Computer Architecture, Lecture 1

What is CISC?

- Dominant style through the mid 80's. - Included instructions to perform "typical" programming tasks. -Uses a stack to pass arguments and save the program counter. -Explicit push and pop instructions. -Arithmetic instructions may access memory.

What is a Bus?

A bus is a collection of parallel wires that carry address, data, and a control signals. Buses are typically shared by multiple devices.

What about nowadays?

Code compatibility is often more important that run speed, since we can often compensate with hardware for deficiencies in ISA design.

What is Computer Hardware?

Computer hardware refers to the specifics of a machine, including the detailed logic design and the packaging technology of the machine.

What do CISC proponents argue?

Easier for the compiler, fewer code bytes.

IA-32 instruction format

IA-32 instruction (length from 1 to 17 bytes) consist of: -optional instruction prefixes (in any order). -1-3 opcode bytes - determines the action of the statement. -an addressing-form specifier (if required) consisting of: 1) the modR/M byte - addressing modes register/memory. 2)sometimes the SIB (Scale-Index-Base) byte. 3) a displacement (if required). -an immediate data field (if required).

What is Instruction set architecture (ISA)?

ISA is an abstract interface between the hardware and the lowest level of software.

What is the ISA made up of?

ISA is made up of the programmer visible machine interface such as the instruction set, all registers, memory organization, I/O, and exception handling. It contains all information necessary to write a machine language program.

How does Memory Write work?

Memory write happens during the execution of a store instruction. 1. The processor calculates the address from parameters in the store instruction and places the address on the bus. 2. Main memory reads it and waits for the corresponding data word to arrive. 3. The CPU places the data from the register specified by the store instruction on the bus. 4. Memory reads the data from the bus and then stores it at the given address.

What is the focus of embedded computing?

Often driven by price and power demands. Code size is important as well, the smaller the better.

How do RISC instructions apply to data in registers?

RISC instructions apply to data in registers and typically change the entire register.

What is the focus of personal computing?

The focus is on optimizing cost-performance as measured by a single user. We want general purpose with a variety of software.

Sample Instruction

addl %edx, 2048888(%ebp, %ebx, 8) adds a content of register %edx and a content of memory location with address calculated as: [%ebp]+[%ebx]*8+204888, and stores the result in the same memory location.

What three aspects of computer design does "computer architecture" cover?

-Instruction Set Architecture -Computer organization -Computer Hardware

What is RISC?

-Internal project at IBM -Fewer and simpler instructions, fewer addressing modes: -might take more instructions to get a task done -they can be executed quickly due to the specific hardware. -Register oriented instruction set -more than the typical 32 registers. -used for arguments pointers and temporaries. -only a few load and store instructions can access memory. -simply memory addressing -all instructions are one size and few instruction formats -three address instructions

What three properties lead to dramatic simplifications in the implementation of advanced design techniques?

1) A smaller number of memory addressing modes. 2) The instruction formats are few in number with all instructions typically being one size. 3) Large number of registers.

What are the three cache levels in the Intel Core i7 Cache Hierarchy?

1. L1 cache, size: 32kb, Access: 4 cycles. 2. L2 cache: size:256kb, access: 11 cycles. 3. L3 cache: size: 8MB, access: 30-40 cycles.

What are the 4 major classes of computers?

1. Personal computers. 2. Server computers. 3. Embedded computers. 4. Supercomputers.

What are the two focal points of the PostPC era?

1. Personal mobile devices: Battery operated, connected to the internet, relatively cheap, everything from smartphones to tablets. 2. Cloud Computing: Warehouse scale computers, portion of software runs on the personal mobile device and a portion runs in the cloud.

What are the two main approaches to designing ISA

1. RISC (reduced instruction set computer) architecture. 2. CISC (complex instruction set computer) architecture.

How does reading a Disc Sector work?

1. The CPU initiates a disk read by writing a read command, logical sector number, and destination memory address to an address associated with the disk controller. 2. Then the Disk controller reads the sector and performs a direct memory access (DMA) transfer into main memory. 3. When the DMA transfer completes the disk controller notifies the CPU with an interrupt.

How does Instruction Fetch work (4 steps)?

1. The Cpu places an address from the program counter on the memroy bus. 2. Main memory reads the address from the memory bus, retrieves the instruction at the given address and then places it on the bus. 3. The processor reads the given instruction from the bus and then copies it in the instruction register. 4. Main memory is accessed again when the load instruction is executed with the following difference: The address is calculated by the processor, and the content from memory is loaded into one of the registers.

What are the 3 limits in increasing computer performance?

1. The memory wall. 2. The Instruction Level Parallelism wall. 3. The power wall.

What are the four goals of computer architects?

1. To meet functional requirements. 2. To meet cost goals. 3. To meet performance goals. 4. To meet power goals.

What do RISC proponents argue?

Better for optimizing compilers, you can make the code run fist with simple chip design.

What is the focus of supercomputing?

High-end science and engineering calculations. These are the pinnacle of computing performance but only take up a small percentage of the computing market.

What is the focus of server computing?

The focus is on network base, high capacity, performance and reliability. Size is not as much an issue as performance is.

What is the Instruction level Parallelism wall?

The increasing difficulty of finding enough parallelism in a single instructions stream to keep a high-performance processor busy.

What is the Memory Wall?

The increasing gap between processor and memory speeds. This effect pushes cache sizes larger in order to mask the latency of memory.

Which instructions affect memory?

The only instructions that affect memory are load and store instructions that move data from memory to a register or to memory from a register, respectively.

What is the power wall?

The trend of consuming exponentially more power with each factorial increase of operating frequency. The power wall poses manufacturing, system design and deployment problems that have not been justified in the face of diminished gains in performance due to the memory wall and the ILP wall.

What is computer organization?

These are the high-level aspects of design, including the design of the CPU, and the design of the memory systems and bus structure.

How is the IA-32 memory address calculated?

[Rb] + S*[Ri] + D, where: D --> Constant displacement using 1, 2, or 4 bytes. Rb --> base register: any of 8 integer registers. Ri --> index register: any of 8 integer registers, except %esp S --> scale 1, 2, 4, or 8 bytes.