Write Back Cache Example - Georgia Tech HPCA Part 3
ฝัง
- เผยแพร่เมื่อ 22 ก.พ. 2015
- Watch on Udacity: www.udacity.com/course/viewer...
Check out the full High Performance Computer Architecture course for free at: www.udacity.com/course/ud007
Georgia Tech online Master's program: www.udacity.com/georgia-tech
Cleared all my questions up in 3:19 minutes. Thank you!
Man, I love the straightforward explanation. I wish you covered more cases, but I love it. If anyone else has any more video recommendations to find further cases with a writeback cache, let me know.
cleared up all my questions, and a GREAT explanation of dirty bits. Thank you.
Really good explanation, thanks!
Thanks a lot, Sir! You saved hours of my time!
awesome explanation
excellent explanation
Sir how did you make these toutrials. i mean which softwere did you use?
One question, would not the dirty bit mostly be set to one all the time in a computer? Because it is modified a lot. Or am I wrong? Hehe.
Thanks so much!
Why replace entries before the cache is full?
Cache explained in the video is using a map data structure. In this case, you calculate hash of the key. Assume the hash (after applying some more mathematical operations on it) gives you 1, you need to store the entry at index 1 (map is actually backed by an array). So if two keys give you the same hash and thus the same index, you need to replace the existing entry. Now when you need the value corresponding to a particular key, you just calculate the hash of that key and look for the value in that location. Now tag helps you to confirm if the index (corresponding to the calculated hash) indeed has the value you are looking for.
It depends on replacement policy, most time, we will use LRU, which remove the oldest entry, en.wikipedia.org/wiki/Cache_replacement_policies
italiano?
dirty bit martini...