Issue #17031 has been updated by marcandre (Marc-Andre Lafortune).
Sure
```
Calculating -------------------------------------
short_backtrace 28.315k (=B1 7.1%) i/s - 141.984k in 5.044733s
long_backtrace 24.168k (=B1 8.7%) i/s - 120.900k in 5.050243s
no_caller_locations 29.288k (=B1 2.5%) i/s - 148.359k in 5.068723s
Total allocated: 1.58 kB (3 objects)
Total retained: 0 B (0 objects)
Total allocated: 19.58 kB (3 objects)
Total retained: 0 B (0 objects)
```
I got a factor 6.2 this time: (1/24.168-1/29.288)/(1/28.315-1/29.288)
=
=
----------------------------------------
Bug #17031: `Kernel#caller_locations(m, n)` should be optimized
https://bugs.ruby-lang.org/issues/17031#change-86559
* Author: marcandre (Marc-Andre Lafortune)
* Status: Open
* Priority: Normal
* Backport: 2.5: UNKNOWN, 2.6: UNKNOWN, 2.7: UNKNOWN
----------------------------------------
`Kernel#caller_locations(1, 1)` currently appears to needlessly allocate me=
mory for the whole backtrace.
It allocates ~20kB for a 800-deep stacktrace, vs 1.6 kB for a shallow backt=
race.
It is also much slower for long stacktraces: about 7x slower for a 800-deep=
backtrace than for a shallow one.
Test used:
```ruby
def do_something
location =3D caller_locations(1, 1).first
end
def test(depth, trigger)
do_something if depth =3D=3D trigger
test(depth - 1, trigger) unless depth =3D=3D 0
end
require 'benchmark/ips'
Benchmark.ips do |x|
x.report (:short_backtrace ) {test(800,800)}
x.report (:long_backtrace ) {test(800, 0)}
x.report (:no_caller_locations) {test(800, -1)}
end
require 'memory_profiler'
MemoryProfiler.report { test(800,800) }.pretty_print(scale_bytes: true, det=
ailed_report: false)
MemoryProfiler.report { test(800, 0) }.pretty_print(scale_bytes: true, det=
ailed_report: false)
```
Found when checking memory usage on RuboCop.
-- =
https://bugs.ruby-lang.org/
Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=3Dunsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>