanyika.

724
00:41:07,664 --> 00:41:09,580
Na kisha, kama wewe aina
ya kuja chini hapa, kuna

725
00:41:09,580 --> 00:41:12,110
makundi mengine kwamba
si mbio ajira nyingi zaidi.

726
00:41:12,110 --> 00:41:13,410
>> Na kwa nini ni kwamba?

727
00:41:13,410 --> 00:41:15,080
Je, ni kwa sababu kompyuta ni ngumu sana?

728
00:41:15,080 --> 00:41:19,580
Je, ni kwa sababu hawajui jinsi ya?

729
00:41:19,580 --> 00:41:22,880
Hatujui, kwa sababu
tumeenda na inaonekana.

730
00:41:22,880 --> 00:41:25,620
Na hivyo kwamba ni nini hii
mradi ni wote kuhusu,

731
00:41:25,620 --> 00:41:27,830
ni ndani ya nchi, ndani ya
kila moja ya mikoa hii,

732
00:41:27,830 --> 00:41:32,660
kuangalia kwa fursa ambapo tunaweza kushiriki
na Kitivo na watafiti

733
00:41:32,660 --> 00:41:36,400
kweli katika mwisho chini ya mkia,
na kuelewa nini re kufanya.

734
00:41:36,400 --> 00:41:37,920
>> Hivyo hiyo ni kitu ambacho tuko
kweli shauku.

735
00:41:37,920 --> 00:41:39,920
Na kwamba ni jambo ambalo
sayansi si kuendelea

736
00:41:39,920 --> 00:41:44,260
kusonga mbele mpaka sisi kutatua
baadhi ya kesi hizi makali.

737
00:41:44,260 --> 00:41:46,590
Bits mengine ya sayansi
kwamba kinaendelea up-- kila mtu

738
00:41:46,590 --> 00:41:48,260
kuonekana Kubwa Hadron Collider.

739
00:41:48,260 --> 00:41:49,540
Ajabu, haki?

740
00:41:49,540 --> 00:41:52,960
Mambo haya yote mbio nje katika Holyoke.

741
00:41:52,960 --> 00:41:56,510
Sisi built-- sana kwanza
sayansi yaliyotokea katika Holyoke

742
00:41:56,510 --> 00:41:59,130
ilikuwa ushirikiano kati ya
wenyewe na Chuo Kikuu cha Boston.

743
00:41:59,130 --> 00:42:01,510
Hivyo ni kweli, kweli cool.

744
00:42:01,510 --> 00:42:04,410
>> Hii ni furaha kipande
ya sayansi kwa ajili ya wadogo.

745
00:42:04,410 --> 00:42:07,650
Hii ni upatikanaji digital kwa
anga karne katika Harvard.

746
00:42:07,650 --> 00:42:09,170
Kimsingi, ni archive sahani.

747
00:42:09,170 --> 00:42:13,350
Kama kwenda chini Oxford--
Garden Street, sorry,

748
00:42:13,350 --> 00:42:16,560
utapata moja ya uchunguzi
majengo ni kimsingi full

749
00:42:16,560 --> 00:42:19,480
ya karibu nusu milioni sahani.

750
00:42:19,480 --> 00:42:24,410
>> Na hawa ni picha za
anga wakati wa usiku, zaidi ya miaka 100.

751
00:42:24,410 --> 00:42:28,760
Hivyo kuna rig nzima kuanzisha
hapa digitize sahani hizo,

752
00:42:28,760 --> 00:42:32,100
kuchukua picha wao, kujiandikisha
yao, kuziweka kwenye kompyuta.

753
00:42:32,100 --> 00:42:36,410
Na kwamba ni petabyte na nusu,
haki tu there-- mradi mmoja mdogo.

754
00:42:36,410 --> 00:42:37,530
>> Hizi ni miradi mingine.

755
00:42:37,530 --> 00:42:42,800
Mradi huu Pan-STARRS ni kufanya
full mbalimbali panoramic utafiti,

756
00:42:42,800 --> 00:42:47,390
kuangalia kwa asteroids karibu Dunia
na muda mfupi mbinguni matukio.

