Silvermont. Introducing Next Generation Low Power Microarchitecture: Dadi Perlmutter

Transcription

1 Introducing Next Generation Low Power Microarchitecture: Silvermont Dadi Perlmutter Executive Vice President General Manager, Intel Architecture Group Chief Product Officer

2 Risk Factors Today s presentations contain forward-looking statements. All statements made that are not historical facts are subject to a number of risks and uncertainties, and actual results may differ materially. Please refer to our most recent earnings release, Form 10-Q and 10-K filing available for more information on the risk factors that could cause actual results to differ. If we use any non-gaap financial measures during the presentations, you will find on our website, intc.com, the required reconciliation to the most directly comparable GAAP financial measure. Rev. 4/16/13

3 Legal Disclaimers Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to: Intel, Intel Atom and the Intel logo are trademarks of Intel Corporation in the United States and other countries. 1 Based on the geometric mean of a variety of power and performance measurements across various benchmarks. Benchmarks included in this geomean are measurements on browsing benchmarks and workloads including SunSpider* and page load tests on Internet Explorer*, FireFox*, & Chrome*; Dhrystone*; EEMBC* workloads including CoreMark*; Android* workloads including CaffineMark*, AnTutu*, Linpack* and Quadrant* as well as measured estimates on SPECint* rate_base2000 & SPECfp* rate_base2000; on Silvermont preproduction systems compared to Atom processor Z2580. Individual results will vary. SPEC* CPU2000* is a retired benchmark. * Other names and brands may be claimed as the property of others. 2 Based on a geometric mean of the projected power and performance of SPECint* rate_base2000 on Silvermont compared to expected configurations of main ARM*-based mobile competitors using descriptions of the architectures; assumes similar configurations. Numbers may be subject to change once verified with the actual parts. Individual results will vary. SPEC* CPU2000* is a retired benchmark; results are estimates. * Other names and brands may be claimed as the property of others 3 Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance.

4 Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to: Introducing Silvermont Next Gen Microarchitecture 22nm SoC Process ~3X Higher Performance 1 ~5X Lower Power 1 1 Based on the geometric mean of a variety of power and performance measurements across various benchmarks. Benchmarks included in this geomean are measurements on browsing benchmarks and workloads including SunSpider* and page load tests on Internet Explorer*, FireFox*, & Chrome*; Dhrystone*; EEMBC* workloads including CoreMark*; Android* workloads including CaffineMark*, AnTutu*, Linpack* and Quadrant* as well as measured estimates on SPECint* rate_base2000 & SPECfp* rate_base2000; on Silvermont preproduction systems compared to Atom processor Z2580. Individual results will vary. SPEC* CPU2000* is a retired benchmark. * Other names and brands may be claimed as the property of others.

5 Best In Class Performance Silvermont Out of Order Execution engine New multi-core and system fabric architecture Capable of delivering 8 cores Higher single thread performance New IA instructions and core technologies Best In Class Power & Power Efficiency Wider dynamic power operating range Enhanced power management Fast standby entry/exit Great Scalability Across Multiple Market Segments

6 Architecture & Microarchitecture Definitions Architecture Instruction Set Definition Example Intel 64 IA-32 Analogy Engine Type Microarchitecture HW Implementation for ISA Ivy Bridge Haswell Saltwell Silvermont Engine Schematic SoC (System on a Chip) Single die with integrated IPs Haswell Crypto Storage Graphics Audio CPU Video Display Cloverview Type of Vehicle

7 The Constantly Evolving Intel Architecture Performance Power Efficiency Media Manageability Security SSE Extensions AVX Instructions AES-NI Smart Caches Turbo Boost Technology Intel Hyper-Threading Intel Matrix Storage... Thermal Monitoring Technologies Idle States Intel Smart Idle Intel SpeedStep Technology Intel Demand Based Switching... Multimedia Instructions Intel Quick Sync Video Intel HD Graphics Intel HD Audio Intel Clear Video HD Intel Flexible Display Technolgy Intel Insider... Intel Virtualization Technology Intel vpro Technology Intel CIRA Technology Intel AMT Intel Remote PC Assist Intel Remote Wake... Intel Anti-Theft Technology Intel Trusted Execution Technology Intel Secure Key Intel Run Secure Execute Disable Bit...

