Read after write and write after write - data hazards, Computer Engineering

RAW and WAW - Data hazards:

RAW (read after write) - j tries to read a source before i writes it, hence j wrongly gets the old value .This is the most usual type of hazard and the kind that we use forwarding to solve it.

WAW (write after write) - j tries to write an operand before i. The writes end up being performed in the incorrect order, leaving the value written by  i rather than the value written by  j in the destination. This hazard is there only in pipelines that write in more than 1 pipe stage or permit an instruction to proceed even when a preceding instruction is stalled. DLX integer pipeline writes a register just in WB and ignores this class of hazards.

WAW hazards would be possible if we perform the following 2 changes to the DLX pipeline:1)move write back for an ALU operation into the MEM stage, while the data value is available by then. Assume that the data memory access took 2 pipe stages.

Here is a sequence of 2 instructions which showing the execution in this revised pipeline, importance the pipe stage that writes the consequence:

1753_Read after write and write after write - data hazards.png

 Unless this type of hazard is ignored, execution of this sequence on this revised pipeline will leave the result of the first write (the LW) in R1, before the result of the instruction ADD.

Let writes in different pipe stages introduces other problems, since 2 instructions may try to write at the similar clock cycle. DLX FP pipeline, which has both writes in different pipeline lengths and different stages and will deal with both write conflicts and WAW hazards in detail.

Posted Date: 10/13/2012 4:07:06 AM | Location : United States







Related Discussions:- Read after write and write after write - data hazards, Assignment Help, Ask Question on Read after write and write after write - data hazards, Get Answer, Expert's Help, Read after write and write after write - data hazards Discussions

Write discussion on Read after write and write after write - data hazards
Your posts are moderated
Related Questions
What are the Main aspects available in all word processors -  Ability to set page size and page orientation -  Ability to change font size and font style (for example Arial,

What is the purpose of DOTNET? Dot Net Is a collection of products like C#, ASp.Net, Vb.Net, XML. The purpose of .Net is that we can simply upgrade and degrade our programs whi

Describe the various characteristics of UDP protocol. The characteristics of the UDP are as follows: End to end: UDP is transport protocols that can distinguish between

What is multiplicity? Multiplicity is applied to attributes for data base application. Multiplicity for an attribute specifies the number of possible value for every instantiat

program in c for scanline seed fill

Backward Chaining - Artificial intelligence Given that we are just interested in constructing the path, we may set our initial state to be the theorem statement and search backw

Advantage to depth first search: It just looks like it will be a long period it finds 'DAN' until. This highlights an important drawback for depth first search. It can regular

Explain the E-cheques verses Credit Cards in brief. E-cheques: E-cheques are utilized for business dealing into e-commerce. Transactions of such cheques take place onto Inter

MIPS - computer architecture: The MIPS ISA, so far 3 instruction formats Fixed 32-bit instruction 3-operand, load-store architecture 32 general-purpose register

The structure of Logical Databases reflects which dependencies of hierarchical tables in the SAP system. Foreign Key dependencies