diff --git a/Rakefile b/Rakefile
index 49733ed4af..4a4458070d 100644
--- a/Rakefile
+++ b/Rakefile
@@ -131,165 +131,4 @@ namespace :docs do
   end
 end
 
-require_relative "profile/benchmarking"
-
-# Some require data files, available from the drivers team. See the comments above each task for details."
-namespace :benchmark do
-  desc "Run the bson benchmarking tests"
-  task :bson do
-    puts "BSON BENCHMARK"
-    Mongo::Benchmarking.report({
-      bson: Mongo::Benchmarking::BSON.run_all({
-        flat: %i[ encode decode ],
-        deep: %i[ encode decode ],
-        full: %i[ encode decode ],
-      })
-    })
-  end
-
-  namespace :bson do
-    namespace :flat do
-      desc "Benchmarking for flat bson documents."
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called flat_bson.json.
-      task :encode do
-        puts "BSON BENCHMARK :: FLAT :: ENCODE"
-        Mongo::Benchmarking.report({ bson: { flat: { encode: Mongo::Benchmarking::BSON.run(:flat, :encode) } } })
-      end
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called flat_bson.json.
-      task :decode do
-        puts "BSON BENCHMARK :: FLAT :: DECODE"
-        Mongo::Benchmarking.report({ bson: { flat: { decode: Mongo::Benchmarking::BSON.run(:flat, :decode) } } })
-      end
-    end
-
-    namespace :deep do
-      desc "Benchmarking for deep bson documents."
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called deep_bson.json.
-      task :encode do
-        puts "BSON BENCHMARK :: DEEP :: ENCODE"
-        Mongo::Benchmarking.report({ bson: { deep: { encode: Mongo::Benchmarking::BSON.run(:deep, :encode) } } })
-      end
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called deep_bson.json.
-      task :decode do
-        puts "BSON BENCHMARK :: DEEP :: DECODE"
-        Mongo::Benchmarking.report({ bson: { deep: { decode: Mongo::Benchmarking::BSON.run(:deep, :decode) } } })
-      end
-    end
-
-    namespace :full do
-      desc "Benchmarking for full bson documents."
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called full_bson.json.
-      task :encode do
-        puts "BSON BENCHMARK :: FULL :: ENCODE"
-        Mongo::Benchmarking.report({ bson: { full: { encode: Mongo::Benchmarking::BSON.run(:full, :encode) } } })
-      end
-
-      # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called full_bson.json.
-      task :decode do
-        puts "BSON BENCHMARK :: FULL :: DECODE"
-        Mongo::Benchmarking.report({ bson: { full: { decode: Mongo::Benchmarking::BSON.run(:full, :decode) } } })
-      end
-    end
-  end
-
-  namespace :single_doc do
-    desc "Run the common driver single-document benchmarking tests"
-    task :command do
-      puts "SINGLE DOC BENCHMARK:: COMMAND"
-      Mongo::Benchmarking::SingleDoc.run(:command)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called TWEET.json.
-    task :find_one do
-      puts "SINGLE DOC BENCHMARK:: FIND ONE BY ID"
-      Mongo::Benchmarking::SingleDoc.run(:find_one)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called SMALL_DOC.json.
-    task :insert_one_small do
-      puts "SINGLE DOC BENCHMARK:: INSERT ONE SMALL DOCUMENT"
-      Mongo::Benchmarking::SingleDoc.run(:insert_one_small)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called LARGE_DOC.json.
-    task :insert_one_large do
-      puts "SINGLE DOC BENCHMARK:: INSERT ONE LARGE DOCUMENT"
-      Mongo::Benchmarking::SingleDoc.run(:insert_one_large)
-    end
-  end
-
-  namespace :multi_doc do
-    desc "Run the common driver multi-document benchmarking tests"
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called TWEET.json.
-    task :find_many do
-      puts "MULTI DOCUMENT BENCHMARK:: FIND MANY"
-      Mongo::Benchmarking::MultiDoc.run(:find_many)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called SMALL_DOC.json.
-    task :bulk_insert_small do
-      puts "MULTI DOCUMENT BENCHMARK:: BULK INSERT SMALL"
-      Mongo::Benchmarking::MultiDoc.run(:bulk_insert_small)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called LARGE_DOC.json.
