1Stdlib.Gc(3) OCaml library Stdlib.Gc(3)
2
6 Stdlib.Gc - no description
7
9 Module Stdlib.Gc
10
12 Module Gc
13 : (module Stdlib__gc)
14 type stat = {
22 minor_words : float ; (* Number of words allocated in the minor heap
23 since the program was started.
24 *)
25 promoted_words : float ; (* Number of words allocated in the minor
26 heap that survived a minor collection and were moved to the major heap
27 since the program was started.
28 *)
29 major_words : float ; (* Number of words allocated in the major heap,
30 including the promoted words, since the program was started.
31 *)
32 minor_collections : int ; (* Number of minor collections since the
33 program was started.
34 *)
35 major_collections : int ; (* Number of major collection cycles com‐
36 pleted since the program was started.
37 *)
38 heap_words : int ; (* Total size of the major heap, in words.
39 *)
40 heap_chunks : int ; (* Number of contiguous pieces of memory that
41 make up the major heap.
42 *)
43 live_words : int ; (* Number of words of live data in the major heap,
44 including the header words.
45 *)
46 live_blocks : int ; (* Number of live blocks in the major heap.
47 *)
48 free_words : int ; (* Number of words in the free list.
49 *)
50 free_blocks : int ; (* Number of blocks in the free list.
51 *)
52 largest_free : int ; (* Size (in words) of the largest block in the
53 free list.
54 *)
55 fragments : int ; (* Number of wasted words due to fragmentation.
56 These are 1-words free blocks placed between two live blocks. They are
57 not available for allocation.
58 *)
59 compactions : int ; (* Number of heap compactions since the program
60 was started.
61 *)
62 top_heap_words : int ; (* Maximum size reached by the major heap, in
63 words.
64 *)
65 stack_size : int ; (* Current size of the stack, in words.
66 Since 3.12.0
69 *)
70 forced_major_collections : int ; (* Number of forced full major col‐
71 lections completed since the program was started.
72 Since 4.12.0
75 *)
76 }
77 The memory management counters are returned in a stat record.
80 The total amount of memory allocated by the program since it was
82 started is (in words) minor_words + major_words - promoted_words .
83 Multiply by the word size (4 on a 32-bit machine, 8 on a 64-bit ma‐
84 chine) to get the number of bytes.
85 type control = {
88 mutable minor_heap_size : int ; (* The size (in words) of the minor
90 heap. Changing this parameter will trigger a minor collection. De‐
91 fault: 256k.
92 *)
93 mutable major_heap_increment : int ; (* How much to add to the major
95 heap when increasing it. If this number is less than or equal to 1000,
96 it is a percentage of the current heap size (i.e. setting it to 100
97 will double the heap size at each increase). If it is more than 1000,
98 it is a fixed number of words that will be added to the heap. Default:
99 15.
100 *)
101 mutable space_overhead : int ; (* The major GC speed is computed from
103 this parameter. This is the memory that will be "wasted" because the
104 GC does not immediately collect unreachable blocks. It is expressed as
105 a percentage of the memory used for live data. The GC will work more
106 (use more CPU time and collect blocks more eagerly) if space_overhead
107 is smaller. Default: 80.
108 *)
109 mutable verbose : int ; (* This value controls the GC messages on
111 standard error output. It is a sum of some of the following flags, to
112 print messages on the corresponding events:
113 - 0x001 Start and end of major GC cycle.
115 - 0x002 Minor collection and major GC slice.
117 - 0x004 Growing and shrinking of the heap.
119 - 0x008 Resizing of stacks and memory manager tables.
121 - 0x010 Heap compaction.
123 - 0x020 Change of GC parameters.
125 - 0x040 Computation of major GC slice size.
127 - 0x080 Calling of finalisation functions.
129 - 0x100 Bytecode executable and shared library search at start-up.
131 - 0x200 Computation of compaction-triggering condition.
133 - 0x400 Output GC statistics at program exit. Default: 0.
