gballet
core, trie: cache the code chunking in contract
Since ChunkifyCode and touchEachChunksOnReadAndChargeGas are dominating the execution time, this PR attempts to cache the values as much as possible, in order to reduce the amount of time these functions are called.
Before: ChunkifyCode dominates execution time
Current state: execution sped up by; a factor 2, and touchEachChunksOnReadAndChargeGas now dominates
Both functions were called too often, because chunking is performed in PUSHn instructions, which are among the most common instructions found in a standard EVM bytecode.
This is still useful when calling CODECOPY, but contract execution is no longer a problem and EXTCODECOPY won't work. I think a better approach is to create a global code cache. I'm creating an issue for that.