CO-Write Through

Question

 my doubt

1) default is hierarchal or simultaneous.

2)will we use hierarchical or simultaneous here..also explain.

Comments

Is the answer 277.5ns?

---

i am also getting 277.5

but answer is given as 274.125

---

For write through we take the effective access time for write operation as the memory access time only.. how have they done?

---

they have considered simultaneous arrangement...

---

https://gateoverflow.in/3653/gate2004-it-12-isro2016-77

Check here..Arjun Sir has asked to use hierarchical access as the default case.

---

i have read that link... but in the video solution  the instructor is using simultaneous ...totally confused.

---

But I think what we are using is the case of simultaneous only..I just read from a link in which it said that for write through, cache and memory are updated simultaneously so their access is simultaneous. Hence max(cache time, memory time) is considered which is certainly memory time.

If we would have added the cache time then answer would be greater than 277.5ns..can you share the solution?

---

I think until unless they don't mention in the question about the hierarchal access we don't consider it. In terms of hierarchal access following words may be used:- "on the miss in L_i cache the data is transferred from L_i+1 to L_i"

I did:- $0.75 *(0.9*45+0.1*750) + 0.25*750 = 274.125$

---

Okay so you considered simultaneous access in case of read operation as well?

---

I have considered simultaneous access for "read operation" only because for write you have to employ write-through protocol.

Answer 1

Memory request for read $=0.75$

Memory request for write $=0.25$

Now, it is given for read request $90$% is hit and write through protocol is used.

In write through protocol, cache and main memory updated simultaneously.

 

• So, for Read time required $=0.75(0.9*45+0.1*750)=86.625ns.$

 

• for Write time required (as hit ratio of write is $1$)$=0.25*750=187.5$

Total time required $=86.625+187.5=274.125ns.$