- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
What is the feature of "dual-sectored line" in the Intel Xeron processor L1/L2/L3 caches? Is there any diagram?
Thanks.
Link Copied
3 Replies
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
You may be referring to the default setting of "Adjacent Sector Prefetch", which brings pairs of 64-byte cache lines into cache on read. When you have a read in either the odd or even numbered cache line, prefetch of the other line of the pair is triggered.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
This is my calculation, L1 size=8192bytes, the cacheline size is 64k, so there are 128 cachelines with 4-way associatives in L1, i.e. 32 sets, why there are 64 entries in Data TLB instead of 32 entries? is this why was called "dual sectored line" for prefetch to gain the performance?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
sorry, correction the cacheline size is 64 bytes, not 64k.
Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page