8 Intel Solutions For All Segments Intel Atom Datacenter and HPC Ultrabook Tablets Phones Intelligent Systems From TERAFLOPS to MILLIWATTS

9 Intel Solutions For All Segments Intel Atom Datacenter and HPC Ultrabook Tablets Phones Intelligent Systems From TERAFLOPS to MILLIWATTS

10 Enabling a Full Spectrum of Mobile Products Ultrabooks Convertible Tablet Smartphone Mobility Spectrum Best Performance and Outstanding Battery Life Best Visuals Best Battery Life Thinnest and Lightest Form Factors Intel Atom

11 As Well as Datacenter and Intelligent Systems Solutions Server Network Storage Embedded IVI Highest Performance Best Performance/Watt Highest Density Designs Lowest Power Intel Atom

12 Silvermont Microarchitecture in Next Generation Intel Products From DATACENTER to DEVICES Avoton Rangeley Baytrail Merrifield TBA Port of Choice Enabling Exceptional Experiences * Other names and brands may be claimed as the property of others.

13 Committed Leadership Roadmap 45nm 32nm 22nm 14nm 14nm Silvermont; The First in a Family of Yearly Refreshes

14 Intel s IDM Advantage PROCESS TECHNOLOGY INTEL ARCHITECTURE PRODUCT DESIGN Common Tools Common Goals SOFTWARE PACKAGING MANUFACTURING

15 Silvermont Technical Overview Belli Kuttanna Intel Fellow Intel Architecture Group

16 Key Messages Next Generation Low Power Microarchitecture Manufactured in custom Intel 22nm SOC process technology Present in a wide range of low power products from Phones to Servers ~3x peak performance 1 improvement or the same performance at ~5x lower power 1 over the current generation Atom core Leading performance and performance per watt efficiency 2 First in a family of cores that will be refreshed every year 1 Based on the geometric mean of a variety of power and performance measurements across various benchmarks. Benchmarks included in this geomean are measurements on browsing benchmarks and workloads including SunSpider* and page load tests on Internet Explorer*, FireFox*, & Chrome*; Dhrystone*; EEMBC* workloads including CoreMark*; Android* workloads including CaffineMark*, AnTutu*, Linpack* and Quadrant* as well as measured estimates on SPECint* rate_base2000 & SPECfp* rate_base2000; on Silvermont preproduction systems compared to Atom processor Z2580. Individual results will vary. SPEC* CPU2000* is a retired benchmark. * Other names and brands may be claimed as the property of others. 2 Based on a geometric mean of the projected power and performance of SPECint* rate_base2000 on Silvermont compared to expected configurations of main ARM*-based mobile competitors using descriptions of the architectures; assumes similar configurations. Numbers may be subject to change once verified with the actual parts. Individual results will vary. SPEC* CPU2000* is a retired benchmark; results are estimates. * Other names and brands may be claimed as the property of others. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

17 Better Performance Silvermont Highlights Better Power Efficiency Full Advantage of Intel 22nm SoC Process Technology Out of Order Execution engine New multi-core and system fabric architecture New IA instructions extensions (Core Westmere level) New security and virtualization technologies Wider dynamic operating range Enhanced active and idle power management ~3X the Performance Or ~5X Lower Power 1 3-D Tri-gate transistors tuned for SoC products Architecture and design cooptimized with the process 1 Based on the geometric mean of a variety of power and performance measurements across various benchmarks. Benchmarks included in this geomean are measurements on browsing benchmarks and workloads including SunSpider* and page load tests on Internet Explorer*, FireFox*, & Chrome*; Dhrystone*; EEMBC* workloads including CoreMark*; Android* workloads including CaffineMark*, AnTutu*, Linpack* and Quadrant* as well as measured estimates on SPECint* rate_base2000 & SPECfp* rate_base2000; on Silvermont preproduction systems compared to Atom processor Z2580. Individual results will vary. SPEC* CPU2000* is a retired benchmark. * Other names and brands may be claimed as the property of others. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

18 Agenda

19 Agenda

20 Silvermont Micro-Architecture BENEFITS High Performance Without Sacrificing Power Efficiency FEATURES Out-of-Order Execution Pipeline Macro operation execution pipeline Improved instruction latencies and throughput Smart pipeline resource management Power and Performance Improvements Efficient Branch Processing Accurate branch predictors Fast recovery pipeline Faster and More Efficient Access to Memory Low Latency, high bandwidth caches Out of order memory transactions Multiple advanced hardware prefetchers Balanced core and memory subsystems Building upon Intel s Strengths and Expertise In Defining Microarchitectures

