1Gc(3) OCaml library Gc(3)
2
6 Gc - Memory management control and statistics; finalised values.
7
9 Module Gc
10
12 Module Gc
13 : sig end
14 Memory management control and statistics; finalised values.
17 type stat = {
23 minor_words : float ; (* Number of words allocated in the minor heap
24 since the program was started.
25 *)
26 promoted_words : float ; (* Number of words allocated in the minor
27 heap that survived a minor collection and were moved to the major heap
28 since the program was started.
29 *)
30 major_words : float ; (* Number of words allocated in the major heap,
31 including the promoted words, since the program was started.
32 *)
33 minor_collections : int ; (* Number of minor collections since the
34 program was started.
35 *)
36 major_collections : int ; (* Number of major collection cycles com‐
37 pleted since the program was started.
38 *)
39 heap_words : int ; (* Total size of the major heap, in words.
40 *)
41 heap_chunks : int ; (* Number of contiguous pieces of memory that
42 make up the major heap.
43 *)
44 live_words : int ; (* Number of words of live data in the major heap,
45 including the header words.
46 Note that "live" words refers to every word in the major heap that
48 isn't currently known to be collectable, which includes words that have
49 become unreachable by the program after the start of the previous gc
50 cycle. It is typically much simpler and more predictable to call
51 Gc.full_major (or Gc.compact ) then computing gc stats, as then "live"
52 words has the simple meaning of "reachable by the program". One caveat
53 is that a single call to Gc.full_major will not reclaim values that
54 have a finaliser from Gc.finalise (this does not apply to Gc.fi‐
55 nalise_last ). If this caveat matters, simply call Gc.full_major twice
56 instead of once.
57 *)
58 live_blocks : int ; (* Number of live blocks in the major heap.
59 See live_words for a caveat about what "live" means.
61 *)
62 free_words : int ; (* Number of words in the free list.
63 *)
64 free_blocks : int ; (* Number of blocks in the free list.
65 *)
66 largest_free : int ; (* Size (in words) of the largest block in the
67 free list.
68 *)
69 fragments : int ; (* Number of wasted words due to fragmentation.
70 These are 1-words free blocks placed between two live blocks. They are
71 not available for allocation.
72 *)
73 compactions : int ; (* Number of heap compactions since the program
74 was started.
75 *)
76 top_heap_words : int ; (* Maximum size reached by the major heap, in
77 words.
78 *)
79 stack_size : int ; (* Current size of the stack, in words.
80 Since 3.12.0
83 *)
84 forced_major_collections : int ; (* Number of forced full major col‐
85 lections completed since the program was started.
86 Since 4.12.0
89 *)
90 }
91 The memory management counters are returned in a stat record.
94 The total amount of memory allocated by the program since it was
96 started is (in words) minor_words + major_words - promoted_words .
97 Multiply by the word size (4 on a 32-bit machine, 8 on a 64-bit ma‐
98 chine) to get the number of bytes.
99 type control = {
102 mutable minor_heap_size : int ; (* The size (in words) of the minor
104 heap. Changing this parameter will trigger a minor collection. De‐
105 fault: 256k.
106 *)
107 mutable major_heap_increment : int ; (* How much to add to the major
109 heap when increasing it. If this number is less than or equal to 1000,
110 it is a percentage of the current heap size (i.e. setting it to 100
111 will double the heap size at each increase). If it is more than 1000,
112 it is a fixed number of words that will be added to the heap. Default:
113 15.
114 *)
115 mutable space_overhead : int ; (* The major GC speed is computed from
117 this parameter. This is the memory that will be "wasted" because the
118 GC does not immediately collect unreachable blocks. It is expressed as
119 a percentage of the memory used for live data. The GC will work more
120 (use more CPU time and collect blocks more eagerly) if space_overhead
121 is smaller. Default: 120.
122 *)
123 mutable verbose : int ; (* This value controls the GC messages on
125 standard error output. It is a sum of some of the following flags, to
126 print messages on the corresponding events:
127 - 0x001 Start and end of major GC cycle.
129 - 0x002 Minor collection and major GC slice.
131 - 0x004 Growing and shrinking of the heap.
133 - 0x008 Resizing of stacks and memory manager tables.
135 - 0x010 Heap compaction.
