WebCore: Architectural Support for Mobile Web Browsing

ISCA 2014 Main talk

Transcript

1. WebCore: Architectural Support for Mobile Web Browsing Yuhao Zhu, Vijay Janapa Reddi Department of Electrical and Computer Engineering The University of Texas at Austin ISCA MainTalk — June 18th, 2014
7. Swift
9. The Fundamental Challenges 4
10. The Fundamental Challenges 4 Achieving High Performance Demanded by End-User
11. The Fundamental Challenges 4 Achieving High Performance Demanded by End-User Conserving Energy Due to Limited Battery Capacity
12. The Fundamental Challenges 4 Achieving High Performance Demanded by End-User Conserving Energy Due to Limited Battery Capacity Conflicting requirements
13. The Fundamental Challenges How to achieve high performance with low energy? 4 Achieving High Performance Demanded by End-User Conserving Energy Due to Limited Battery Capacity Conflicting requirements
14. The Fundamental Challenges How to achieve high performance with low energy? 4 Achieving High Performance Demanded by End-User Conserving Energy Due to Limited Battery Capacity Conflicting requirements A mobile architecture
15. The Fundamental Challenges How to achieve high performance with low energy? 4 Achieving High Performance Demanded by End-User Conserving Energy Due to Limited Battery Capacity Conflicting requirements A mobile architecture WebCore:
16. Executive Summary 5 Time Energy General Purpose Designs
17. Executive Summary 5 Time Energy General Purpose Designs Diminishing return
18. Executive Summary 5 Time Energy General Purpose Designs ASIC?
19. Executive Summary 5 Time Energy General Purpose Designs ASIC? Extremely challenging ‣Chrome: 7M LoC, 29 languages ‣Firefox: 10M LoC, 33 languages
20. Executive Summary 5 Time Energy General Purpose Designs ASIC?
21. Executive Summary 5 Time Energy General Purpose Designs ASIC? WebCore Goal
22. Executive Summary 5 Time Energy General Purpose Designs ??? ASIC? WebCore Goal
23. Executive Summary 6 Time Energy General Purpose Designs WebCore Goal
24. Executive Summary 6 Time Energy General Purpose Designs WebCore Goal
25. Executive Summary 6 Time Energy General Purpose Designs Customizing µarch Parameters WebCore Goal
26. Executive Summary 6 Time Energy General Purpose Designs Customizing µarch Parameters Specialized FU and Memory WebCore Goal
27. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web 7
28. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization 7
29. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization ▸Evaluation Results 7
30. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization ▸Evaluation Results ▸Related Work 7
31. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization ▸Evaluation Results ▸Related Work 8
32. Customization: Find the Ideal General Purpose Baseline Architecture
33. ▸Why customization?!? Customization: Find the Ideal General Purpose Baseline Architecture
34. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? Customization: Find the Ideal General Purpose Baseline Architecture
35. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? ▹Out-of-order (Silvermont, A15) or in-order (Saltwell, A7)? Customization: Find the Ideal General Purpose Baseline Architecture
36. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? ▹Out-of-order (Silvermont, A15) or in-order (Saltwell, A7)? ▹Are existing general purpose mobile designs ideal? Customization: Find the Ideal General Purpose Baseline Architecture
37. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? ▹Out-of-order (Silvermont, A15) or in-order (Saltwell, A7)? ▹Are existing general purpose mobile designs ideal? ▸Exhaustive design space exploration Customization: Find the Ideal General Purpose Baseline Architecture
38. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? ▹Out-of-order (Silvermont, A15) or in-order (Saltwell, A7)? ▹Are existing general purpose mobile designs ideal? ▸Exhaustive design space exploration Customization: Find the Ideal General Purpose Baseline Architecture
39. ▸Why customization?!? ▸What is a proper general purpose baseline architecture? ▹Out-of-order (Silvermont, A15) or in-order (Saltwell, A7)? ▹Are existing general purpose mobile designs ideal? ▸Exhaustive design space exploration Customization: Find the Ideal General Purpose Baseline Architecture
40. Design Space Exploration (DSE) Setup ▸Integrated power (McPAT) and performance x86 full-system simulator (Marss86) ▸WebKit engine in the Chromium Web browser 10
41. Design Space Exploration (DSE) Setup ▸Integrated power (McPAT) and performance x86 full-system simulator (Marss86) ▸WebKit engine in the Chromium Web browser 10
42. Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA
43. ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA
44. ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA 10-4 10-3 10-2 10-1 100 101 PC2 (log) -5 0 5 PC1
45. ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA 10-4 10-3 10-2 10-1 100 101 PC2 (log) -5 0 5 PC1 dominated by # webpage elements
46. ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA 10-4 10-3 10-2 10-1 100 101 PC2 (log) -5 0 5 PC1 dominated by IPC
47. ▹PCs calculated from webpage-inherent and µarch-dependent features (~400 in total) 10-4 10-3 10-2 10-1 100 101 PC2 (log) -5 0 5 PC1 Design Space Exploration (DSE) Setup 11 ▸Webpages selection using PCA
48. Design Space Exploration (DSE) Findings 12
49. Design Space Exploration (DSE) Findings 12
50. Design Space Exploration (DSE) Findings 12
51. Design Space Exploration (DSE) Findings ▸Out-of-order µarchitecture is much more flexible 12
