Term
| What does the memory management unit (MMU) do? |
|
Definition
| Translates the logical address to the physical address |
|
|
Term
| What is the general purpose of cache memory? |
|
Definition
| To reduce the problems of the Van Neumann bottle neck |
|
|
Term
| What type of memory is the cache made out of? |
|
Definition
|
|
Term
| Where is cache typically located and what is that called? |
|
Definition
| It is typically located on the same piece of silicon as the CPU. This is called "on-chip cache?" |
|
|
Term
| True or False: Cache is an order of magnitude faster than main memory |
|
Definition
|
|
Term
| True or False: Locality of reference is also referred to as principle of locality. |
|
Definition
|
|
Term
| What is locality of reference/principle of locality? |
|
Definition
| It is the tendency of the processor to access the same set of memory locations repetitively over a short period of time |
|
|
Term
| What are the two types of reference locality? |
|
Definition
|
|
Term
| What is temporal locality? |
|
Definition
| It refers to the reuse of specific data/resource elements within a relatively short time duration |
|
|
Term
| What is spatial locality? |
|
Definition
| It refers to the use of data elements within a relatively close storage location |
|
|
Term
| What is sequential locality? |
|
Definition
| It's a type of spatial locality which occurs when data elements are arranged and accessed linearly (like traversing a 1-dimensional array) |
|
|
Term
| What is the cache speed up ratio? |
|
Definition
| It's the ratio of the memory system's access time without cache to its access time with cache |
|
|
Term
| In the cache speed formula, what does the h stand for? |
|
Definition
| It's the hit ratio. This is the number of hits (it's in cache) verses the number of all memory references |
|
|
Term
| In the cache speed formula, what does the m stand for? |
|
Definition
| The miss ratio, which is 1-h (h is hit ratio) |
|
|
Term
| In the cache speed formula, what does tm stand for? |
|
Definition
| Access time of the main store |
|
|
Term
| In the cache speed formula, what does the tc stand for? |
|
Definition
| Access time of the cache memory |
|
|
Term
| What is the memory process in a parallel memory-cache system? |
|
Definition
| In this system a request is sent to both the cache and main memory at the same time. If there is a cache hit, the call to the main memory is terminated [this provides some speed up if there is a cache miss] |
|
|
Term
|
Definition
| It's when the cache is watching the address lines for transactions |
|
|
Term
|
Definition
| When the cache takes the data for the data lines. Allows the cache to stay updated and maintains consistency |
|
|
Term
|
Definition
| It's when data has been modified in the cache but not in main memory |
|
|
Term
|
Definition
| It's when the data has been updated in main memory but not in the cache |
|
|
Term
|
Definition
| it consists of several consecutive words that is transferred from main memory to the cache at one time |
|
|
Term
| What two things determine the size of a cache line? |
|
Definition
| The processor and the cache design |
|
|
Term
| What does it mean if a cache is fully associative? |
|
Definition
| It means that any memory line can be placed anywhere in the cache |
|
|
Term
|
Definition
| It is a truncated field that relates to the start of the address line of cache since the least significant bits aren't necessary. Produces a hit or miss quickly |
|
|
Term
| What are the 3 line-replacement strategies if the cache is full? |
|
Definition
1. Least recently used (LRU)
2. First-in first-out (FIFO)
3. Random |
|
|
Term
| What are the two types of cache misses for fully associative cache? |
|
Definition
1. Compulsory- application has just started and cache is empty
2. Capacity miss- the cache is full, have to replace existing cache |
|
|
Term
| What are the 4 things to know about fully associative cache? |
|
Definition
1. Any line in memory can be in any line of the cache
2. No cache pages
3. Best performance of any other cache
4. Complex and expensive |
|
|
Term
| What is a direct-map cache? |
|
Definition
| Where the line in memory may only be stored in the same line in the cache |
|
|
Term
| True or False: Direct Map cache is the same as 1-way set associative cache |
|
Definition
|
|
Term
| What to know about Direct Map/1-way set associative cache: |
|
Definition
1. Main memory is divided into pages that are the same size as the cache
2. May only store the specific line of memory in the same line in the cache
3. Simplest, cheapest, least flexible
4. Lower performance- can have unused cache |
|
|
Term
| True or False: Set-Associative cache is also known as X-Way set associative cache |
|
Definition
|
|
Term
| How does set associative/X way associative cache work? |
|
Definition
| The cache is divided in multiple cache ways which each way matches the cache page size). This allows for the same number from multiple pages to be stored in the cache at the same time |
|
|
Term
| What are the 3 types of cache misses for set associative cache? |
|
Definition
1. Compulsory- cache is empty/program just started
2. Capacity- the cache is full
3. Conflict- The cache isn't yet full but the line needs to be replaced |
|
|
Term
| True or False: DRAM is too slow to be used for cache |
|
Definition
|
|
Term
| True or False: SRAM is used for cache |
|
Definition
|
|
Term
|
Definition
| Only fetch a line from memory when there is a reference from the CPU |
|
|
Term
|
Definition
| When no there isn't a specific request, but it brings in lines of memory from main memory into the cache based on educated guesses |
|
|
Term
|
Definition
| This is when you have one cache for instructions and one cache for data |
|
|
Term
| When the cache has been updated, what are the two write policies for writing back to main memory? |
|
Definition
1. Write back- cache is like a buffer. When the bus is available, the information is written back to main memory (makes things faster, but more complex)
2. Write through- the cache is updated and immediately updates the main memory (slows the process down, but easier) |
|
|
Term
| What are some issues with cache and security? |
|
Definition
| Some prefetched instructions can create vulnerabilities. The cache can be cleared or invalidated, however an only be done by the operating system |
|
|
Term
| What is the logical address? |
|
Definition
| This is the address calculated by the SPU |
|
|
Term
| What is the physical address? |
|
Definition
| The actual location of data in main memory |
|
|
Term
| Main memory is divided into _________ |
|
Definition
|
|
Term
| If the logical address is in the MMU, this is called ________ |
|
Definition
|
|
Term
| If the referenced logical address is not in the MMU, this is a ______________ |
|
Definition
|
|
Term
| What 2 things does virtual memory support? |
|
Definition
1. Multitasking
2. Programs that are larger than the current main memory |
|
|