757
00:42:47,390 --> 00:42:52,100
Kama biophysicist molecular, I love
neno tukio muda mfupi mbinguni.

758
00:42:52,100 --> 00:42:55,050
Mimi nina uhakika kabisa ni nini, lakini
anyway, sisi ni kuangalia kwa ajili yao.

759
00:42:55,050 --> 00:43:00,372
>> Na sisi ni kuzalisha 30 terabytes
usiku nje ya darubini hizo.

760
00:43:00,372 --> 00:43:03,330
Na kwamba si kweli Bandwidth
tatizo, kwamba kama tatizo FedEx.

761
00:43:03,330 --> 00:43:08,420
Hivyo kuweka kuhifadhi juu ya van
na wewe kutuma chochote ni.

762
00:43:08,420 --> 00:43:10,570
>> Bicep ni kweli interesting--
hivyo background imaging

763
00:43:10,570 --> 00:43:13,850
ya cosmic ziada galactic ubaguzi.

764
00:43:13,850 --> 00:43:16,880
Wakati mimi kwanza kuanza kufanya kazi
katika Harvard saba au hivyo,

765
00:43:16,880 --> 00:43:21,440
miaka nane iliyopita, Nakumbuka
kazi juu ya mradi huu

766
00:43:21,440 --> 00:43:26,010
na haikuwa kweli kuzama
nyumbani kwa nini polarized mwanga

767
00:43:26,010 --> 00:43:29,770
kutoka microwave cosmic
background itakuwa muhimu,

768
00:43:29,770 --> 00:43:30,800
mpaka hii kilichotokea.

769
00:43:30,800 --> 00:43:34,580
>> Na hii ilikuwa John Kovac,
ambaye mimi aliongea na kabla,

770
00:43:34,580 --> 00:43:42,030
kutumia mamilioni juu ya mamilioni ya CPU
masaa, katika kituo wetu na wengine,

771
00:43:42,030 --> 00:43:46,600
kimsingi kuwaangalia ndani
ya ulimwengu wa kwanza wakati

772
00:43:46,600 --> 00:43:49,150
baada ya Big Bang, na
kujaribu kuelewa

773
00:43:49,150 --> 00:43:51,290
Ujumla nadharia Einstein ya relativity.

774
00:43:51,290 --> 00:43:56,040
Ni akili ya kupiga kwamba kompyuta yetu
wanatusaidia unravel na staré

775
00:43:56,040 --> 00:43:59,280
katika mwanzo kabisa wa nini tuko hapa.

776
00:43:59,280 --> 00:44:03,450
>> Hivyo wakati wewe majadiliano juu ya wadogo,
hii ni wadogo baadhi kubwa.

777
00:44:03,450 --> 00:44:09,260
Jambo jingine ya wadogo, kwamba
mradi fulani hit guys haya.

778
00:44:09,260 --> 00:44:15,320
Na hili ndilo jibu Curve kwa bicep
[Inaudible] Hii ilikuwa utafiti wetu kidogo.

779
00:44:15,320 --> 00:44:19,220
>> Na unaweza kuona hapa, maisha
ilikuwa nzuri mpaka juu hapa,

780
00:44:19,220 --> 00:44:21,200
ambayo ilikuwa wakati
Tangazo hilo limekuja nje.

781
00:44:21,200 --> 00:44:24,120
Na una got literally
sekunde kujibu

782
00:44:24,120 --> 00:44:29,020
tukio kuongeza ambayo
sambamba na hii dot kidogo hapa,

783
00:44:29,020 --> 00:44:32,200
ambayo kuishia kuhama
terabytes nne au zaidi ya data

784
00:44:32,200 --> 00:44:36,370
kupitia mtandao wa kompyuta
kwamba day-- pretty nywele.

785
00:44:36,370 --> 00:44:38,210
>> Na hivyo, haya ni
aina ya mambo ambayo

786
00:44:38,210 --> 00:44:43,040
yanaweza kutokea kwa wewe katika miundombinu yako
kama huna kubuni kwa wadogo.