137 - 0x020 Change of GC parameters.
139 - 0x040 Computation of major GC slice size.
141 - 0x080 Calling of finalisation functions.
143 - 0x100 Bytecode executable and shared library search at start-up.
145 - 0x200 Computation of compaction-triggering condition.
147 - 0x400 Output GC statistics at program exit. Default: 0.
149 *)
151 mutable max_overhead : int ; (* Heap compaction is triggered when the
153 estimated amount of "wasted" memory is more than max_overhead percent
154 of the amount of live data. If max_overhead is set to 0, heap com‐
155 paction is triggered at the end of each major GC cycle (this setting is
156 intended for testing purposes only). If max_overhead >= 1000000 , com‐
157 paction is never triggered. If compaction is permanently disabled, it
158 is strongly suggested to set allocation_policy to 2. Default: 500.
159 *)
160 mutable stack_limit : int ; (* The maximum size of the stack (in
162 words). This is only relevant to the byte-code runtime, as the native
163 code runtime uses the operating system's stack. Default: 1024k.
164 *)
165 mutable allocation_policy : int ; (* The policy used for allocating in
167 the major heap. Possible values are 0, 1 and 2.
168 -0 is the next-fit policy, which is usually fast but can result in
171 fragmentation, increasing memory consumption.
172 -1 is the first-fit policy, which avoids fragmentation but has corner
175 cases (in certain realistic workloads) where it is sensibly slower.
176 -2 is the best-fit policy, which is fast and avoids fragmentation. In
179 our experiments it is faster and uses less memory than both next-fit
180 and first-fit. (since OCaml 4.10)
181 The default is best-fit.
183 On one example that was known to be bad for next-fit and first-fit,
185 next-fit takes 28s using 855Mio of memory, first-fit takes 47s using
186 566Mio of memory, best-fit takes 27s using 545Mio of memory.
187 Note: If you change to next-fit, you may need to reduce the space_over‐
189 head setting, for example using 80 instead of the default 120 which is
190 tuned for best-fit. Otherwise, your program will need more memory.
191 Note: changing the allocation policy at run-time forces a heap com‐
193 paction, which is a lengthy operation unless the heap is small (e.g. at
194 the start of the program).
195 Default: 2.
197 Since 3.11.0
200 *)
201 window_size : int ; (* The size of the window used by the major GC
202 for smoothing out variations in its workload. This is an integer be‐
203 tween 1 and 50. Default: 1.
204 Since 4.03.0
207 *)
208 custom_major_ratio : int ; (* Target ratio of floating garbage to ma‐
209 jor heap size for out-of-heap memory held by custom values located in
210 the major heap. The GC speed is adjusted to try to use this much memory
211 for dead values that are not yet collected. Expressed as a percentage
212 of major heap size. The default value keeps the out-of-heap floating
213 garbage about the same size as the in-heap overhead. Note: this only
214 applies to values allocated with caml_alloc_custom_mem (e.g. bigar‐
215 rays). Default: 44.
216 Since 4.08.0
219 *)
220 custom_minor_ratio : int ; (* Bound on floating garbage for
221 out-of-heap memory held by custom values in the minor heap. A minor GC
222 is triggered when this much memory is held by custom values located in
223 the minor heap. Expressed as a percentage of minor heap size. Note:
224 this only applies to values allocated with caml_alloc_custom_mem (e.g.
225 bigarrays). Default: 100.
226 Since 4.08.0
229 *)
230 custom_minor_max_size : int ; (* Maximum amount of out-of-heap memory
231 for each custom value allocated in the minor heap. When a custom value
232 is allocated on the minor heap and holds more than this many bytes,
233 only this value is counted against custom_minor_ratio and the rest is
234 directly counted against custom_major_ratio . Note: this only applies
235 to values allocated with caml_alloc_custom_mem (e.g. bigarrays). De‐
236 fault: 8192 bytes.
237 Since 4.08.0
240 *)
241 }
242 The GC parameters are given as a control record. Note that these pa‐
245 rameters can also be initialised by setting the OCAMLRUNPARAM environ‐
246 ment variable. See the documentation of ocamlrun .
247 val stat : unit -> stat
251 Return the current values of the memory management counters in a stat
253 record. This function examines every heap block to get the statistics.
