This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Some basic principles (cache line, cache misses, set associativity) about the cache of C64x+

1. What is the Pros and Cons of multiple-way set associative cache?

My understanding after reading the TI's documents is follows:

Pros: Reduce the conflict cache misses because the most-recently-used data has less possibility to be evicted.

Cons: Some overhead due to the conditioning of the multiple ways and setting the LRU bit.

So if the cache misses is too much, we tend to use more-way set associative caches, e.g., L2 has 4-way set associative caches. L1D use 2-way set-associative caches.

Please correct my understanding.

 

2. What is the cache misses for L2?

3. Why L1D (64 bytes) has smaller cache line than L2's (128 bytes)?

My understanding: Smaller cache line is fast in fetching new data but suffers from more conflict misses. Please correct my understanding.