52. Design Space Exploration (DSE) Findings ▸Out-of-order µarchitecture is much more flexible 12
53. Design Space Exploration (DSE) Findings ▸Out-of-order µarchitecture is much more flexible 12 ▸In-order cores are acceptable if end-users can tolerate latency
54. Understand the Difference Using Kernel Knowledge 13
55. Understand the Difference Using Kernel Knowledge 13 Execution time breakdown
56. Understand the Difference Using Kernel Knowledge In-order design 13
57. Understand the Difference Using Kernel Knowledge In-order design 13
58. ▸In-order designs show strong kernel variance Understand the Difference Using Kernel Knowledge In-order design 13
59. ▸In-order designs show strong kernel variance Understand the Difference Using Kernel Knowledge In-order design 13
60. ▸In-order designs show strong kernel variance Understand the Difference Using Kernel Knowledge In-order design 13
61. ▸In-order designs show strong kernel variance Understand the Difference Using Kernel Knowledge In-order design 13 Out-of-order design
62. ▸In-order designs show strong kernel variance Understand the Difference Using Kernel Knowledge In-order design 13 Out-of-order design ▸An Out-of-order design can accommodate kernel variance
63. 14 Customization: Identifying Major Sources of Energy Inefficiency
64. 14 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
65. 14 Customization: Identifying Major Sources of Energy Inefficiency P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096
66. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
67. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
68. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
69. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 ▸Instruction delivery 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
70. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 ▸Instruction delivery 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
71. P1 P2 ARM A15 Issue width 1 3 3 # Function units 2 3 8 Load queue size 4 16 16 Store queue size 4 16 16 BTB size 1024 128 256 ROB size 128 128 40+ L1 I-$ size (KB) 64 128 32 # Physical registers 128 140 ? L1 D-$ size (KB) 8 64 32 L2-$ size (KB) 256 1024 <4096 ▸Instruction delivery ▸Data feeding 15 P2 P1 Customization: Identifying Major Sources of Energy Inefficiency
72. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization -Mitigate instruction delivery: Style resolution unit (SRU) -Improving data feeding: Browser engine cache ▸Evaluation Results ▸Related Work 16
73. WebCore Specialization Overview 17 Customized core IF ID EX MEM WB Hardware Layer
74. WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU Hardware Layer
75. WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Hardware Layer
76. L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Hardware Layer
77. L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Hardware Layer Browser Engine Cache
78. L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Hardware Layer API Layer Browser Engine Cache
79. L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Browser Engine Cache
80. DOM_LD(Id, &attr); DOM_ST(Id, &attr); L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Browser Engine Cache
81. DOM_LD(Id, &attr); DOM_ST(Id, &attr); L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Runtime Layer Browser Engine Cache
82. DOM_LD(Id, &attr); DOM_ST(Id, &attr); L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Runtime Layer Cache Management Browser Engine Cache
83. DOM_LD(Id, &attr); DOM_ST(Id, &attr); L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Runtime Layer Cache Management SRU Access Browser Engine Cache
84. DOM_LD(Id, &attr); DOM_ST(Id, &attr); L1 D-cache WebCore Specialization Overview 17 Customized core IF ID MEM WB ALU MUL FPU SRU Style_apply(Id); Hardware Layer API Layer Runtime Layer Cache Management Software Failsafe SRU Access Browser Engine Cache
85. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization -Mitigate instruction delivery: Style resolution unit (SRU) -Improving data feeding: Browser engine cache ▸Evaluation Results ▸Related Work 18
86. ▸Style kernel is the most critical kernel Style Resolution Unit 19
87. ▸Style kernel is the most critical kernel Style Resolution Unit 19 Execution time breakdown Energy consumption breakdown
88. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}}
89. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}}
90. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}} Rule-level Parallelism (RLP)
91. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}} Rule-level Parallelism (RLP)
92. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}} Rule-level Parallelism (RLP) Property-level Parallelism (PLP)
93. ▸Style kernel is the most critical kernel Style Resolution Unit 19 for (each rule in matchedRules) { for (each property in rule) { switch (property.id) { case Font: Style[Font] = Handler(property.value, DOMNode); break; case N: ...}}} Rule-level Parallelism (RLP) Property-level Parallelism (PLP) ▸Exploiting the parallelism to increase the arithmetic intensity and reduce instruction footprint
94. ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0
95. ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
96. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
97. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
98. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
99. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
100. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
101. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px margin 0 High priority
102. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px ▸Order Matters in RLP ▸Order Does Not Matter in PLP margin 0 High priority
103. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px ▸Order Matters in RLP ▸Order Does Not Matter in PLP margin 0 High priority
104. Property 1 Property 1 Property 2 Property 2 Property 3 Property 3 id value id value id value Final Style Info ▸A running example from www.cnn.com Style Resolution Unit (2) Rule Property 1 Property 1 Property 2 Property 2 Rule id value id value 1 padding 0 margin 0 2 padding 6 px width 36 px Style Rules padding 0 width 6 px 36 px ▸Order Matters in RLP ▸Order Does Not Matter in PLP margin 0 High priority
105. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 ▸Order Matters in RLP ▸Order Does Not Matter in PLP
106. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory ▸Order Matters in RLP ▸Order Does Not Matter in PLP
107. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority
108. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory Conflict Resolution ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority
109. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory Conflict Resolution ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority Prop m Prop m
110. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory Conflict Resolution ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority Prop m
111. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory Conflict Resolution Compute Lanes ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority
112. ... ... Rule j ... ... Prop l ... ... Rule i.id ... Prop m ... Prop k ... Rule j.id ... ... ... ... ... start end start end Rule i Prop k Prop m Prop m Prop l Style l Style m Style k Style Resolution Unit (3) 21 Input Scratchpad Memory Output Scratchpad Memory Conflict Resolution Compute Lanes ▸Order Matters in RLP ▸Order Does Not Matter in PLP Higher Priority
113. Agenda of Today’s Talk ▸Motivation of our work: energy-efficiency of the mobile Web ▸How does WebCore improve the energy-efficiency? ▹Customization ▹Specialization ▸Evaluation Results ▸Related Work 22
114. Evaluations 23 ▸Fully synthesized using Synopsys 28 nm toolchain
115. Evaluations 23 ▸Fully synthesized using Synopsys 28 nm toolchain ▸24 representative webpages
116. Evaluations 23 ▸Fully synthesized using Synopsys 28 nm toolchain ▸24 representative webpages www.amazon.com www.cnn.com www.msn.com www.google.com.hk www.twitter.com www.espn.go.com www.bbc.co.uk www.slashdot.org www.youtube.com www.ebay.com www.sina.com.cn www.163.com Desktop and mobile versions
117. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s)
118. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design
119. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization
120. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) 18.6% A15-like design Customization
121. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) 18.6% 22.2% A15-like design Customization
122. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) 18.6% 22.2% A15-like design Customization Specialization
123. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) 18.6% 22.2% 22.2% A15-like design Customization Specialization
124. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) 18.6% 22.2% 9.2% 22.2% A15-like design Customization Specialization
125. Evaluations 24 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization 29.2% 47.0%
126. Evaluations 25 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization Cost of specialization: 0.59 mm2 area overhead
127. Evaluations 25 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization Cost of specialization: 0.59 mm2 area overhead Better than scaling- up approaches
128. Evaluations 25 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization Cost of specialization: 0.59 mm2 area overhead Better than scaling- up approaches I$
129. Evaluations 25 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization Cost of specialization: 0.59 mm2 area overhead Better than scaling- up approaches D$
130. Evaluations 25 0.55 0.688 0.825 0.963 1.1 1.6 1.8 2 2.2 2.4 Energy (J) Load Time (s) A15-like design Customization Specialization Cost of specialization: 0.59 mm2 area overhead Better than scaling- up approaches I+D$
131. Related Work 26 Hardware Software Focus on Performance Focus on Energy-Efficiency
132. Related Work 26 Hardware Software Focus on Performance Focus on Energy-Efficiency Parallelization Algorithm- level Zoomm Mozilla Servo
133. Related Work 26 Hardware Software Focus on Performance Focus on Energy-Efficiency Parallelization Algorithm- level Zoomm Mozilla Servo System- level Optimizations Redundancy Removal Prefetching Big/little Scheduling
134. Related Work 26 Hardware Software Focus on Performance Focus on Energy-Efficiency Parallelization Algorithm- level Zoomm Mozilla Servo ASIC Tegra 4 WebRTC accelerator SiChrome System- level Optimizations Redundancy Removal Prefetching Big/little Scheduling
135. Related Work 26 Hardware Software Focus on Performance Focus on Energy-Efficiency Parallelization Algorithm- level Zoomm Mozilla Servo ASIC Tegra 4 WebRTC accelerator SiChrome System- level Optimizations Redundancy Removal Prefetching Big/little Scheduling WebCore
136. Conclusions 27 The Web browser has become a general purpose platform that supports a wide range of mobile Web applications Customization allows us to find the ideal general-purpose baseline architecture Hardware/software collaborative specialization leverages application knowledge to mitigate inefficiencies in general-purpose architectures
137. Thank you