254 val quick_stat : unit -> stat
258 Same as stat except that live_words , live_blocks , free_words ,
260 free_blocks , largest_free , and fragments are set to 0. This function
261 is much faster than stat because it does not need to go through the
262 heap.
263 val counters : unit -> float * float * float
267 Return (minor_words, promoted_words, major_words) . This function is
269 as fast as quick_stat .
270 val minor_words : unit -> float
274 Number of words allocated in the minor heap since the program was
276 started. This number is accurate in byte-code programs, but only an ap‐
277 proximation in programs compiled to native code.
278 In native code this function does not allocate.
280 Since 4.04
283 val get : unit -> control
287 Return the current values of the GC parameters in a control record.
289 val set : control -> unit
293 set r changes the GC parameters according to the control record r .
296 The normal usage is: Gc.set { (Gc.get()) with Gc.verbose = 0x00d }
297 val minor : unit -> unit
302 Trigger a minor collection.
304 val major_slice : int -> int
308 major_slice n Do a minor collection and a slice of major collection. n
311 is the size of the slice: the GC will do enough work to free (on aver‐
312 age) n words of memory. If n = 0, the GC will try to do enough work to
313 ensure that the next automatic slice has no work to do. This function
314 returns an unspecified integer (currently: 0).
315 val major : unit -> unit
319 Do a minor collection and finish the current major collection cycle.
321 val full_major : unit -> unit
325 Do a minor collection, finish the current major collection cycle, and
327 perform a complete new cycle. This will collect all currently unreach‐
328 able blocks.
329 val compact : unit -> unit
333 Perform a full major collection and compact the heap. Note that heap
335 compaction is a lengthy operation.
336 val print_stat : out_channel -> unit
340 Print the current values of the memory management counters (in hu‐
342 man-readable form) into the channel argument.
343 val allocated_bytes : unit -> float
347 Return the total number of bytes allocated since the program was
349 started. It is returned as a float to avoid overflow problems with int
350 on 32-bit machines.
351 val get_minor_free : unit -> int
355 Return the current size of the free space inside the minor heap.
357 Since 4.03.0
360 val get_bucket : int -> int
364 get_bucket n returns the current size of the n -th future bucket of the
367 GC smoothing system. The unit is one millionth of a full GC.
368 Since 4.03.0
371 Raises Invalid_argument if n is negative, return 0 if n is larger than
374 the smoothing window.
375 val get_credit : unit -> int
379 get_credit () returns the current size of the "work done in advance"
382 counter of the GC smoothing system. The unit is one millionth of a full
383 GC.
384 Since 4.03.0
387 val huge_fallback_count : unit -> int
391 Return the number of times we tried to map huge pages and had to fall
393 back to small pages. This is always 0 if OCAMLRUNPARAM contains H=1 .
394 Since 4.03.0
397 val finalise : ('a -> unit) -> 'a -> unit
401 finalise f v registers f as a finalisation function for v . v must be
404 heap-allocated. f will be called with v as argument at some point be‐
405 tween the first time v becomes unreachable (including through weak
406 pointers) and the time v is collected by the GC. Several functions can
407 be registered for the same value, or even several instances of the same
408 function. Each instance will be called once (or never, if the program
409 terminates before v becomes unreachable).
410 The GC will call the finalisation functions in the order of dealloca‐
412 tion. When several values become unreachable at the same time (i.e.
413 during the same GC cycle), the finalisation functions will be called in
414 the reverse order of the corresponding calls to finalise . If finalise
415 is called in the same order as the values are allocated, that means
416 each value is finalised before the values it depends upon. Of course,
417 this becomes false if additional dependencies are introduced by assign‐
418 ments.
419 In the presence of multiple OCaml threads it should be assumed that any
421 particular finaliser may be executed in any of the threads.
422 Anything reachable from the closure of finalisation functions is con‐
424 sidered reachable, so the following code will not work as expected:
425 - let v = ... in Gc.finalise (fun _ -> ...v...) v
427 Instead you should make sure that v is not in the closure of the final‐
429 isation function by writing:
430 - let f = fun x -> ... let v = ... in Gc.finalise f v
432 The f function can use all features of OCaml, including assignments
434 that make the value reachable again. It can also loop forever (in this
435 case, the other finalisation functions will not be called during the
436 execution of f, unless it calls finalise_release ). It can call fi‐
437 nalise on v or other values to register other functions or even itself.
