Week 1 - MondayGlenn Reinman4731G Boelter [email protected] 1 - Computer Abstractions and Technology1_1IntroductionComputer Revolution Use small processors to do computations Moore's Law - doubling of transistor density More capacity to do more stuff Increase in capability makes novel applications feasible Computers in automobiles Cell phones WWW Search enginesClasses of Computers What kind of computer depends on what you're trying to use it for Types: Personal computers Designed for one individual at a time General purpose Cost and performance is directly related Server computers High capacity in resources Network based Reliable Supercomputers High-end computers for scientific and engineering calculations Highest capability Small fraction of the market (not used by everyone) Embedded computers Hidden Stringent power/performance/cost constraints Have a performance target to reach Fixed battery lifePost-PC Era Smart phones are getting increasingly more popular Smart phones != Cell phones Personal Mobile Device (PMD) Battery-operated Connects to the internet Examples: Smart phones Tablets Electronic glasses Cloud computing Warehouse Scale Computers (WSC) Software as a Service (SaaS) Portion of software run on a PMD and a portion run in the Cloud Examples: iCloud Google DriveObjectives How programs are translated into the machine language How hardware executes and interprets them Hardware/software interface Program performance and how to improve it Both software and hardwarePerformance Algorithm Number of operations executed Efficiency Programming language, compiler, architecture Number of machine instructions executed per operation Processor and memory system How fast instructions are executed I/O system (including OS) How fast I/O operations are executed8 Gr8 Ideas Design for Moore's Law Lots of silicon chips -> Lots of computation power Use abstraction to simplify design Make the common case fast Performance via parallelism Performance via pipelining Special case of parallelism Have stages in processing Performance via prediction Hard to figure out what the software does If statement Will go to then Or else Hierarchy of memories Instead of having one type of memory, we will have one large array of memory And make it look like we have multiple types of memory Dependability via redundancy Does the same functionality as something else Though wasteful, it may be reliable (in case of crashes)Below Your Program Application software Written in high-level language Humans can read it easily Machines need to interpret it System software Compiler translates the HLL code to machin code Operating system handles the service code Handling I/O Manage memory and storage Scheduling tasks & sharing resources Diagram: Three concentric circles: Innermost -> Outermost Hardware, Systems software, Applications softwareLevels of Program Code High-level language Level of abstraction closer to problem domain Provides for productivity and portability Example: swap(int v[], int k) { int temp; temp = v[l]; v[k] = v[k+1]; v[k+1] = temp; } Assembly language Textual representation of instructions Example:muli $2, $5, 4 add $2, $4, $2 lw $15, 0($2) lw $16, 4($2) lw $16, 0($2) lw $15, 4($2) jr $31 Hardware language Binary digits (bits) Encoded instructions and data) Example 00000000101000010000000000011000.........1_2Components of a Computer Same components for all kinds of computer I/O includes User-interface devices Storage devices Network adaptersiPad Capactitive multitouch LCD screen 3.8 V, 25 Watt-hour battery Computer board A5 - chipInside the System on Chip (SoC) Snapdragon 810Inside the Processor (CPU) Datapath: performs operations on data Control: sequences datapath, memory, ... Cache memory Small, fast SRAM memory for immediate access to data Close to the processor Holds variables Top of the memory hierarchy MORE ON HIERARCHY: Latency Storage Quick but small Large but slow Illusion of one large memory Examples of Processor: Cortex-A57 Cortex-A53Abstractions Instruction set architecture (ISA) Most important abstraction Language used by the compiler to speak to the processor Hardware/software interfaceTechnology Trends Electronics technology continues to evolve Increased capacity and performance Reduced costSemiconductor Techonlogy Silicon Add materials to transform properties:Conductors Insulators SwitchManufacturing Integrated Circuits (ICs) Yield: for a given wafer, how much of those dies will be functioningIntegrated Circuit Cost Cost per die = Cost per wafer / (Dies per wafer * Yield) Dies per wafer ≈ Wafer area / Die area Yield = 1 / (1 + (Defects area * Die area / 2))^2 Nonlinear relation to area and defect rate
View Full Document
Unlocking...