787
00:44:43,040 --> 00:44:45,630
Tulikuwa kidogo ya
kinyang'anyiro siku hiyo, kuwa

788
00:44:45,630 --> 00:44:50,440
uwezo wa span nje ya mtandao kutosha huduma
kuweka tovuti juu na kukimbia.

789
00:44:50,440 --> 00:44:53,399
Na sisi walikuwa na mafanikio.

790
00:44:53,399 --> 00:44:55,190
Hii ni email kidogo
hiyo ni aina ya cute.

791
00:44:55,190 --> 00:45:00,245
Hii ni barua ya Mark Vogelsberger,
na Lars Hernquist, ambaye ni

792
00:45:00,245 --> 00:45:02,650
Kitivo mwanachama hapa katika Harvard.

793
00:45:02,650 --> 00:45:03,570
Zaidi kuhusu Mark baadaye.

794
00:45:03,570 --> 00:45:05,990
Lakini nadhani hii ni moja
aina ya anavyohitimisha aina

795
00:45:05,990 --> 00:45:09,920
ya ambapo kompyuta
ni katika kompyuta utafiti.

796
00:45:09,920 --> 00:45:12,070
Hey, timu, tangu jana
Jumanne, nyie racked up

797
00:45:12,070 --> 00:45:15,470
zaidi ya 28% ya mpya
nguzo, ambayo pamoja

798
00:45:15,470 --> 00:45:20,040
ni zaidi ya miaka 78 ya
CPU katika siku tatu tu.

799
00:45:20,040 --> 00:45:22,502
Na mimi alisema, bado
tu Ijumaa asubuhi.

800
00:45:22,502 --> 00:45:23,460
Hii ni pretty kushangaza!

801
00:45:23,460 --> 00:45:24,740
Furaha Ijumaa!

802
00:45:24,740 --> 00:45:27,450
>> Kisha mimi kuwapa pointi data.

803
00:45:27,450 --> 00:45:30,260
Na ili kwamba ilikuwa aina ya kuvutia.

804
00:45:30,260 --> 00:45:34,840
Basi kumbuka kuhusu Mark, yeye itabidi kuja
nyuma katika picha katika kidogo.

805
00:45:34,840 --> 00:45:36,935
Hivyo wadogo-out kompyuta ni kila mahali.

806
00:45:36,935 --> 00:45:41,080
>> Sisi ni hata kusaidia folks kuangalia
jinsi kazi NBA,

807
00:45:41,080 --> 00:45:43,140
na ambapo watu ni
kutupa mipira kutoka.

808
00:45:43,140 --> 00:45:47,580
Mimi si kweli kuelewa mchezo huu pia
vizuri, lakini inaonekana, ni mpango kubwa.

809
00:45:47,580 --> 00:45:50,610
Kuna hoops na mabakuli, na fedha.

810
00:45:50,610 --> 00:45:55,300
>> Na hivyo, database yetu, sisi
kujengwa kidogo 500 [inaudible]

811
00:45:55,300 --> 00:45:58,170
processor sambamba nguzo,
michache ya terabytes ya RAM,

812
00:45:58,170 --> 00:46:03,590
kuwa na uwezo wa kujenga hii
kwa Kirk na timu yake.

813
00:46:03,590 --> 00:46:08,524
Na wao ni kufanya kompyuta
kwa njia nyingine kwa ujumla.

814
00:46:08,524 --> 00:46:10,440
Sasa huu ni mradi tuko
wanaohusika na kwamba

815
00:46:10,440 --> 00:46:14,880
kuvutia kabisa, karibu neural
kinamu connectomics na genomic

816
00:46:14,880 --> 00:46:20,960
imprinting-- tatu nzito sana
kupiga maeneo ya utafiti

817
00:46:20,960 --> 00:46:24,650
kwamba sisi kupambana na juu ya
siku-kwa-siku.

818
00:46:24,650 --> 00:46:30,670
wazo kwamba akili zetu ni chini ya
stress plastiki wakati sisi ni vijana.

