Use GetAncestor to compute new target #5948

pull sipa wants to merge 1 commits into bitcoin:master from sipa:fastpow changing 1 files +3 −3

sipa commented at 6:15 PM on March 26, 2015: member

I don't expect a measurable performance improvement from this, but result seems more readable too.
Trivial optimization: use GetAncestor to compute new target 1cc0e96e9c

in src/pow.cpp:None in 1cc0e96e9c

  41 | @@ -42,9 +42,9 @@ unsigned int GetNextWorkRequired(const CBlockIndex* pindexLast, const CBlockHead
  42 |      }
  43 |  
  44 |      // Go back by what we want to be 14 days worth of blocks
  45 | -    const CBlockIndex* pindexFirst = pindexLast;
  46 | -    for (int i = 0; pindexFirst && i < params.DifficultyAdjustmentInterval()-1; i++)
  47 | -        pindexFirst = pindexFirst->pprev;
  48 | +    int nHeightFirst = pindexLast->nHeight - (params.DifficultyAdjustmentInterval()-1);
  49 | +    assert(nHeightFirst >= 0);

laanwj commented at 7:57 AM on March 27, 2015:

Looks good to me. This assert could even be left out as GetAncestor already returns NULL for heights <0.

sipa commented at 6:14 PM on March 27, 2015:

To mimick the exact old code, it would need to be nHeightFirst = std::max(0, pindex->nHeight - (... - 1));. I added the assert to be sure for now that dropping the std::max is safe (which it should be afaict).

laanwj commented at 11:21 AM on April 1, 2015:

OK, fair enough

laanwj added the label Improvement on Mar 27, 2015
laanwj commented at 11:23 AM on April 1, 2015: member

Going to test this on my test node.
laanwj commented at 10:50 AM on April 6, 2015: member

Tested ACK
sipa merged this on Apr 7, 2015
sipa closed this on Apr 7, 2015
sipa referenced this in commit b36f1ce87e on Apr 7, 2015
MarcoFalke locked this on Sep 8, 2021

Contributors

sipa

laanwj

Labels

Refactoring