21 Multi-Core Support Module building block-based expansion from 1 to 8 cores Module contains: Two cores Tightly coupled second-level cache (up to 1MB): very low latency, high bandwidth Dedicated point-to-point interface (IDI) to SOC Fabric Independent read, write channels Module A Core Core L2 Cache System Agent Module B Core Core L2 Cache Higher bandwidth, Lower Latency, OOO transaction support Per-core frequency and power management support Balanced Core and Memory Subsystem Design IDI IDI Memory Controller DRAM

22 Evolving Architecture: New Instructions and Technologies Performance Virtualization Security VM 1 VM 2 VMM New Instructions Intel Core 2 64b ISA + Core Westmere SSE4.1, SSE4.2, POPCNT Intel Core Westmere AES-NI, Intel Secure Key New Technologies Real Time Instruction Tracing TSC Deadline Timer Intel VT-x2: Extended Page Tables Virtual Processor IDs Intel OS Guard Support for McAfee DeepSAFE Unrestricted Guest Fully Compatible with the Breadth of IA Software Installed Base

23 Agenda

24 Process/Design/Architecture Co-optimization + + = 22nm SoC process 3D Transistors that offer unprecedented improvement in performance and power Metal stack co-optimized with key SOC IPs for density and performance Multiple 22nm process versions optimized for specific products Architecture for Low Power Completely redefined for efficiency and scalability Design for Low Power Custom arrays and libraries optimized for power and performance Power-Aware design and automation Wide Dynamic Range of Operation + High Efficiency Unique Intel Leadership in Process, Design and Architecture

25 Burst Mode Improvements Prior Atom cores Opportunistically exposed additional P-states based on available thermal headroom Silvermont enhancements Burst frequency managed in hardware based on Thermal, Electrical and Power Delivery constraints Power sharing between CPU cores and SOC IPs (e.g. Graphics) is supported Burst operating points can be dynamically adjusted Power Sharing Between Cores Core0 Core1 GFX Maximize Performance within Platform Capabilities

26 Burst Mode Improvements Prior Atom cores Opportunistically exposed additional P-states based on available thermal headroom Silvermont enhancements Burst frequency managed in hardware based on Thermal, Electrical and Power Delivery constraints Power sharing between CPU cores and SOC IPs (e.g. Graphics) is supported Burst operating points can be dynamically adjusted Power Sharing Between Cores Core0 Core1 GFX Maximize Performance within Platform Capabilities

27 Burst Mode Improvements Prior Atom cores Opportunistically exposed additional P-states based on available thermal headroom Silvermont enhancements Burst frequency managed in hardware based on Thermal, Electrical and Power Delivery constraints Power sharing between CPU cores and SOC IPs (e.g. Graphics) is supported Burst operating points can be dynamically adjusted Power Sharing Between Cores Power Sharing Between Cores and GFX Core0 Core1 GFX Core0 Core1 GFX Maximize Performance within Platform Capabilities

28 Burst Mode Improvements Prior Atom cores Opportunistically exposed additional P-states based on available thermal headroom Silvermont enhancements Burst frequency managed in hardware based on Thermal, Electrical and Power Delivery constraints Power sharing between CPU cores and SOC IPs (e.g. Graphics) is supported Burst operating points can be dynamically adjusted Power Sharing Between Cores Power Sharing Between Cores and GFX Dynamic Burst Core0 Core1 GFX Core0 Core1 GFX Core0 Core1 GFX Maximize Performance within Platform Capabilities

29 Core State Power C-States Core Clock L1 Caches Core Voltage Wakeup Time C0 C1/C2 C6 OFF OFF off active

30 Module State Core State Power C-States Core Clock L1 Caches Core Voltage Wakeup Time L2 Cache PLL Wakeup Time C0 C1/C2 C6 active active OFF active partial flush OFF off retention OFF off OFF Improved C6 Modes with Cache State Retention Fast Standby Mode Entry/Exit Fine Grained Power Gating 3D Transistors Enable Low Leakage and Low Voltage Operation Power

