When to Run the garbage collector?

ruby-talk
1

hi,I want to use ruby's garbage collector

GC.start

but when use it,and when not use it?

···

--
Posted via http://www.ruby-forum.com/.

2

hi,I want to use ruby's garbage collector

GC.start

but when use it,and when not use it?

Garbage collection in ruby is automatic. You can turn it off
(GC.disable) or turn it on (GC.enable) or tell it to run "right now"
(GC.start). But normally you just leave it on by default and
let it collect automatically. You normally don't need to mess with
it unless you have some special situation.

Regards,

Bill

···

From: "gz zz" <gpygood@126.com>

3

Thank you.
I only see rails use the 'GC.start' in my all 'gems'.e.g.
def test_time_recognition
    n = 10000
    if RunTimeTests
      GC.start
      rectime = Benchmark.realtime do
        n.times do
          rs.recognize_path("content")
          rs.recognize_path("content/list")
          rs.recognize_path("content/show/10")
          rs.recognize_path("admin/user")
          rs.recognize_path("admin/user/list")
          rs.recognize_path("admin/user/show/10")
        end
      end
      puts "\n\nRecognition (RouteSet):"
      per_url = rectime / (n * 6)
      puts "#{per_url * 1000} ms/url"
      puts "#{1 / per_url} url/s\n\n"
    end
  end

···

--
Posted via http://www.ruby-forum.com/.

4

Thank you.
I only see rails use the 'GC.start' in my all 'gems'.e.g.
def test_time_recognition
   n = 10000
   if RunTimeTests
     GC.start
     rectime = Benchmark.realtime do
       n.times do
         rs.recognize_path("content")
         rs.recognize_path("content/list")
         rs.recognize_path("content/show/10")
         rs.recognize_path("admin/user")
         rs.recognize_path("admin/user/list")
         rs.recognize_path("admin/user/show/10")
       end
     end
     puts "\n\nRecognition (RouteSet):"
     per_url = rectime / (n * 6)
     puts "#{per_url * 1000} ms/url"
     puts "#{1 / per_url} url/s\n\n"
   end
end

It's doing GC.start there right before the Benchmark starts, just to
help achieve a more consistent starting point, for measuring the
elapsed time.

Notice it only does the GC.start when benchmarks are being run.

Regards,

Bill

···

From: "gz zz" <gpygood@126.com>

5

It may be difficult to benchmark (or profile for that matter) code under
GC: if the GC runs, you can get more than 200ms accounted to whichever
method was being run.

If you can afford the memory allocation of your benchmarked code, try
GC.start ; GC.disable before starting the bench.

···

On Thursday 28 June 2007, Bill Kelly wrote:

From: "gz zz" <gpygood@126.com>

> Thank you.
> I only see rails use the 'GC.start' in my all 'gems'.e.g.
> def test_time_recognition
> n = 10000
> if RunTimeTests
> GC.start
> rectime = Benchmark.realtime do
> n.times do
> rs.recognize_path("content")
> rs.recognize_path("content/list")
> rs.recognize_path("content/show/10")
> rs.recognize_path("admin/user")
> rs.recognize_path("admin/user/list")
> rs.recognize_path("admin/user/show/10")
> end
> end
> puts "\n\nRecognition (RouteSet):"
> per_url = rectime / (n * 6)
> puts "#{per_url * 1000} ms/url"
> puts "#{1 / per_url} url/s\n\n"
> end
> end

It's doing GC.start there right before the Benchmark starts, just to
help achieve a more consistent starting point, for measuring the
elapsed time.

Notice it only does the GC.start when benchmarks are being run.

--
Sylvain Joyeux

6

Sylvain Joyeux wrote:

It may be difficult to benchmark (or profile for that matter) code under GC: if the GC runs, you can get more than 200ms accounted to whichever method was being run.

If you can afford the memory allocation of your benchmarked code, try GC.start ; GC.disable before starting the bench.

Given that the GC runs during real world operation, and that benchmarks are generally about determining whether some code will perform acceptebley in a real situation, I'm not sure the above is great advice. Seems to me it would be better to write benchmarks that run for an amount of time that is sufficient to ensure that the overhead of GC is consistent between runs of the benchmark. For example, don't write a benchmark that takes 0.001 seconds. Write one that takes 30 seconds, or whatever. I'm not experienced enough with ruby yet, to know what a sensible duration is.

cheers,
mick

7

vice.

Seems to me it
would be better to write benchmarks that run for an amount of time that is
sufficient to ensure that the overhead of GC is consistent between runs of
the benchmark.

The problem is that, in real programs, the code snippet will not be the only
part of the code allocating objects. When the GC runs, you pay for *all* the
objects present in the system. Measuring the time the GC takes in a
controlled situation like a benchmark will not give you any information on
how much impact your code will have w.r.t. GC in a real program.

For example, don't write a benchmark that takes 0.001
seconds. Write one that takes 30 seconds, or whatever. I'm not experienced
enough with ruby yet, to know what a sensible duration is.

The problem is not the duration, but the amount of objects created.
IMO, if you're interested in benchmarking GC effects, benchmark your code in
both time and object allocation. You'll see the kind of tradeoff that one
solution has w.r.t. another. (ok: it runs 10 times faster but allocates a lot
of objects)

···

--
Sylvain