-    task :bulk_insert_large do
-      puts "MULTI DOCUMENT BENCHMARK:: BULK INSERT LARGE"
-      Mongo::Benchmarking::MultiDoc.run(:bulk_insert_large)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called GRIDFS_LARGE.
-    task :gridfs_upload do
-      puts "MULTI DOCUMENT BENCHMARK:: GRIDFS UPLOAD"
-      Mongo::Benchmarking::MultiDoc.run(:gridfs_upload)
-    end
-
-    # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called GRIDFS_LARGE.
-    task :gridfs_download do
-      puts "MULTI DOCUMENT BENCHMARK:: GRIDFS DOWNLOAD"
-      Mongo::Benchmarking::MultiDoc.run(:gridfs_download)
-    end
-  end
-
-  namespace :parallel do
-    desc "Run the common driver paralell ETL benchmarking tests"
-
-    # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called LDJSON_MULTI,
-    # with the files used in this task.
-    task :import do
-      puts "PARALLEL ETL BENCHMARK:: IMPORT"
-      Mongo::Benchmarking::Parallel.run(:import)
-    end
-
-    # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called LDJSON_MULTI,
-    # with the files used in this task.
-    # Requirement: Another directory in "#{Mongo::Benchmarking::DATA_PATH}/LDJSON_MULTI"
-    # called 'output'.
-    task :export do
-      puts "PARALLEL ETL BENCHMARK:: EXPORT"
-      Mongo::Benchmarking::Parallel.run(:export)
-    end
-
-    # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called GRIDFS_MULTI,
-    # with the files used in this task.
-    task :gridfs_upload do
-      puts "PARALLEL ETL BENCHMARK:: GRIDFS UPLOAD"
-      Mongo::Benchmarking::Parallel.run(:gridfs_upload)
-    end
-
-    # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called GRIDFS_MULTI,
-    # with the files used in this task.
-    # Requirement: Another directory in "#{Mongo::Benchmarking::DATA_PATH}/GRIDFS_MULTI"
-    # called 'output'.
-    task :gridfs_download do
-      puts "PARALLEL ETL BENCHMARK:: GRIDFS DOWNLOAD"
-      Mongo::Benchmarking::Parallel.run(:gridfs_download)
-    end
-  end
-end
+load 'profile/benchmarking/rake/tasks.rake'
diff --git a/profile/benchmarking/bson.rb b/profile/benchmarking/bson.rb
index 88cb979ab2..415e58114e 100644
--- a/profile/benchmarking/bson.rb
+++ b/profile/benchmarking/bson.rb
@@ -71,7 +71,7 @@ def score_for(type, percentiles, scale: 10_000)
       # @return [ Hash<:timings,:percentiles,:score> ] The test results for
       #    the requested benchmark.
       def run(type, action)
-        timings = Benchmarking.without_gc { send(action, file_for(type)) }
+        timings = send(action, file_for(type))
         percentiles = Percentiles.new(timings)
         score = score_for(type, percentiles)
 
diff --git a/profile/benchmarking/helper.rb b/profile/benchmarking/helper.rb
index 856cb7659b..70d4e34312 100644
--- a/profile/benchmarking/helper.rb
+++ b/profile/benchmarking/helper.rb
@@ -58,13 +58,21 @@ def parse_json(document)
     #   iterating.
     #
     # @return [ Array<Float> ] the timings for each iteration
-    def benchmark(max_iterations: Benchmarking::TEST_REPETITIONS, min_time: 60, max_time: 5 * 60, &block)
+    def benchmark(max_iterations: Benchmarking::TEST_REPETITIONS,
+                  min_time: 60,
+                  max_time: 5 * 60,
+                  progress: default_progress_callback,
+                  &block)
+      progress ||= ->(state) {} # fallback to a no-op callback
+      progress[:start]
+
       [].tap do |results|
         iteration_count = 0
         cumulative_time = 0
 
         loop do
-          timing = Benchmark.realtime(&block)
+          timing = without_gc { Benchmark.realtime(&block) }
+          progress[:step]
 
           iteration_count += 1
           cumulative_time += timing
@@ -78,6 +86,8 @@ def benchmark(max_iterations: Benchmarking::TEST_REPETITIONS, min_time: 60, max_
           # number of iterations have been reached.
           break if cumulative_time >= min_time && iteration_count >= max_iterations
         end
+
+        progress[:end]
       end
     end
 
@@ -98,32 +108,6 @@ def report(results, indent: 0, percentiles: [ 10, 25, 50, 75, 90, 95, 98, 99 ])
       end
     end
 
-    # A utility class for returning the list item at a given percentile
-    # value.