135 *)
137 mutable max_overhead : int ; (* Heap compaction is triggered when the
139 estimated amount of "wasted" memory is more than max_overhead percent
140 of the amount of live data. If max_overhead is set to 0, heap com‐
141 paction is triggered at the end of each major GC cycle (this setting is
142 intended for testing purposes only). If max_overhead >= 1000000 , com‐
143 paction is never triggered. If compaction is permanently disabled, it
144 is strongly suggested to set allocation_policy to 2. Default: 500.
145 *)
146 mutable stack_limit : int ; (* The maximum size of the stack (in
148 words). This is only relevant to the byte-code runtime, as the native
149 code runtime uses the operating system's stack. Default: 1024k.
150 *)
151 mutable allocation_policy : int ; (* The policy used for allocating in
153 the major heap. Possible values are 0, 1 and 2.
154 -0 is the next-fit policy, which is usually fast but can result in
157 fragmentation, increasing memory consumption.
158 -1 is the first-fit policy, which avoids fragmentation but has corner
161 cases (in certain realistic workloads) where it is sensibly slower.
162 -2 is the best-fit policy, which is fast and avoids fragmentation. In
165 our experiments it is faster and uses less memory than both next-fit
166 and first-fit. (since OCaml 4.10)
167 The current default is next-fit, as the best-fit policy is new and not
169 yet widely tested. We expect best-fit to become the default in the fu‐
170 ture.
171 On one example that was known to be bad for next-fit and first-fit,
173 next-fit takes 28s using 855Mio of memory, first-fit takes 47s using
174 566Mio of memory, best-fit takes 27s using 545Mio of memory.
175 Note: When changing to a low-fragmentation policy, you may need to aug‐
177 ment the space_overhead setting, for example using 100 instead of the
178 default 80 which is tuned for next-fit. Indeed, the difference in frag‐
179 mentation behavior means that different policies will have different
180 proportion of "wasted space" for a given program. Less fragmentation
181 means a smaller heap so, for the same amount of wasted space, a higher
182 proportion of wasted space. This makes the GC work harder, unless you
183 relax it by increasing space_overhead .
184 Note: changing the allocation policy at run-time forces a heap com‐
186 paction, which is a lengthy operation unless the heap is small (e.g. at
187 the start of the program).
188 Default: 0.
190 Since 3.11.0
193 *)
194 window_size : int ; (* The size of the window used by the major GC
195 for smoothing out variations in its workload. This is an integer be‐
196 tween 1 and 50. Default: 1.
197 Since 4.03.0
200 *)
201 custom_major_ratio : int ; (* Target ratio of floating garbage to ma‐
202 jor heap size for out-of-heap memory held by custom values located in
203 the major heap. The GC speed is adjusted to try to use this much memory
204 for dead values that are not yet collected. Expressed as a percentage
205 of major heap size. The default value keeps the out-of-heap floating
206 garbage about the same size as the in-heap overhead. Note: this only
207 applies to values allocated with caml_alloc_custom_mem (e.g. bigar‐
208 rays). Default: 44.
209 Since 4.08.0
212 *)
213 custom_minor_ratio : int ; (* Bound on floating garbage for
214 out-of-heap memory held by custom values in the minor heap. A minor GC
215 is triggered when this much memory is held by custom values located in
216 the minor heap. Expressed as a percentage of minor heap size. Note:
217 this only applies to values allocated with caml_alloc_custom_mem (e.g.
218 bigarrays). Default: 100.
219 Since 4.08.0
222 *)
223 custom_minor_max_size : int ; (* Maximum amount of out-of-heap memory
224 for each custom value allocated in the minor heap. When a custom value
225 is allocated on the minor heap and holds more than this many bytes,
226 only this value is counted against custom_minor_ratio and the rest is
227 directly counted against custom_major_ratio . Note: this only applies
228 to values allocated with caml_alloc_custom_mem (e.g. bigarrays). De‐
229 fault: 8192 bytes.
230 Since 4.08.0
233 *)
234 }
235 The GC parameters are given as a control record. Note that these pa‐
238 rameters can also be initialised by setting the OCAMLRUNPARAM environ‐
239 ment variable. See the documentation of ocamlrun .
240 val stat : unit -> stat
244 Return the current values of the memory management counters in a stat
246 record. This function examines every heap block to get the statistics.
