retagged by
893 views
1 1 vote

@ Arjun Sir, @ Habib Sir

https://gateoverflow.in/13484/coa 

I got a slightly different answer(69.12 ns) by using the concepts given by one of the professor and now I am in great dilemma when it comes to write access problems particularly in write through policy. But using that  GATE '14 and '15 were solved accurately.

There he told that write miss penalty is always going to be same(as we have to every time update the memory word) ( i.e memory time if considering only one cache) and when he calculated the average access time he even included the write access time of cache also .

For example= 

one of the formula given is 

AMAT= 1 + %read*(1-h)*(M) + % write*(M)  ( 1 is the access time of cache in clock cycles )

actually it is  AMAT = 1 + avg. stalls per access.

whereas formula used here is --->  AMAT=  %reads*(1+(1-h)*M) + %writes*(M)= "%reads*1" + %reads*(1-h)*M + %writes*M 

  so if we compare then  only that "quoted"  thing in the formula is different and  b'coz of that I got slightly greater value.

And this was given under the concept Write Through , No write Allocate , No Write Buffer... Please refer it and tell me whether any thing wrong in this ... I am not able to upload the photo . I tried many times..Waiting for your response Sir. Please solve it. Many sleepless nights because of this . Thank You.. 

Please log in or register to answer this question.

Position:
Show:

Related questions

0 0 votes
1 1 answer
1.2k
1.2k views
sushmita asked Dec 12, 2018
1,221 views
In this question can someone plz explain to me the write back part? Why are we taking write back time only for cache misses? Why not for hits? How to know when a block i...
4 4 votes
1 1 answer
1.8k
1.8k views
rahul sharma 5 asked Jan 3, 2017
1,785 views
Please some one share good resource to read Simultaneous and heirarchial memory access with write back and write through cache.There are already some posts asking the sam...
0 0 votes
0 0 answers
1.6k
1.6k views
Raj Singh 1 asked Dec 19, 2018
1,610 views
$h_1→L1$ hit ratio $h_2→L2$ hit ratio$C_1→ L1$ access time$C_2→ $Miss penalty to transfer information from L2 to L1$M→$ Miss penalty to transfer information from main mem...
1 1 vote
1 1 answer
1.0k
1.0k views
dragonball asked Jan 1, 2018
1,030 views
Compute the average access time for a machine with 80%cache hit ratio.The cache access time and memory access time are 20 ns and 200 ns .DOUBT:: While solving these type ...