819
00:46:30,670 --> 00:46:34,980
Na mengi ya tabia yetu ya watu wazima ni
sculpted na uzoefu katika uchanga.

820
00:46:34,980 --> 00:46:37,040
Hivyo hii ni dealio kubwa.

821
00:46:37,040 --> 00:46:41,360
>> Na hivyo hii ni kazi hiyo unaofadhiliwa na
Taasisi ya Taifa ya Afya ya Akili.

822
00:46:41,360 --> 00:46:46,860
Na sisi ni kujaribu kimsingi,
njia ya kura ya data kubwa

823
00:46:46,860 --> 00:46:51,970
na kubwa uchambuzi wa data, aina
ya rika katika ubongo wetu wa kibinadamu

824
00:46:51,970 --> 00:46:54,870
njia ya aina ya
mbinu mbalimbali.

825
00:46:54,870 --> 00:47:00,360
>> Kwa hiyo nilitaka kuacha na aina ya
pause tu kwa muda kidogo.

826
00:47:00,360 --> 00:47:04,160
changamoto na kijijini
vituo vya data ni mbali mbali.

827
00:47:04,160 --> 00:47:05,520
Ni hawawezi kufanya kazi.

828
00:47:05,520 --> 00:47:07,590
Nahitaji data yangu karibu.

829
00:47:07,590 --> 00:47:10,730
Mimi haja ya kufanya utafiti wangu katika maabara yangu.

830
00:47:10,730 --> 00:47:18,620
>> Na hivyo mimi aina ya alichukua mfano wa
kazi magnetic resonance imaging

831
00:47:18,620 --> 00:47:22,260
kuweka data kutoka takwimu zetu
katika kituo cha Western Misa.

832
00:47:22,260 --> 00:47:24,660
na kushikamana kwa yangu
desktop katika Cambridge.

833
00:47:24,660 --> 00:47:27,440
Na mimi itabidi kucheza video kidogo.

834
00:47:27,440 --> 00:47:29,750
Nadhani itakuwa aina ya kazi.

835
00:47:29,750 --> 00:47:33,480
>> Hivyo hii ni mimi kwenda kwa
kuangalia GPUs yangu ni kazi.

836
00:47:33,480 --> 00:47:35,430
Na mimi nina kuangalia kwamba VNC ya juu.

837
00:47:35,430 --> 00:47:36,810
Na hii ni VNC wajanja.

838
00:47:36,810 --> 00:47:38,970
Hii ni VNC na vipande 3D.

839
00:47:38,970 --> 00:47:41,975
Na hivyo, kama unaweza kuona muda mfupi, hii
ni mimi inazunguka ubongo huu.

840
00:47:41,975 --> 00:47:44,460
Mimi nina kujaribu aina ya kupata ni oriented.

841
00:47:44,460 --> 00:47:49,574
Na kisha naweza hoja kwa njia ya wengi
vipande mbalimbali ya data MRI.

842
00:47:49,574 --> 00:47:51,490
Na kitu pekee kwamba
mbalimbali kuhusu hili

843
00:47:51,490 --> 00:47:55,160
ni, ni kuja juu ya waya
kutoka Magharibi Misa. desktop yangu.

844
00:47:55,160 --> 00:47:57,300
Na utoaji wake
kasi zaidi kuliko desktop yangu,

845
00:47:57,300 --> 00:48:02,840
kwa sababu sina $ 4,000
graphics kadi katika desktop yangu, ambayo

846
00:48:02,840 --> 00:48:04,262
tuna nje Western Misa.

847
00:48:04,262 --> 00:48:05,720
Bila shaka, mimi nina kujaribu kuwa wajanja.

848
00:48:05,720 --> 00:48:08,859
Mimi mbio gia GLX katika
background, wakati akifanya yote haya,

849
00:48:08,859 --> 00:48:10,900
kuhakikisha kwamba naweza
kusisitiza graphics kadi,

850
00:48:10,900 --> 00:48:14,140
na kwamba kila aina ya
kazi na wengine yote.