247 val quick_stat : unit -> stat
251 Same as stat except that live_words , live_blocks , free_words ,
253 free_blocks , largest_free , and fragments are set to 0. This function
254 is much faster than stat because it does not need to go through the
255 heap.
256 val counters : unit -> float * float * float
260 Return (minor_words, promoted_words, major_words) . This function is
262 as fast as quick_stat .
263 val minor_words : unit -> float
267 Number of words allocated in the minor heap since the program was
269 started. This number is accurate in byte-code programs, but only an ap‐
270 proximation in programs compiled to native code.
271 In native code this function does not allocate.
273 Since 4.04
276 val get : unit -> control
280 Return the current values of the GC parameters in a control record.
282 val set : control -> unit
286 set r changes the GC parameters according to the control record r .
289 The normal usage is: Gc.set { (Gc.get()) with Gc.verbose = 0x00d }
290 val minor : unit -> unit
295 Trigger a minor collection.
297 val major_slice : int -> int
301 major_slice n Do a minor collection and a slice of major collection. n
304 is the size of the slice: the GC will do enough work to free (on aver‐
305 age) n words of memory. If n = 0, the GC will try to do enough work to
306 ensure that the next automatic slice has no work to do. This function
307 returns an unspecified integer (currently: 0).
308 val major : unit -> unit
312 Do a minor collection and finish the current major collection cycle.
314 val full_major : unit -> unit
318 Do a minor collection, finish the current major collection cycle, and
320 perform a complete new cycle. This will collect all currently unreach‐
321 able blocks.
322 val compact : unit -> unit
326 Perform a full major collection and compact the heap. Note that heap
328 compaction is a lengthy operation.
329 val print_stat : out_channel -> unit
333 Print the current values of the memory management counters (in hu‐
335 man-readable form) into the channel argument.
336 val allocated_bytes : unit -> float
340 Return the total number of bytes allocated since the program was
342 started. It is returned as a float to avoid overflow problems with int
343 on 32-bit machines.
344 val get_minor_free : unit -> int
348 Return the current size of the free space inside the minor heap.
350 Since 4.03.0
353 val get_bucket : int -> int
357 get_bucket n returns the current size of the n -th future bucket of the
360 GC smoothing system. The unit is one millionth of a full GC.
361 Since 4.03.0
364 Raises Invalid_argument if n is negative, return 0 if n is larger than
367 the smoothing window.
368 val get_credit : unit -> int
372 get_credit () returns the current size of the "work done in advance"
375 counter of the GC smoothing system. The unit is one millionth of a full
376 GC.
377 Since 4.03.0
380 val huge_fallback_count : unit -> int
384 Return the number of times we tried to map huge pages and had to fall
386 back to small pages. This is always 0 if OCAMLRUNPARAM contains H=1 .
387 Since 4.03.0
390 val finalise : ('a -> unit) -> 'a -> unit
394 finalise f v registers f as a finalisation function for v . v must be
397 heap-allocated. f will be called with v as argument at some point be‐
398 tween the first time v becomes unreachable (including through weak
399 pointers) and the time v is collected by the GC. Several functions can
400 be registered for the same value, or even several instances of the same
401 function. Each instance will be called once (or never, if the program
402 terminates before v becomes unreachable).
403 The GC will call the finalisation functions in the order of dealloca‐
405 tion. When several values become unreachable at the same time (i.e.
406 during the same GC cycle), the finalisation functions will be called in
407 the reverse order of the corresponding calls to finalise . If finalise
408 is called in the same order as the values are allocated, that means
409 each value is finalised before the values it depends upon. Of course,
410 this becomes false if additional dependencies are introduced by assign‐
411 ments.
412 In the presence of multiple OCaml threads it should be assumed that any
414 particular finaliser may be executed in any of the threads.
415 Anything reachable from the closure of finalisation functions is con‐
417 sidered reachable, so the following code will not work as expected:
418 - let v = ... in Gc.finalise (fun _ -> ...v...) v
420 Instead you should make sure that v is not in the closure of the final‐
422 isation function by writing:
423 - let f = fun x -> ... let v = ... in Gc.finalise f v
425 The f function can use all features of OCaml, including assignments
427 that make the value reachable again. It can also loop forever (in this
428 case, the other finalisation functions will not be called during the
429 execution of f, unless it calls finalise_release ). It can call fi‐
430 nalise on v or other values to register other functions or even itself.
