Quick Benchmark to check what kind of array 'slicing' performs well

Array with instances benchmark

require 'benchmark'

class User
  def initialize
    @instance_var = rand(100)
  end
end

users = []
10000.times do |i|
  users.push User.new
end
SPLIT_AT = users.size / 2

# Using bmbm to account for GC
# http://www.ruby-doc.org/stdlib-1.9.3/libdoc/benchmark/rdoc/Benchmark.html

Benchmark.bmbm do |b|
  b.report('next: ') do
    users.each_with_index do |user, index|
      next if users.index(user) > SPLIT_AT
      local_var = user
    end
  end

  b.report('reject: ') do
    users.reject{|i| users.index(i) > SPLIT_AT}.each do |user|
      local_var = user
    end
  end

  b.report('range: ') do
    users[0..SPLIT_AT].each do |user|
      local_var = user
    end
  end

  b.report('first(n): ') do
    users.first(SPLIT_AT).each do |user|
      local_var = user
    end
  end

  b.report('shift(n): ') do
    users.shift(SPLIT_AT).each do |user|
      local_var = user
    end
  end
end

Array with instances result

Rehearsal ----------------------------------------------
next:        3.520000   0.070000   3.590000 (  3.839394)
reject:      3.560000   0.090000   3.650000 (  3.812247)
range:       0.000000   0.000000   0.000000 (  0.000587)
first(n):    0.000000   0.000000   0.000000 (  0.000601)
shift(n):    0.000000   0.000000   0.000000 (  0.000541)
------------------------------------- total: 7.240000sec

                 user     system      total        real
next:        0.880000   0.010000   0.890000 (  0.914190)
reject:      0.920000   0.010000   0.930000 (  0.945588)
range:       0.000000   0.000000   0.000000 (  0.000700)
first(n):    0.000000   0.000000   0.000000 (  0.000709)
shift(n):    0.000000   0.000000   0.000000 (  0.000683)

Array with integers benchmark

require 'benchmark'

ARRAY = 1.upto(1000000).to_a
SPLIT_AT = 500000

# Using bmbm to account for GC

Benchmark.bmbm do |b|
  b.report('next: ') do
    ARRAY.each do |i|
      next if i > SPLIT_AT
      local_var = i
    end
  end

  b.report('reject: ') do
    ARRAY.reject{|i| i > SPLIT_AT}.each do |i|
      local_var = i
    end
  end

  b.report('range: ') do
    ARRAY[0..SPLIT_AT].each do |i|
      local_var = i
    end
  end

  b.report('first(n): ') do
    ARRAY.first(SPLIT_AT).each do |i|
      local_var = i
    end
  end

  b.report('shift(n): ') do
    ARRAY.shift(SPLIT_AT).each do |i|
      local_var = i
    end
  end
end

Array with integers result

Rehearsal ----------------------------------------------
next:        0.110000   0.010000   0.120000 (  0.120553)
reject:      0.150000   0.010000   0.160000 (  0.169044)
range:       0.060000   0.000000   0.060000 (  0.067574)
first(n):    0.050000   0.000000   0.050000 (  0.053314)
shift(n):    0.050000   0.000000   0.050000 (  0.055513)
------------------------------------- total: 0.440000sec

                 user     system      total        real
next:        0.060000   0.000000   0.060000 (  0.063703)
reject:      0.050000   0.000000   0.050000 (  0.058155)
range:       0.050000   0.000000   0.050000 (  0.054786)
first(n):    0.050000   0.000000   0.050000 (  0.057561)
shift(n):    0.060000   0.010000   0.070000 (  0.058572)