851
00:48:14,140 --> 00:48:16,700
Lakini jambo muhimu ni,
ni hii ni 100 maili.

852
00:48:16,700 --> 00:48:20,460
Na unaweza kuona kutokana na hili kwamba
hakuna latency dhahiri.

853
00:48:20,460 --> 00:48:24,600
Mambo kufanya pamoja uungwana vizuri.

854
00:48:24,600 --> 00:48:28,907
>> Na hivyo kwamba, katika yenyewe,
ni mfano na baadhi ya ufahamu

855
00:48:28,907 --> 00:48:31,490
katika jinsi kompyuta na wadogo-nje
kompyuta kinaenda kutokea.

856
00:48:31,490 --> 00:48:35,330
Tuko wote kazi ya
wakondefu na wakondefu vifaa.

857
00:48:35,330 --> 00:48:36,870
Matumizi yetu ya vidonge ni kuongezeka.

858
00:48:36,870 --> 00:48:39,160
>> Hivyo basi, dioksidi yangu
footprint kimsingi ni

859
00:48:39,160 --> 00:48:42,060
kuhama kutoka kile kutumika
ya kufanya hivyo itakuwa wameweza

860
00:48:42,060 --> 00:48:46,060
wamekuwa mashine kubwa
chini ya dawati yangu, kwa nini

861
00:48:46,060 --> 00:48:49,550
sasa ni facility-- inaweza kuwa mahali popote.

862
00:48:49,550 --> 00:48:50,800
Ni inaweza kuwa mahali popote wakati wote.

863
00:48:50,800 --> 00:48:54,790
Lakini hata hivyo, bado uwezo wa kuleta
nyuma juu ya utendaji graphics

864
00:48:54,790 --> 00:48:56,630
desktop yangu.

865
00:48:56,630 --> 00:49:00,900
>> Hivyo, kupata karibu
end-- kumbuka Mark?

866
00:49:00,900 --> 00:49:04,480
Naam, smart kijana ni Mark.

867
00:49:04,480 --> 00:49:09,360
Yeye aliamua kwamba alikuwa anaenda
kujenga kweli ulimwengu virtual.

868
00:49:09,360 --> 00:49:12,820
Hiyo ni mradi kabisa, wakati wewe
unafikiri nimepata lami hii.

869
00:49:12,820 --> 00:49:14,740
Mimi nina kwenda kutumia
kompyuta, na mimi nina kwenda

870
00:49:14,740 --> 00:49:21,040
kwa mfano miaka milioni 12 baada ya
Big Bang kuwakilisha siku.

871
00:49:21,040 --> 00:49:27,080
Na kisha mimi nina kwenda kufanya 13.8
miaka bilioni ya cosmic mageuzi.

872
00:49:27,080 --> 00:49:28,270
Wote haki.

873
00:49:28,270 --> 00:49:30,970
>> Hii kwa kweli anatumia kompyuta
ilikuwa kubwa kuliko kompyuta yetu,

874
00:49:30,970 --> 00:49:35,040
na kilichomwagika juu kwenye kitaifa
rasilimali na marafiki zetu chini katika Texas.

875
00:49:35,040 --> 00:49:38,820
Na vifaa vya kitaifa,
hii ilikuwa mengi ya compute.

876
00:49:38,820 --> 00:49:40,750
Lakini sisi alifanya mengi ya
simulation ndani ya nchi

877
00:49:40,750 --> 00:49:44,820
kuhakikisha kwamba programu
kazi na mifumo ya kazi.

878
00:49:44,820 --> 00:49:47,790
>> Na ni siku kama hii wakati wewe
kutambua kwamba wewe ni kusaidia sayansi

879
00:49:47,790 --> 00:49:51,090
katika ngazi hii ya wadogo, kwamba
sasa watu wanaweza kusema mambo

880
00:49:51,090 --> 00:49:52,840
kama, mimi nina kwenda kwa mfano ulimwengu.

881
00:49:52,840 --> 00:49:54,145
Na hii ni mfano wake wa kwanza.