438 It can raise an exception; in this case the exception will interrupt
439 whatever the program was doing when the function was called.
440 finalise will raise Invalid_argument if v is not guaranteed to be
443 heap-allocated. Some examples of values that are not heap-allocated
444 are integers, constant constructors, booleans, the empty array, the
445 empty list, the unit value. The exact list of what is heap-allocated
446 or not is implementation-dependent. Some constant values can be
447 heap-allocated but never deallocated during the lifetime of the pro‐
448 gram, for example a list of integer constants; this is also implementa‐
449 tion-dependent. Note that values of types float are sometimes allo‐
450 cated and sometimes not, so finalising them is unsafe, and finalise
451 will also raise Invalid_argument for them. Values of type 'a Lazy.t
452 (for any 'a ) are like float in this respect, except that the compiler
453 sometimes optimizes them in a way that prevents finalise from detecting
454 them. In this case, it will not raise Invalid_argument , but you should
455 still avoid calling finalise on lazy values.
456 The results of calling String.make , Bytes.make , Bytes.create , Ar‐
458 ray.make , and ref are guaranteed to be heap-allocated and non-constant
459 except when the length argument is 0 .
460 val finalise_last : (unit -> unit) -> 'a -> unit
464 same as Gc.finalise except the value is not given as argument. So you
466 can't use the given value for the computation of the finalisation func‐
467 tion. The benefit is that the function is called after the value is un‐
468 reachable for the last time instead of the first time. So contrary to
469 Gc.finalise the value will never be reachable again or used again. In
470 particular every weak pointer and ephemeron that contained this value
471 as key or data is unset before running the finalisation function. More‐
472 over the finalisation functions attached with Gc.finalise are always
473 called before the finalisation functions attached with Gc.finalise_last
474 .
475 Since 4.04
478 val finalise_release : unit -> unit
482 A finalisation function may call finalise_release to tell the GC that
484 it can launch the next finalisation function without waiting for the
485 current one to return.
486 type alarm
489 An alarm is a piece of data that calls a user function at the end of
492 each major GC cycle. The following functions are provided to create
493 and delete alarms.
494 val create_alarm : (unit -> unit) -> alarm
498 create_alarm f will arrange for f to be called at the end of each major
501 GC cycle, starting with the current cycle or the next one. A value of
502 type alarm is returned that you can use to call delete_alarm .
503 val delete_alarm : alarm -> unit
507 delete_alarm a will stop the calls to the function associated to a .
510 Calling delete_alarm a again has no effect.
511 val eventlog_pause : unit -> unit
515 eventlog_pause () will pause the collection of traces in the runtime.
518 Traces are collected if the program is linked to the instrumented run‐
519 time and started with the environment variable OCAML_EVENTLOG_ENABLED.
520 Events are flushed to disk after pausing, and no new events will be
521 recorded until eventlog_resume is called.
522 Since 4.11
525 val eventlog_resume : unit -> unit
529 eventlog_resume () will resume the collection of traces in the runtime.
532 Traces are collected if the program is linked to the instrumented run‐
533 time and started with the environment variable OCAML_EVENTLOG_ENABLED.
534 This call can be used after calling eventlog_pause , or if the program
535 was started with OCAML_EVENTLOG_ENABLED=p. (which pauses the collection
536 of traces before the first event.)
537 Since 4.11
540 module Memprof : sig end
543 Memprof is a sampling engine for allocated memory words. Every allo‐
547 cated word has a probability of being sampled equal to a configurable
548 sampling rate. Once a block is sampled, it becomes tracked. A tracked
549 block triggers a user-defined callback as soon as it is allocated, pro‐
550 moted or deallocated.
551 Since blocks are composed of several words, a block can potentially be
553 sampled several times. If a block is sampled several times, then each
554 of the callback is called once for each event of this block: the multi‐
555 plicity is given in the n_samples field of the allocation structure.
556 This engine makes it possible to implement a low-overhead memory pro‐
558 filer as an OCaml library.
559 Note: this API is EXPERIMENTAL. It may change without prior notice.
561OCamldoc 2023-01-23 Gc(3)