-    class Percentiles
-      # @return [ Array<Number> ] the sorted list of numbers to consider
-      attr_reader :list
-
-      # Create a new Percentiles object that encapsulates the given list of
-      # numbers.
-      #
-      # @param [ Array<Number> ] list the list of numbers to considier
-      def initialize(list)
-        @list = list.sort
-      end
-
-      # Finds and returns the element in the list that represents the given
-      # percentile value.
-      #
-      # @param [ Number ] percentile a number in the range [1,100]
-      #
-      # @return [ Number ] the element of the list for the given percentile.
-      def [](percentile)
-        i = (list.size * percentile / 100.0).ceil - 1
-        list[i]
-      end
-    end
-
     # Get the median of values in a list.
     #
     # @example Get the median.
@@ -144,5 +128,37 @@ def without_gc
     ensure
       GC.enable
     end
+
+    private
+
+    # Returns the proc object (or nil) corresponding to the "PROGRESS"
+    # environment variable.
+    #
+    # @return [ Proc | nil ] the callback proc to use (or nil if none should
+    #   be used)
+    def default_progress_callback
+      case ENV['PROGRESS']
+      when '0', 'false', 'none'
+        nil
+      when nil, '1', 'true', 'minimal'
+        method(:minimal_progress_callback).to_proc
+      else
+        raise ArgumentError, "unsupported progress callback #{ENV['PROGRESS'].inspect}"
+      end
+    end
+
+    # A minimal progress callback implementation, printing '|' when a benchmark
+    # starts and '.' for each iteration.
+    #
+    # @param [ :start | :step | :end ] state the current progress state
+    def minimal_progress_callback(state)
+      case state
+      when :start then print '|'
+      when :step  then print '.'
+      when :end   then puts
+      end
+
+      $stdout.flush
+    end
   end
 end
diff --git a/profile/benchmarking/rake/bson.rake b/profile/benchmarking/rake/bson.rake
new file mode 100644
index 0000000000..4d5bdb1c04
--- /dev/null
+++ b/profile/benchmarking/rake/bson.rake
@@ -0,0 +1,122 @@
+# frozen_string_literal: true
+
+# rubocop:disable Layout/FirstHashElementIndentation
+
+desc 'Run the full BSON benchmarking suite'
+task :bson do
+  puts 'BSON BENCHMARK SUITE'
+  Mongo::Benchmarking.report({
+    bson: Mongo::Benchmarking::BSON.run_all(
+      flat: %i[ encode decode ],
+      deep: %i[ encode decode ],
+      full: %i[ encode decode ]
+    )
+  })
+end
+
+namespace :bson do # rubocop:disable Metrics/BlockLength
+  # a convenience task for running all of the bson benchmark tasks; this is
+  # only useful for testing that they all work.
