In the book "computer architecture a quantitative approach by hennessy and patterson", it is given that write takes more time than read because tag comparison in cache happens in parallel with block reading (in case of read) and if it is a hit then memory block immediately being transferred to CPU. While this is not possible in case of write, because we can't write a block unless it is a hit, else it would overwrite the present data in cache.
So my question is "is it only a reason behind this or there are others too?"
and what is rationale behind the more power consumption in case of Read/Write?
Hi, I could suggest you the reading of the attached paper to know about R/W operaton cost on flash memories.
https://www.usenix.org/sites/default/files/fast16_full_proceedings_interior.pdf#page=133
Power consumption will be more when an operation takes more time. In this case, write takes much more time than read which means write operation will have much more power consumption. lets think like this: write is doing more work than read, so its power consumption is also becoming more than read
Simultaneous reads are possible as it not going to affect the data consistency where in writes have to happen in a linear way. This might be one of the reasons and the data travel for write is atleast twice but only once in read.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems