caching - Cache-oblivious data structures and dynamic languages - effective? -
I have recently read about cache-forgotten data structures such as an assist buffer pile, in this data structure cache memory Most recently work by keeping access elements, so any subsequent access becomes faster.
More of these data structures are implemented with low-level language like C-C ++. Are these data structures trying to keep the dynamic language on a port like Python, or the upper part of the running on a virtual machine is ruined by all the performance benefits of these data structures? It seems that the latter, but I thought I would ask to ask if it really has some experience in it. You have better control over the exact size of each data structure in C (or C ++);
There is also the possibility of proper control over the storage allocation.
In most dynamic languages (such as Python) you have no control over the exact size, there is very little space
You may have some temporary areas in Python, but You do not have less or no spatial location.
The transient terrain can be extended through the simple memomanization of the result. This is a common speed-up, which is often included in the memoizing decorator, so that the memorandum of the core algorithm (temporary area) is unacquainted.
I do not think translating dynamic languages to implement C or C ++ cache-bets because I do not think you have enough control, instead of just taking advantage of the memoisation for speed-up .
Comments
Post a Comment