+  task test: %w[
+    bson
+    bson:flat bson:flat:encode bson:flat:decode
+    bson:deep bson:deep:encode bson:deep:decode
+    bson:full bson:full:encode bson:full:decode
+  ]
+
+  desc 'Learn how to run the BSON benchmarks'
+  task :help do
+    puts <<~HELP
+      The BSON micro benchmarks require a set of data files that are stored in
+      the specifications repository, here:
+
+        https://github.com/mongodb/specifications/tree/master/source/benchmarking/data
+
+      Download the `extended_bson.tgz` file and extract its contents. It should
+      contain a single folder (`extended_bson`) with several files in it. Move
+      those files to:
+
+        #{Mongo::Benchmarking::DATA_PATH}
+
+      Once there, you may run any of the BSON benchmarking tasks:
+
+        $ rake benchmark:bson:flat:encode
+
+      Tasks may be run in aggregate, as well, by specifying the namespace
+      directly:
+
+        $ rake benchmark:bson:flat # runs all flat BSON benchmarks
+        $ rake benchmark:bson:deep # runs all deep BSON benchmarks
+        $ rake benchmark:bson:full # runs all full BSON benchmarks
+        # rake benchmark:bson      # runs all BSON benchmarks
+    HELP
+  end
+
+  desc 'Run the `flat` BSON benchmarking suite'
+  task :flat do
+    puts 'BSON BENCHMARK :: FLAT'
+    Mongo::Benchmarking.report({
+      bson: Mongo::Benchmarking::BSON.run_all(flat: %i[ encode decode ])
+    })
+  end
+
+  namespace :flat do
+    desc 'Run the `flat` encoding BSON benchmark'
+    task :encode do
+      puts 'BSON BENCHMARK :: FLAT :: ENCODE'
+      Mongo::Benchmarking.report({ bson: { flat: { encode: Mongo::Benchmarking::BSON.run(:flat, :encode) } } })
+    end
+
+    desc 'Run the `flat` decoding BSON benchmark'
+    task :decode do
+      puts 'BSON BENCHMARK :: FLAT :: DECODE'
+      Mongo::Benchmarking.report({ bson: { flat: { decode: Mongo::Benchmarking::BSON.run(:flat, :decode) } } })
+    end
+  end
+
+  desc 'Run the `deep` BSON benchmarking suite'
+  task :deep do
+    puts 'BSON BENCHMARK :: DEEP'
+    Mongo::Benchmarking.report({
+      bson: Mongo::Benchmarking::BSON.run_all(deep: %i[ encode decode ])
+    })
+  end
+
+  namespace :deep do
+    desc 'Run the `deep` encoding BSON benchmark'
+    task :encode do
+      puts 'BSON BENCHMARK :: DEEP :: ENCODE'
+      Mongo::Benchmarking.report({ bson: { deep: { encode: Mongo::Benchmarking::BSON.run(:deep, :encode) } } })
+    end
+
+    desc 'Run the `deep` decoding BSON benchmark'
+    task :decode do
+      puts 'BSON BENCHMARK :: DEEP :: DECODE'
+      Mongo::Benchmarking.report({ bson: { deep: { decode: Mongo::Benchmarking::BSON.run(:deep, :decode) } } })
+    end
+  end
+
+  desc 'Run the `full` BSON benchmarking suite'
+  task :full do
+    puts 'BSON BENCHMARK :: FULL'
+    Mongo::Benchmarking.report({
+      bson: Mongo::Benchmarking::BSON.run_all({ full: %i[ encode decode ] })
+    })
+  end
+
+  namespace :full do
+    desc 'Run the `full` encoding BSON benchmark'
+    task :encode do
+      puts 'BSON BENCHMARK :: FULL :: ENCODE'
+      Mongo::Benchmarking.report({ bson: { full: { encode: Mongo::Benchmarking::BSON.run(:full, :encode) } } })
+    end
+
+    desc 'Run the `full` decoding BSON benchmark'
+    task :decode do
+      puts 'BSON BENCHMARK :: FULL :: DECODE'
+      Mongo::Benchmarking.report({ bson: { full: { decode: Mongo::Benchmarking::BSON.run(:full, :decode) } } })
+    end
+  end
+end
+
+# rubocop:enable Layout/FirstHashElementIndentation
diff --git a/profile/benchmarking/rake/multi_doc.rake b/profile/benchmarking/rake/multi_doc.rake
new file mode 100644
index 0000000000..86c190ef1f
--- /dev/null
+++ b/profile/benchmarking/rake/multi_doc.rake
@@ -0,0 +1,34 @@
+# frozen_string_literal: true
+# rubocop:todo all
+
+namespace :multi_doc do
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called TWEET.json.
+  task :find_many do
+    puts 'MULTI DOCUMENT BENCHMARK :: FIND MANY'
+    Mongo::Benchmarking::MultiDoc.run(:find_many)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called SMALL_DOC.json.
+  task :bulk_insert_small do
+    puts 'MULTI DOCUMENT BENCHMARK :: BULK INSERT SMALL'
+    Mongo::Benchmarking::MultiDoc.run(:bulk_insert_small)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called LARGE_DOC.json.
+  task :bulk_insert_large do
+    puts 'MULTI DOCUMENT BENCHMARK :: BULK INSERT LARGE'
+    Mongo::Benchmarking::MultiDoc.run(:bulk_insert_large)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called GRIDFS_LARGE.