431 It can raise an exception; in this case the exception will interrupt
432 whatever the program was doing when the function was called.
433 finalise will raise Invalid_argument if v is not guaranteed to be
436 heap-allocated. Some examples of values that are not heap-allocated
437 are integers, constant constructors, booleans, the empty array, the
438 empty list, the unit value. The exact list of what is heap-allocated
439 or not is implementation-dependent. Some constant values can be
440 heap-allocated but never deallocated during the lifetime of the pro‐
441 gram, for example a list of integer constants; this is also implementa‐
442 tion-dependent. Note that values of types float are sometimes allo‐
443 cated and sometimes not, so finalising them is unsafe, and finalise
444 will also raise Invalid_argument for them. Values of type 'a Lazy.t
445 (for any 'a ) are like float in this respect, except that the compiler
446 sometimes optimizes them in a way that prevents finalise from detecting
447 them. In this case, it will not raise Invalid_argument , but you should
448 still avoid calling finalise on lazy values.
449 The results of calling String.make , Bytes.make , Bytes.create , Ar‐
451 ray.make , and ref are guaranteed to be heap-allocated and non-constant
452 except when the length argument is 0 .
453 val finalise_last : (unit -> unit) -> 'a -> unit
457 same as Gc.finalise except the value is not given as argument. So you
459 can't use the given value for the computation of the finalisation func‐
460 tion. The benefit is that the function is called after the value is un‐
461 reachable for the last time instead of the first time. So contrary to
462 Gc.finalise the value will never be reachable again or used again. In
463 particular every weak pointer and ephemeron that contained this value
464 as key or data is unset before running the finalisation function. More‐
465 over the finalisation functions attached with Gc.finalise are always
466 called before the finalisation functions attached with Gc.finalise_last
467 .
468 Since 4.04
471 val finalise_release : unit -> unit
475 A finalisation function may call finalise_release to tell the GC that
477 it can launch the next finalisation function without waiting for the
478 current one to return.
479 type alarm
482 An alarm is a piece of data that calls a user function at the end of
485 each major GC cycle. The following functions are provided to create
486 and delete alarms.
487 val create_alarm : (unit -> unit) -> alarm
491 create_alarm f will arrange for f to be called at the end of each major
494 GC cycle, starting with the current cycle or the next one. A value of
495 type alarm is returned that you can use to call delete_alarm .
496 val delete_alarm : alarm -> unit
500 delete_alarm a will stop the calls to the function associated to a .
503 Calling delete_alarm a again has no effect.
504 val eventlog_pause : unit -> unit
508 eventlog_pause () will pause the collection of traces in the runtime.
511 Traces are collected if the program is linked to the instrumented run‐
512 time and started with the environment variable OCAML_EVENTLOG_ENABLED.
513 Events are flushed to disk after pausing, and no new events will be
514 recorded until eventlog_resume is called.
515 val eventlog_resume : unit -> unit
519 eventlog_resume () will resume the collection of traces in the runtime.
522 Traces are collected if the program is linked to the instrumented run‐
523 time and started with the environment variable OCAML_EVENTLOG_ENABLED.
524 This call can be used after calling eventlog_pause , or if the program
525 was started with OCAML_EVENTLOG_ENABLED=p. (which pauses the collection
526 of traces before the first event.)
527 module Memprof : sig end
530 Memprof is a sampling engine for allocated memory words. Every allo‐
534 cated word has a probability of being sampled equal to a configurable
535 sampling rate. Once a block is sampled, it becomes tracked. A tracked
536 block triggers a user-defined callback as soon as it is allocated, pro‐
537 moted or deallocated.
538 Since blocks are composed of several words, a block can potentially be
540 sampled several times. If a block is sampled several times, then each
541 of the callback is called once for each event of this block: the multi‐
542 plicity is given in the n_samples field of the allocation structure.
543 This engine makes it possible to implement a low-overhead memory pro‐
545 filer as an OCaml library.
546 Note: this API is EXPERIMENTAL. It may change without prior notice.
548OCamldoc 2021-07-22 Stdlib.Gc(3)