882
00:49:54,145 --> 00:49:56,422
Na hii ni kielelezo cha kwanza timu yake.

883
00:49:56,422 --> 00:49:58,130
Kuna wengine wengi
folks kwamba ni kwenda

884
00:49:58,130 --> 00:50:01,520
kuja nyuma Mark, ambao ni kwenda
wanataka mfano na azimio juu,

885
00:50:01,520 --> 00:50:04,652
na zaidi maalum,
kwa usahihi zaidi.

886
00:50:04,652 --> 00:50:09,105
>> Na hivyo, katika michache ya mwisho ya dakika,
Mimi nataka tu kuonyesha video hii

887
00:50:09,105 --> 00:50:15,270
ya Mark na Lars ya kwamba mimi, tena,
kama maisha mwanasayansi, ni aina ya cute.

888
00:50:15,270 --> 00:50:17,890

889
00:50:17,890 --> 00:50:20,970
Hivyo hii, chini
hapa, na kuelekeza wewe,

890
00:50:20,970 --> 00:50:23,640
hii ni kuwaambia wewe
muda tangu Big Bang.

891
00:50:23,640 --> 00:50:26,570
Hivyo sisi ni saa kuhusu 0700000000 miaka.

892
00:50:26,570 --> 00:50:28,740
Na hii ni kuonyesha update sasa.

893
00:50:28,740 --> 00:50:33,450
Hivyo wewe ni kuona kwa wakati huu,
giza jambo na mageuzi

894
00:50:33,450 --> 00:50:39,910
muundo wa faini na mapema
miundo katika ulimwengu wetu inayojulikana.

895
00:50:39,910 --> 00:50:45,690
>> Na uhakika na hii ni kwamba hii
yote ni kufanyika ndani ya kompyuta.

896
00:50:45,690 --> 00:50:48,530
Hii ni seti ya vigezo
na seti ya fizikia

897
00:50:48,530 --> 00:50:52,840
na seti ya hisabati
na seti ya mifano

898
00:50:52,840 --> 00:50:59,284
kwamba ni uangalifu, na kisha
makini kushikamana na kila mmoja

899
00:50:59,284 --> 00:51:00,825
kuwa na uwezo wa kubuni mwingiliano.

900
00:51:00,825 --> 00:51:04,850
>> Hivyo unaweza kuona baadhi kuanza ya
baadhi milipuko gaseous hapa.

901
00:51:04,850 --> 00:51:06,880
Na gesi joto ni kubadilika.

902
00:51:06,880 --> 00:51:13,720
Na unaweza kuanza kuona muundo
inayoonekana mabadiliko ya ulimwengu.

903
00:51:13,720 --> 00:51:18,130
Na sehemu muhimu na hii
ni, kila kidogo kidogo, vidogo, vidogo dot

904
00:51:18,130 --> 00:51:21,070
ni kipande ya fizikia na ina
seti ya hisabati kote,

905
00:51:21,070 --> 00:51:23,030
kuwafahamisha rafiki yake na jirani yake.

906
00:51:23,030 --> 00:51:27,245
>> Hivyo kutokana na kuongeza mtazamo, haya
kompyuta na kazi yote katika tamasha

907
00:51:27,245 --> 00:51:29,470
na kuzungumza na kila mmoja kwa ufanisi.

908
00:51:29,470 --> 00:51:31,060
Hivyo hawawezi kuwa pia chatty.

909
00:51:31,060 --> 00:51:33,520
Wana kuhifadhi matokeo yao.

910
00:51:33,520 --> 00:51:37,902
Na wao wana kuendelea
kuwajulisha wote wa marafiki zao.

911
00:51:37,902 --> 00:51:40,860
Hakika, utaona sasa, mtindo huu ya
kupata zaidi na zaidi ngumu.

912
00:51:40,860 --> 00:51:42,590
Kuna mambo zaidi na zaidi kinachoendelea.

913
00:51:42,590 --> 00:51:45,210
Kuna zaidi na zaidi
nyenzo wanaruka kote.