+  task :gridfs_upload do
+    puts 'MULTI DOCUMENT BENCHMARK :: GRIDFS UPLOAD'
+    Mongo::Benchmarking::MultiDoc.run(:gridfs_upload)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called GRIDFS_LARGE.
+  task :gridfs_download do
+    puts 'MULTI DOCUMENT BENCHMARK :: GRIDFS DOWNLOAD'
+    Mongo::Benchmarking::MultiDoc.run(:gridfs_download)
+  end
+end
diff --git a/profile/benchmarking/rake/parallel.rake b/profile/benchmarking/rake/parallel.rake
new file mode 100644
index 0000000000..98752e231e
--- /dev/null
+++ b/profile/benchmarking/rake/parallel.rake
@@ -0,0 +1,36 @@
+# frozen_string_literal: true
+# rubocop:todo all
+
+namespace :parallel do
+  # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called LDJSON_MULTI,
+  # with the files used in this task.
+  task :import do
+    puts 'PARALLEL ETL BENCHMARK :: IMPORT'
+    Mongo::Benchmarking::Parallel.run(:import)
+  end
+
+  # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called LDJSON_MULTI,
+  # with the files used in this task.
+  # Requirement: Another directory in '#{Mongo::Benchmarking::DATA_PATH}/LDJSON_MULTI'
+  # called 'output'.
+  task :export do
+    puts 'PARALLEL ETL BENCHMARK :: EXPORT'
+    Mongo::Benchmarking::Parallel.run(:export)
+  end
+
+  # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called GRIDFS_MULTI,
+  # with the files used in this task.
+  task :gridfs_upload do
+    puts 'PARALLEL ETL BENCHMARK :: GRIDFS UPLOAD'
+    Mongo::Benchmarking::Parallel.run(:gridfs_upload)
+  end
+
+  # Requirement: A directory in Mongo::Benchmarking::DATA_PATH, called GRIDFS_MULTI,
+  # with the files used in this task.
+  # Requirement: Another directory in '#{Mongo::Benchmarking::DATA_PATH}/GRIDFS_MULTI'
+  # called 'output'.
+  task :gridfs_download do
+    puts 'PARALLEL ETL BENCHMARK :: GRIDFS DOWNLOAD'
+    Mongo::Benchmarking::Parallel.run(:gridfs_download)
+  end
+end
diff --git a/profile/benchmarking/rake/single_doc.rake b/profile/benchmarking/rake/single_doc.rake
new file mode 100644
index 0000000000..803e28d593
--- /dev/null
+++ b/profile/benchmarking/rake/single_doc.rake
@@ -0,0 +1,27 @@
+# frozen_string_literal: true
+# rubocop:todo all
+
+namespace :single_doc do
+  task :command do
+    puts 'SINGLE DOC BENCHMARK :: COMMAND'
+    Mongo::Benchmarking::SingleDoc.run(:command)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called TWEET.json.
+  task :find_one do
+    puts 'SINGLE DOC BENCHMARK :: FIND ONE BY ID'
+    Mongo::Benchmarking::SingleDoc.run(:find_one)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called SMALL_DOC.json.
+  task :insert_one_small do
+    puts 'SINGLE DOC BENCHMARK :: INSERT ONE SMALL DOCUMENT'
+    Mongo::Benchmarking::SingleDoc.run(:insert_one_small)
+  end
+
+  # Requirement: A file in Mongo::Benchmarking::DATA_PATH, called LARGE_DOC.json.
+  task :insert_one_large do
+    puts 'SINGLE DOC BENCHMARK :: INSERT ONE LARGE DOCUMENT'
+    Mongo::Benchmarking::SingleDoc.run(:insert_one_large)
+  end
+end
diff --git a/profile/benchmarking/rake/tasks.rake b/profile/benchmarking/rake/tasks.rake
new file mode 100644
index 0000000000..7feae8d2ab
--- /dev/null
+++ b/profile/benchmarking/rake/tasks.rake
@@ -0,0 +1,11 @@
+# frozen_string_literal: true
+
+require_relative '../../benchmarking'
+
+# Some require data files, available from the drivers team.
+# See the comments above each task for details.
+namespace :benchmark do
+  %w[ bson single_doc multi_doc parallel ].each do |group|
+    load File.join(__dir__, "#{group}.rake")
+  end
+end