News for hackers. Stuff that matters.
Re:Memory Bandwidth Napkin Math by crest on Monday February 10, 2020 @05:30PM (#22294116)
The CPU data caches are low latency and high throughput if you hit them. There is nothing more to it. The interesting questions are:

* What is your cache hit rate? * How much of each cache hit is used? * Are you utilizing all available resources e.g. memory channels, cache banks, vector lanes?

Re:Memory Bandwidth Napkin Math by wpietri on Monday February 10, 2020 @07:13PM (#22294806)
I'm sure you already know this, but if I understand hinkley's question correctly, there's one more thing to it. Modern processors may prefetch memory in chunks, greatly speeding up the sequential case: https://software.intel.com/en-us/articles/optimizing-applica...
Re:Memory Bandwidth Napkin Math by hinkley on Monday February 10, 2020 @05:59PM (#22294352)
I took "Random access into the cache" to mean "random access of memory locations that are already in the cache".

That would be a cache hit rate of 100%. Which would imply that pointer chasing has seen none of the improvements that branch prediction has seen over that interval (see also, 'threaded interpreters' are now considered passé).

Is there a different interpretation I was meant to take from that sentence?