914
00:51:45,210 --> 00:51:48,410
>> Na hii ni nini mapema
cosmos ingekuwa wameweza inaonekana kama.

915
00:51:48,410 --> 00:51:49,770
Ilikuwa ni sehemu pretty nywele.

916
00:51:49,770 --> 00:51:55,140
Kuna milipuko yote juu ya
mahali, collisions nguvu.

917
00:51:55,140 --> 00:51:58,620
Na malezi ya nzito
madini na vipengele.

918
00:51:58,620 --> 00:52:03,910
Na mawingu hayo makubwa smashing ndani
kila mmoja kwa nguvu uliokithiri.

919
00:52:03,910 --> 00:52:08,530
>> Na hivyo sasa tuko 9600000000
miaka kutokana na mlipuko huu wa awali.

920
00:52:08,530 --> 00:52:12,310
Wewe ni mapya ya kuona mambo ni
aina ya tulivu kidogo, tu

921
00:52:12,310 --> 00:52:15,660
kidogo, kwa sababu
nishati ni sasa kuanza kupumzika.

922
00:52:15,660 --> 00:52:19,420
Na hivyo hisabati
mifano got kuwa katika nafasi.

923
00:52:19,420 --> 00:52:22,510
Na wewe ni mapya ya kuona
coalescence wa mambo mbalimbali.

924
00:52:22,510 --> 00:52:26,220
Na kuanzia kuona jambo hili aina
ya kuja pamoja na polepole baridi.

925
00:52:26,220 --> 00:52:32,260
>> Na ni mapya ya kuangalia kidogo
zaidi kama anga la usiku, kidogo.

926
00:52:32,260 --> 00:52:37,870
Na ni [? QSing. ?] Sisi ni sasa 30.2
miaka bilioni na sisi ni aina ya kosa.

927
00:52:37,870 --> 00:52:41,130
Na kisha nini alifanya alikuwa
kwamba alichukua mtindo huu,

928
00:52:41,130 --> 00:52:44,580
na kisha inaonekana katika ulimwengu inayoonekana.

929
00:52:44,580 --> 00:52:48,560
Na kimsingi basi, walikuwa
uwezo wa kuchukua kwamba na kuifunika

930
00:52:48,560 --> 00:52:50,580
ni na nini unaweza kuona.

931
00:52:50,580 --> 00:52:56,160
Na uaminifu ni kubwa, kama kwa
jinsi sahihi mifano ya kompyuta ni.

932
00:52:56,160 --> 00:52:58,760
>> Bila shaka, astrophysicists
na makundi ya utafiti

933
00:52:58,760 --> 00:53:02,780
haja uaminifu hata bora
na azimio hata ya juu.

934
00:53:02,780 --> 00:53:06,230
Lakini kama wewe kufikiri juu ya nini
Nimekuwa kuzungumza na wewe leo

935
00:53:06,230 --> 00:53:11,850
njia hii safari kidogo kupitia wawili
kuhifadhi na muundo na mitandao

936
00:53:11,850 --> 00:53:18,000
na mwingi, jambo muhimu ni,
ni ukubwa-out kompyuta muhimu?

937
00:53:18,000 --> 00:53:22,050
Hiyo ilikuwa hypothesis-- yangu ya awali
nyuma kwa njia yetu ya kisayansi.

938
00:53:22,050 --> 00:53:24,810
>> Natumaini kwamba katika mapema
sehemu ya huu napenda

939
00:53:24,810 --> 00:53:29,400
kutabiri kwamba mimi itakuwa na uwezo wa kueleza
na wewe kuhusu wadogo-out kompyuta.

940
00:53:29,400 --> 00:53:32,870
Na sisi aina ya majaribio
baadhi ya hypotheses hizo.

941
00:53:32,870 --> 00:53:34,585
Tulikwenda kwa njia ya mazungumzo haya.

942
00:53:34,585 --> 00:53:38,920
Na mimi nina kwenda tu kusema wadogo-nje
kompyuta ni essential-- oh,

943
00:53:38,920 --> 00:53:42,480
ndiyo, sana ndiyo.

