Benchmarks for new implementation of kmerminhash

newminhash.jl

b1 = open(FASTAReader,
    "/Users/ksb/computation/science/genomes/kv_input/fasta/Brachybacterium_alimentarium_738_10.fna")

seq1 = dna""

for s in b1
    seq1 = seq1 * s.seq
end
length(seq1) # 4160958

b1 = open(FASTAReader,
    "/Users/ksb/computation/science/genomes/kv_input/fasta/Brachybacterium_alimentarium_738_10.fna")
seq2 = DNASequence[]

for s in b1
    push!(seq2, s.seq)
end

using BenchmarkTools
@benchmark sk1 = minhash(seq1, 12, 5000)
@benchmark sk2 = minhash(seq2, 12, 5000)
@benchmark sk3 = minhash(seq1, 12, 5000)

## New Minhash on concatenated sequences
# BenchmarkTools.Trial: 
#   memory estimate:  128.75 KiB
#   allocs estimate:  18
#   --------------
#   minimum time:     104.309 ms (0.00% GC)
#   median time:      107.608 ms (0.00% GC)
#   mean time:        107.805 ms (0.00% GC)
#   maximum time:     111.584 ms (0.00% GC)
#   --------------
#   samples:          47
#   evals/sample:     1
#   time tolerance:   5.00%
#   memory tolerance: 1.00%


## Old Minhash on concatenated sequences
# BenchmarkTools.Trial: 
#   memory estimate:  128.75 KiB
#   allocs estimate:  18
#   --------------
#   minimum time:     106.780 ms (0.00% GC)
#   median time:      111.939 ms (0.00% GC)
#   mean time:        112.147 ms (0.00% GC)
#   maximum time:     117.537 ms (0.00% GC)
#   --------------
#   samples:          45
#   evals/sample:     1
#   time tolerance:   5.00%
#   memory tolerance: 1.00%


## New Minhash on vector of sequences
# BenchmarkTools.Trial: 
#   memory estimate:  134.45 KiB
#   allocs estimate:  310
#   --------------
#   minimum time:     104.206 ms (0.00% GC)
#   median time:      107.434 ms (0.00% GC)
#   mean time:        107.680 ms (0.00% GC)
#   maximum time:     115.070 ms (0.00% GC)
#   --------------
#   samples:          47
#   evals/sample:     1
#   time tolerance:   5.00%
#   memory tolerance: 1.00%