2. [4 points] Discuss the pros and cons of shared vs. private L2 caches for both

Question

2. [4 points] Discuss the pros and cons of shared vs. private L2 caches for both single-threaded, multi-threaded, and multiprogrammed workloads, and reconsider them if having on-chip L3 caches2 3. 6 points] How do write buffers improve CPU performance? Are there any conditions under which write buffers can't help? Explain. 4. [6 points] (a) Explain the tradeoffs between virtually-addressed caches and physically-addressed caches. b) Why is it more difficult to have a virtually-addressed write-back cache? How can this limitation be overcome?

Explanation / Answer

Solution:

3 and 4 is done, please repost 2.

3)

Write buffers work as a medium between the cache and the main memory or the cache which is next to the primary cache in the memory hierarchy, so what they do is the hold the data which is being written on from the primary cache to next main memory or the secondary cache and this caching is called buffered write-through.

So the write buffer is free to process the cache read request when the write is taking place, that is how it increases the CPU performance.

Write buffers won't work when In multi-core systems, write buffers destroy sequential consistency. Some software disciplines, like C11's data-race-freedom

4)

a)

While indexing based on the physical address is possible with ways larger than the page size (e.g., by predicting the more significant bits or a fast translation mechanism providing those bits using the delay of indexing with the known physical address bits to hide translation latency), it is not commonly done.

Using virtual addresses for tagging allows a cache hit to be determined before translation has been done. Permissions still need to be checked before the access can be committed, but for loads, the data can be forwarded to the execution units and computation using the data begun and for stores, the data can be sent to a buffer to allow delayed commitment of state. A permission exception would flush the pipeline, so this does not add design complexity.

b)

In write-back caching the I/o is directed to the host and immediately it has to be confirmed to the host. so, if it is virtually addressed then it decreases the throughput of the system for write-intensive applications.

I hope this helps if you find any problem. Please comment below. Don't forget to give a thumbs up if you liked it. :)

Related Questions

Navigate

• Browse (All)
• Browse 2
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.