Data hazards in computer architecture, Computer Engineering

Data hazards - computer architecture:

A main effect of pipelining is to alter the relative timing of instructions by overlapping their execution. This introduces control hazards  and data.  Data hazards take place  when the pipeline changes the  order of read/write accesses to operands so  that  the order differs from the order seen by in sequence  executing instructions on the un pipelined machine.

Assume the pipelined execution of these instructions:

2262_Data hazards in computer architecture.png

All of the instructions after the ADD instruction use the result of the ADD instruction (in R1). Instruction ADD writes the value of R1 in the WB stage (indicated black), and the SUB instruction reads the value at the time ID stage (ID sub). This difficulty is called a data hazard. Unless safety measures are taken to prevent it, the SUB instruction will read the incorrect value and try to use it.

The AND instruction is also affected by this data hazard. The write of R1 does not finish till the end of cycle 5 (indicated black). Therefore, the AND instruction that reads the registers during cycle 4 (ID and) will receive the incorrect result.

The OR instruction can be made to operate without incurring a hazard by a easy implementation technique. This technique is used to perform register file reads in the second half of the cycle, and writes in the first half. Because  WB  for ADD  both instruction  and ID or for OR are performed in 1 cycle 5, the write to register file by ADD instruction will perform in the first half of the cycle, and the read of registers by OR will perform in the second half of the cycle. The XOR instruction operates correctly, because its register read taken place in cycle 6 after the register write by instruction ADD.

The next page described forwarding, a technique to remove the stalls for the hazard involving the SUB and AND both instructions.

We will also categorize the data hazards and consider the cases when stalls cannot be removed. We will see what compiler can do to schedule the pipeline to ignore stalls.

A hazard is build whenever there is dependence amongst instructions, and they are close sufficient that the overlap caused by pipelining would change the order of access to an operand. Our instance hazards have all been with register operands, but it is also possible to build dependence by reading and writing the similar memory location. In DLX pipeline, though, memory references are always kept in order, prevent this type of hazard from arising.

All the data hazards described here involve registers within the CPU.  By convention, the hazards are named by the ordering in the program that has to be preserved by the pipeline.

797_Data hazards in computer architecture1.png

 

Posted Date: 10/13/2012 4:04:52 AM | Location : United States







Related Discussions:- Data hazards in computer architecture, Assignment Help, Ask Question on Data hazards in computer architecture, Get Answer, Expert's Help, Data hazards in computer architecture Discussions

Write discussion on Data hazards in computer architecture
Your posts are moderated
Related Questions
What are the different methods of passing data? There are three different methods of passing data Calling by reference    Calling by value Calling by value and result

Subtraction of 01000-01001 using 2's complement method. Ans. Firstly 1's complement of 01001 is 10110 and 2's complement is 10110+ 1 =10111. Thus   01000 =  01000 - 01001

How many 128 × 8 RAM chips are required to provide a memory capacity of 2048 bytes. Ans. Available here RAM chips = 128 x 8 Required the memory capacity    = 2048 x 8 No. of chip

Problem 1 (a) Identify and briefly describe the possible roles of Codes of Ethics (b) Describe why is a code of ethics important to stakeholders. (c) Explain how should

Define Top-down parsing Techniques Top-down parsing can be observed as an attempt to find left-most derivations of an input-stream through searching for parse trees using a top

Q. Describe Independent Loops in fortran? HPF offers extra opportunities for parallel execution by employing the INDEPENDENT directive to declare the iterations of a do-loop is

The decimal equivalent of (1100) 2   is ? Ans. (1100) 2 = (12) 10

What are the benefits of using e-commerce? Basic Benefits of E-Commerce The main benefits are enhancing sales and reducing costs. The other advantages are as follows: 1.

Q. What is Enhanced IDE? The principle behind EIDE interface is the same as in IDE interface however this drive has capacities varying from 10.2GB to 20.5GB. The rotation speed

Q. Show layout 84-key Keyboards? PC 83-key was the earliest keyboard provided by IBM with its first Personal Computers (PC). This had 83 keys. Afterwards IBM added one more key