testing: benchmark assemble / transaction pool's uses of recomputeBlockEvaluator

[cce]

Summary

First draft at attempting to profile CPU and benchmark calling recomputeBlockEvaluator on large pool & block sizes to examine performance bottlenecks updating the transaction pool when a new block arrives.

I was running with GOGC=off to try and declutter the profiles, especially when b.N > 1.

Test Plan

Adds a test.

[codecov-commenter]

Codecov Report

Merging #3138 (2f79d39) into master (e83acec) will decrease coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #3138      +/-   ##
==========================================
- Coverage   55.25%   55.23%   -0.02%     
==========================================
  Files         398      398              
  Lines       50297    50297              
==========================================
- Hits        27793    27784       -9     
- Misses      20171    20180       +9     
  Partials     2333     2333              

Impacted Files	Coverage Δ
util/db/dbutil.go	44.24% <0.00%> (-4.25%)	⬇️
ledger/blockqueue.go	85.63% <0.00%> (-2.88%)	⬇️
catchup/peerSelector.go	98.95% <0.00%> (-1.05%)	⬇️
cmd/tealdbg/debugger.go	72.69% <0.00%> (-0.81%)	⬇️
network/wsPeer.go	67.94% <0.00%> (-0.55%)	⬇️
network/wsNetwork.go	64.63% <0.00%> (-0.20%)	⬇️
catchup/service.go	69.38% <0.00%> (+0.49%)	⬆️
data/transactions/verify/txn.go	44.64% <0.00%> (+0.89%)	⬆️
crypto/merkletrie/node.go	93.48% <0.00%> (+1.86%)	⬆️
crypto/merkletrie/trie.go	68.61% <0.00%> (+2.18%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[brianolson]

a little dead code to cleanup, otherwise looks reasonable

[algonautshant]

Looks good.