0
$\begingroup$
Average memory access time = Hit time + Miss rate * miss penalty

Assume a computer with only one cache level. What is the exact meaning of hit time? Is it the number of clock cycles to access data from cache? OR Clock cycles to execute an instruction? How does the number of clock cycle per instruction come into this equation?

$\endgroup$
  • $\begingroup$ It's usually number of clock cycles to access the data and is given in ns. $\endgroup$ – Gokul Nov 22 '18 at 11:34
1
$\begingroup$

It's back to the mechanism of cache. When the cpu wants a data in cache, try to read data from cache. If there is the data in cache, It will fetch data from cache. This time of reading data from cache (the different between the speed of cache memory and register!) will be denoted by Hit time.

If the wanted memory in the related instruction does not exist in the cache memory, it should be read from RAM, hence, it takes more cpu clock to read the specified address and this time is calculated by the multiplication of the miss rate and the penalty of this miss in time of reading data from RAM which is denoted by miss penalty.

$\endgroup$
0
$\begingroup$

First of all your equation is for the hierarchical cache where you have to search entire cache in any case irrespective of hit or miss.
For simultaneous cache the first term in right hand side will be multiplied by hit time as well. Hit time is nothing but the time taken to sense the presence of data in cache if there is.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.