31 Core Power Silvermont s Wide Dynamic Range of Operation Comp - large Silvermont Comp - small Performance 3 Wide Dynamic Range of Silvermont Makes It More Efficient Than Asymmetric Cores 3 Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

32 Core Power Silvermont s Wide Dynamic Range of Operation Higher power Lower efficiency Comp - large Silvermont Long switching times Lost performance & Lower efficiency High Burst frequencies at lower power Higher efficiency longer residencies higher performance Complex switching algorithms Comp - small Low frequencies + Lower voltage Higher efficiency Performance 3 Fast, Seamless Transitions Wide Dynamic Range of Silvermont Makes It More Efficient Than Asymmetric Cores 3 Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

33 Agenda

34 Tablet: Silvermont vs. Saltwell Core Power and Performance Perf Peak-to-Peak Perf Iso Power Lower Power iso Perf Perf Peak-to-Peak Perf Iso Power Lower Power iso Perf STW 1C1T SLM 1C1T STW 1C1T SLM 1C1T STW 1C1T SLM 1C1T STW 2C4T SLM 4C4T STW 2C4T SLM 4C4T STW 2C4T SLM 4C4T peak to peak iso-power iso-perf peak to peak iso-power iso-perf Single Threaded Multi-Threaded 1 Based on the geometric mean of a variety of power and performance measurements across various benchmarks. Benchmarks included in this geomean are measurements on browsing benchmarks and workloads including SunSpider* and page load tests on Internet Explorer*, FireFox*, & Chrome*; Dhrystone*; EEMBC* workloads including CoreMark*; Android* workloads including CaffineMark*, AnTutu*, Linpack* and Quadrant* as well as measured estimates on SPECint* rate_base2000 & SPECfp* rate_base2000; on Silvermont preproduction systems compared to Atom processor Z2580. Individual results will vary. SPEC* CPU2000* is a retired benchmark. * Other names and brands may be claimed as the property of others. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

35 Core Power Not All Cores Are Created Equal DC Competition QC Competition DC Silvermont Phone Power Target Performance 3 Dual-Core Silvermont is more efficient than Dual-Core Competition 2 Efficient dual-cores outperform 2 inefficient quad-cores under power constraints 2 Based on a geometric mean of the projected power and performance of SPECint* rate_base2000 on Silvermont compared to expected configurations of main ARM*-based mobile competitors using descriptions of the architectures; assumes similar configurations. Numbers may be subject to change once verified with the actual parts. Individual results will vary. SPEC* CPU2000* is a retired benchmark; results are estimates. * Other names and brands may be claimed as the property of others. 3 Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

36 Not All Cores Are Created Equal Dual Core Silvermont vs. Quad Core Competition 2 Silvermont Performance Speedup at 1W Core Power 2.1x Dual Core Silvermont vs. Quad Core Competition 2 X Less Core Power at 1W Comp Performance 3.1x 2.8x 1.5x 1.4x 1.6x Comp1 Comp2 Comp3 Comp1 Comp2 Comp3 1.6x The Performance 2.4x Lower Power 2 Based on a geometric mean of the projected power and performance of SPECint* rate_base2000 on Silvermont compared to expected configurations of main ARM*-based mobile competitors using descriptions of the architectures; assumes similar configurations. Numbers may be subject to change once verified with the actual parts. Individual results will vary. SPEC* CPU2000* is a retired benchmark; results are estimates. * Other names and brands may be claimed as the property of others. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

37 Iso core-count comparisons Competitive View Tablets Silvermont vs Competition 2 Silvermont Performance Speedup at 1.5W Core Power Silvermont vs Competition 2 X Less Core Power at Peak Comp Performance 2.3X 1.6X 1.7X 2.3X 5.2X 5.8X 3.0X 3.6X Comp1 Comp2 Comp3 Comp4 Comp1 Comp2 Comp3 Comp4 2.0x The Performance 4.3x Lower Power 2 Based on a geometric mean of the projected power and performance of SPECint* rate_base2000 on Silvermont compared to expected configurations of main ARM*-based mobile competitors using descriptions of the architectures; assumes similar configurations. Numbers may be subject to change once verified with the actual parts. Individual results will vary. SPEC* CPU2000* is a retired benchmark; results are estimates. * Other names and brands may be claimed as the property of others. Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to:

38