944
00:53:42,480 --> 00:53:44,790
>> Hivyo wakati wewe ni kufikiri
kuhusu codes yako, wakati

945
00:53:44,790 --> 00:53:49,230
unafanya CS50 miradi mwisho,
wakati wewe ni kufikiri kuhusu urithi wako

946
00:53:49,230 --> 00:53:52,990
kwa binadamu na rasilimali kwamba sisi
haja ya kuwa na uwezo wa kuendesha kompyuta hizi

947
00:53:52,990 --> 00:53:56,650
mifumo, kufikiri kwa makini sana
kuhusu flops kwa watt,

948
00:53:56,650 --> 00:53:58,560
na kufikiri kuhusu machafuko Monkey.

949
00:53:58,560 --> 00:54:02,240
>> Fikiria juu ya snowflakes yako, kufanya si
kufanya moja-awamu ya pili, maktaba reuse,

950
00:54:02,240 --> 00:54:06,453
kujenga reusable codes-- mambo yote
kwamba wakufunzi wamekuwa kufundisha

951
00:54:06,453 --> 00:54:08,630
katika darasa hili.

952
00:54:08,630 --> 00:54:11,942
Hizi ni mambo ya msingi.

953
00:54:11,942 --> 00:54:13,150
Wao siyo tu mdomo huduma.

954
00:54:13,150 --> 00:54:15,660
Haya ni mambo ya kweli.

955
00:54:15,660 --> 00:54:20,680
>> Na kama yoyote ya unataka nifuate
Mimi ni obsessive na jambo Twitter.

956
00:54:20,680 --> 00:54:22,770
Mimi nimepata kwa namna fulani kutoa kwamba up.

957
00:54:22,770 --> 00:54:24,960
Lakini mengi ya
taarifa za msingi ni

958
00:54:24,960 --> 00:54:29,260
juu ya kompyuta utafiti wetu
Tovuti rc.fas.harvard.edu.

959
00:54:29,260 --> 00:54:34,010
>> Mimi kujaribu na kuendelea blog hadi
tarehe na teknolojia ya kisasa

960
00:54:34,010 --> 00:54:38,390
na jinsi sisi kufanya distributive
kompyuta na kadhalika.

961
00:54:38,390 --> 00:54:43,600
Na kisha wafanyakazi wetu ni daima
inapatikana kwa njia ya odybot.org.

962
00:54:43,600 --> 00:54:46,270
Na odybot ni msaidizi wetu kidogo.

963
00:54:46,270 --> 00:54:49,280
Yeye mara nyingi ina kidogo
mashindano kwenye tovuti yake

964
00:54:49,280 --> 00:54:51,630
pia, ambapo unaweza kujaribu na
doa yake kuzunguka chuo.

965
00:54:51,630 --> 00:54:55,200
Yeye ni rafiki kidogo
uso wa kompyuta utafiti.

966
00:54:55,200 --> 00:54:59,730
>> Na mimi itabidi aina ya wrap kuna up
na kuwashukuru wote kwa muda wako.

967
00:54:59,730 --> 00:55:05,660
Na natumaini unakumbuka kuwa
wadogo-out kompyuta ni kitu halisi.

968
00:55:05,660 --> 00:55:08,162
Na kuna mengi ya watu
ambao wameweza got mengi ya sanaa kabla

969
00:55:08,162 --> 00:55:09,370
ambao watakuwa na uwezo wa kukusaidia.

970
00:55:09,370 --> 00:55:14,330
Na wote wa bora wa bahati na
juhudi yako ya baadaye katika kufanya

971
00:55:14,330 --> 00:55:18,280
kuhakikisha kwamba yetu kompyuta wote
mizani, ni juu performing,

972
00:55:18,280 --> 00:55:20,370
na husaidia utu zaidi
kuliko kitu kingine chochote.

973
00:55:20,370 --> 00:55:22,850
Hivyo, asante kwa muda wako.

974
00:55:22,850 --> 00:55:23,947