* vm_core.h (rb_thread_t::tag_state): move to "rb_vm_tag::state".
Lifetime of "state" should be same as current tag.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59159 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_core.h (rb_thread_t): rename rb_thread_t::state to tag_state
to make it clear.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59158 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (vm_stack_consistency_error): extracted from
insns.def for further info in the future.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59149 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_core.h (OPT_CALL_CFUNC_WITHOUT_FRAME): removed because nobody use it.
This optimization tries to call C-methods without pushing VM frames,
however no big improvements compare with this complexity.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59076 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* compile.c (iseq_compile_each0): save target child_iseq in the catch-table
for break. This iseq is not for continuation, but for search key at
vm_throw_start().
* vm_insnhelper.c (vm_throw_start): check saved iseq first.
* iseq.h: add comment for it.
* test/ruby/test_iterator.rb (test_ljump): add a test for the issue:
def call b; b.call; end
call(Proc.new{break}){} #=> (1) should raise LocalJumpError
call(Proc.new{break}) #=> (2) shoudd raies LocalJumpError, too.
but (1) doesn't raise LocalJumpError.
This issue is reported by Matz.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59043 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_args.c (enum arg_setup_type): remove arg_setup_lambda because it is
same as arg_setup_method.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59014 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
VM_FRAME_FLAG_LAMBDA. So that lambda block frame is
VM_FRAME_MAGIC_BLOCK | VM_FRAME_FLAG_LAMBDA.
* vm_core.h: remvoe VM_FRAME_MAGIC_PROC because nobody uses it.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@59006 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
To convert the object implicitly, it has had two parts in convert_type() which are
1. lookink up the method's id
2. calling the method
Seems that strncmp() and strcmp() in convert_type() are slightly heavy to look up
the method's id for type conversion.
This patch will add and use internal APIs (rb_convert_type_with_id, rb_check_convert_type_with_id)
to call the method without looking up the method's id when convert the object.
Array#flatten -> 19 % up
Array#+ -> 3 % up
[ruby-dev:50024] [Bug #13341] [Fix GH-1537]
### Before
Array#flatten 104.119k (± 1.1%) i/s - 525.690k in 5.049517s
Array#+ 1.993M (± 1.8%) i/s - 10.010M in 5.024258s
### After
Array#flatten 124.005k (± 1.0%) i/s - 624.240k in 5.034477s
Array#+ 2.058M (± 4.8%) i/s - 10.302M in 5.019328s
### Test Code
require 'benchmark/ips'
class Foo
def to_ary
[1,2,3]
end
end
Benchmark.ips do |x|
ary = []
100.times { |i| ary << i }
array = [ary]
x.report "Array#flatten" do |i|
i.times { array.flatten }
end
x.report "Array#+" do |i|
obj = Foo.new
i.times { array + obj }
end
end
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58978 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* debug_counter.h: add the following counters:
* lvar_get: counter for lvar get.
* lvar_get_dynamic: counter for lvar get from upper frames.
* lvar_set: coutner for lvar set.
* lvar_set_dynamic: coutner for lvar set from upper frames.
* lvar_set_slowpath: counter for lavr set using slowpath.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58977 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (comparable_by_identity): extract the condition
where comparable by identity. currently both are same types,
Fixnum, Flonum, or Symbol.
* vm_insnhelper.c (opt_eql_func): support Symbol too.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58896 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (rb_eql_opt): should call #eql? on Float and
String, not #==.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58882 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This improvement is similar with https://github.com/ruby/ruby/pull/1552
internal.h: add declaration of rb_eql_opt() API.
vm_insnhelper.c (rb_eql_opt): add rb_eql_opt() API which provides optimized
path for #eql? method such as rb_equal_opt().
object.c (rb_eql): optimize using rb_eql_opt() such as rb_equal().
Array#eql? and some methods have used rb_eql() and Array#eql? will be faster
around 20%.
[ruby-core:80761] [Bug #13447] [Fix GH-#1589]
### Before
user system total real
1.570000 0.000000 1.570000 ( 1.569754)
### After
user system total real
1.300000 0.000000 1.300000 ( 1.303624)
### Test code
require 'benchmark'
Benchmark.bmbm do |x|
ary1 = Array.new(1000) { rand(1000) }
ary2 = Array.new(1000) { rand(1000) }
x.report do
5000000.times do
ary1.eql?(ary2)
end
end
end
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58881 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
The goal is to reduce rb_context_t and rb_fiber_t size
by removing the need to store the entire rb_thread_t in
there.
[ruby-core:81045] Work-in-progress: soon, we will move more fields here.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58614 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (vm_once_dispatch): no guarantee that tail call
is always optimized away.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58405 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Those opt_something instructions are worth refactoring. They tend to
have similar ways of executions. By extracting the common part,
generated vm_exec_core function shrinks from 26,816 bytes to 26,256
bytes (according to nm(1)).
This changeset introduces negligible performance impact. 3 repeated
runs of optcarrot benchmark on my machine resulted in:
before this: 28.813363684823557, 27.523907198440366, 27.292766121965400
after this: 28.174038497265080, 28.999513875020405, 29.621399800428065
in fps (greater==faster).
