Then it would still need to fetch the memory from the RAM, the whole point of using caches instead of RAM is access speed. Besides we already have the adress of the RAM block we want to reach, so why would we need a pointer .
This definitely deserves some explanation as it is a technical detail of CPU caches. The last three bits (000-111) are used as an index into the cached data. The first 29 bits are commonly referred to as the 'tag.' Here is how I imagine it working: CPU: Give me address [0...tag_bits...28]xxx CACHE: Let's see if I have data for the 'tag' [0...28] if (TAG_FOUND) : CACHE: Great, is it valid data? if (CAHCE_LINE_VALID) : CACHE: Amaze balls. Read the byte at offset xxx (this is the index mentioned above) The professor probably glossed over these details intentionally to avoid leaking too much about cache mapping strategies into this discussion, which is understandable. Still, almost every book I've read about the subject does this and as a beginner, it is so frustrating to not know why as it's being explained.
"That block is no longer a virgin, and is now dirty so to speak"- lmao
Damn this gentleman is a wizard indeed.
These videos are fantastic. Thank you!
Thanks, this was exactly what I was interested in
great content, thank you for this!
thank you harry putter
many thanks for u please can give me this is material(slides)
Why don't cache memories have pointers to the main memory instead of data?
Then it would still need to fetch the memory from the RAM, the whole point of using caches instead of RAM is access speed. Besides we already have the adress of the RAM block we want to reach, so why would we need a pointer .
the harry potter name will certainly bring more views :). I wish my teacher explained like you
please explain. the last 3 bits of an address. 1 byte is 8bits. how does 3 bits represent 8 bytes?
2^3 = 8
This definitely deserves some explanation as it is a technical detail of CPU caches. The last three bits (000-111) are used as an index into the cached data. The first 29 bits are commonly referred to as the 'tag.'
Here is how I imagine it working:
CPU: Give me address [0...tag_bits...28]xxx
CACHE: Let's see if I have data for the 'tag' [0...28]
if (TAG_FOUND) : CACHE: Great, is it valid data?
if (CAHCE_LINE_VALID) : CACHE: Amaze balls. Read the byte at offset xxx (this is the index mentioned above)
The professor probably glossed over these details intentionally to avoid leaking too much about cache mapping strategies into this discussion, which is understandable. Still, almost every book I've read about the subject does this and as a beginner, it is so frustrating to not know why as it's being explained.
The point is, to store 8 bytes you need 64 bits, but to address 8 bytes you only need to count them, so 3 bits are enough.
People who disliked this video must be some jealous lecturers😂
1 byte is 8bits. why did you say 8bytes are stored in last 3 bits 000? :(
"8 bytes are stored in last 3 bits" means - 3 bits can be used to point to 8 bytes
000 --> First Byte
001 --> Second Byte
.
.
.
111 --> Eighth Byte
Just realized his name is Harry Porter..
bouta go take a bong rip brb
harry potter
the man is giving very high quality explanation yet you insult hum
this world is a joke
@@songoku6425 why you think it a joke ? you think harry potter is a joke ?