2 votes 2 votes Do we consider hierarchical model or simultaneous access model for write through ? CO and Architecture cache-memory co-and-architecture write-through + – pa-try asked Dec 14, 2017 pa-try 1.3k views answer comment Share Follow See all 8 Comments See all 8 8 Comments reply Show 5 previous comments pa-try commented Dec 15, 2017 reply Follow Share Fine. What would be the ans if WRITE BACK policy used with block size 4 words and write hit 0.6 ? Can you please let me tell. 0 votes 0 votes pa-try commented Dec 15, 2017 reply Follow Share Is this the case for write thruogh..?? 0 votes 0 votes Kaluti commented Sep 24, 2018 reply Follow Share Yes if hit ratio for write is one it would be simultaneous mode . If nothing is given then by default take hierarchical mode 0 votes 0 votes Please log in or register to add a comment.
4 votes 4 votes We will go with Hierarchical Model in Write-through, Tavg= Avg time for Read + Avg time for Write =0.75(45+ 0.1*750)+ 0.25(1*750) =90+ 187.5 =277.5 ns Rajnish Kumar 1 answered Dec 15, 2017 Rajnish Kumar 1 comment Share Follow See all 7 Comments See all 7 7 Comments reply pa-try commented Dec 15, 2017 reply Follow Share Thank you. What would be the ans if WRITE BACK policy used with block size 4 words and write hit 0.6 ? 0 votes 0 votes pa-try commented Dec 15, 2017 reply Follow Share I think you missed 45*0.9 in calculations. 0 votes 0 votes Rajnish Kumar 1 commented Dec 15, 2017 reply Follow Share I have written in short form only, you can see the below calculation.Hope, it will help you Avg Time for Read=0.9*45+0.1*(45+750) =0.9*45+0.1*45+0.1*750 =45(0.1+0.9)+0.1*750 =45+0.1*750 1 votes 1 votes Arpit Dhuriya commented Dec 25, 2017 reply Follow Share they have mention the hit of write though case is 1 means no page in the cache to write it back to main memory and hence we can update the cache and main memory simultaneously. 0 votes 0 votes Sona Barman commented Dec 31, 2017 reply Follow Share Many people use to solve this type of question in short cut way.The problem is that they end up in trap or raised debate.Thank you for this particular solution.Logic is clear.We can easily understand that the implementation of formula. 0 votes 0 votes hem chandra joshi commented Jan 1, 2018 reply Follow Share as we use AMAT in write through AMAT = L1 HIT TIME + L1 MISS RATE * L1 MISS PENALTY AS IN L1 HIT TIME = 45 NS AND MISS RATE = 1-HIT RATE = 1-1 =0 SO WHY WE MULTIPLY WITH 750 ? 0 votes 0 votes Kaluti commented Sep 24, 2018 reply Follow Share <p>As hit ratio is one in case of write through it means updating to main memory and cache is parallel .</p> 0 votes 0 votes Please log in or register to add a comment.
0 votes 0 votes = 0.75 * ( 0.9 * (12) + 0.1 * (750)) + 0.25 * ( 0.9 * 750 + 0.1 * 750) = 86.625 + 187.5 = 274.125 ns Yogesh Mandge answered Dec 31, 2017 Yogesh Mandge comment Share Follow See all 0 reply Please log in or register to add a comment.
0 votes 0 votes Simultaneous access is consider: T-avg=.75(.9(45)+.1(750))+.25(750) =274.125 Sona Barman answered Jan 1, 2018 Sona Barman comment Share Follow See all 0 reply Please log in or register to add a comment.