----
* vm_insnhelper.c (vm_opt_binop_dispatch): new function.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58402 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Contemporary C compilers are good at function inlining. They fold
multiple functions into one. However they are not yet smart enough to
unfold a function into several ones. So generally speaking, it is
wiser for a C programmer to manually split C functions whenever
possible. That should make rooms for compilers to optimize at will.
Before this changeset insns.def was converted into single HUGE
function called vm_exec_core(). By moving each instruction's core
into individual functions, generated C source code is reduced from
3,428 lines to 2,847 lines. Looking at the generated assembly
however, it seems my compiler (gcc 6.2) is extraordinary smart so that
it inlines almost all functions I introduced in this changeset back
into that vm_exec_core. On my machine compiled machine binary of the
function does not shrink very much in size (28,432 bytes to 26,816
bytes, according to nm(1)).
I believe this change is zero-cost. Several benchmarks I exercised
showed no significant difference beyond error mergin. For instance
3 repeated runs of optcarrot benchmark on my machine resulted in:
before this: 28.330329285707490, 27.513378371065920, 29.40420215754537
after this: 27.107195867280414, 25.549324021385907, 30.31581919050884
in fps (greater==faster).
----
* internal.h (rb_obj_not_equal): used from vm_insnhelper.c
* insns.def: move vast majority of lines into vm_insnhelper.c
* vm_insnhelper.c: moved here.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58390 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (rb_threadptr_stack_overflow): move from
thread.c and integrate with vm_stackoverflow.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58379 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_eval.c (rb_yield_lambda): new function which yields an array
to a proc and splat to a lambda. mainly for Enumerable only.
* vm_args.c (setup_parameters_complex): remove special lambda
splatting for [Bug #9605]. [ruby-core:77065] [Bug #12705]
* vm_insnhelper.c (vm_callee_setup_block_arg): ditto.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58019 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm.c (invoke_iseq_block_from_c): guard arguments on stack, not
to be clobbered during splatting.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@58012 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (opt_eq_func): method to dispatch is resolved by
only the receiver's class, not including the argument class.
even if basic operation is redefined, other class conditions
never meet. optimize Float and non-Float case, delegate to
rb_float_equal directly.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57785 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* How to enable this feature?
* define USE_DEBUG_COUNTER as 1.
* you can disable to output the result with
RUBY_DEBUG_COUNTER_DISABLE environment variable
even if USE_DEBUG_COUNTER == 1.
* How to add new counter?
* add COUNTER(<name>) line on debug_counter.h.
* include "debug_counter.h"
* insert RB_DEBUG_COUNTER_INC(<name>) line on your favorite place.
* counter output example:
[RUBY_DEBUG_COUNTER] mc_inline_hit 999
[RUBY_DEBUG_COUNTER] mc_inline_miss 3
[RUBY_DEBUG_COUNTER] mc_global_hit 23
[RUBY_DEBUG_COUNTER] mc_global_miss 273
[RUBY_DEBUG_COUNTER] mc_global_state_miss 3
[RUBY_DEBUG_COUNTER] mc_class_serial_miss 0
[RUBY_DEBUG_COUNTER] mc_cme_complement 0
[RUBY_DEBUG_COUNTER] mc_cme_complement_hit 0
[RUBY_DEBUG_COUNTER] mc_search_super 1384
[RUBY_DEBUG_COUNTER] ivar_get_hit 0
[RUBY_DEBUG_COUNTER] ivar_get_miss 0
[RUBY_DEBUG_COUNTER] ivar_set_hit 0
[RUBY_DEBUG_COUNTER] ivar_set_miss 0
[RUBY_DEBUG_COUNTER] ivar_get 431
[RUBY_DEBUG_COUNTER] ivar_set 465
* mc_... is related to method caching.
* ivar_... is related to instance variable accesses.
* compare with dtrace/system tap features, there are completely
no performacne penalties when it is disabled.
* This feature is supported only on __GNUC__ compilers.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57676 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (vm_call_zsuper): method defined in module in
refinement is not callable as-is. dispatch again.
[ruby-core:79588] [Bug #13227]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57655 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* eval.c (setup_exception): make unfrozen copy of special
exception before setting up a cause.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57415 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (vm_call_iseq_setup_tailcall): check interrupts
after set up the new frame, not the passed block to be clobbered
by invoked finalizers and so on. [ruby-core:78981] [Bug #13107]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@57293 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (callable_class_p): update assertion as callable
class may be T_MODULE or I_ICLASS which refines a module since
r56213. [Feature #12534]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@56381 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
* vm_insnhelper.c (CHECK): remove extra semicolon and fold too
long line.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@55806 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
and VM_FRAME_CFRAME_P().
Most of case, RUBY_VM_NORMAL_ISEQ_P() is no
longer needed.
* vm_core.h: introduce rb_obj_is_iseq().
* cont.c, vm.c: VM_FRAME_MAGIC_DUMMY with
VM_FRAME_FLAG_CFRAME.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@55804 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
ko1
nobu
watson1978
normal
shyouhei
svn
kazu
ktsj
hsbt
shugo
naruse