| Message ID | 20191011070353.c5d7qbrxliw3vkom@kam.mff.cuni.cz |
|---|---|
| State | New |
| Headers | show |
| Series | Implement ggc_trim | expand |
On October 11, 2019 9:03:53 AM GMT+02:00, Jan Hubicka <hubicka@ucw.cz> wrote: >Hi, >this patch adds ggc_trim that releases free pages used by GGC memory >to system. This is useful to reduce memory footprint of WPA streaming: >WPA streaming ought to not use any more GGC memory (patches in testing >for that) and trimming the memory makes it available to fork&malloc >used >by stream out machinery. > >I collected some stats for cc1 for both GGC and heap (using mallinfo). >Memory footprints are as follows: > >After streaming in global stream: 123MB GGC; 25MB of heap. >After streaming in callgraph : 228MB GGC; 45MB of heap. >After streaming in summaries : 373MB GGC; 126MB of heap. >After symbol merging : 348MB GGC; 130MB of heap. >After IPA-ICF : 501MB GGC; 160MB of heap. (this is all ICF) >After IPA-CP : 528MB GGC; 163MB of heap. >After IPA-SRA : 532MB GGC; 163MB of heap. >After Inline : 644MB GGC; 173MB of heap > This is after collecting of 118MB of > garbage and returning 740k to system > by madvise_dontneed >After ipa-reference : 644MB GGC; 370MB of heap > I checked this all goes into the > bitmaps; I have WIP patch for that >After releasing summariess : 431MB GGC; 383MB of heap > Trim releases 43MB by unmap > and 321MB by madvise_dontneed > >At least i learnt new fact about ipa-reference consuming 200MB of >memory which was not obvious from our detailed mem stats. > >I think the lowest hanging fruit after this patch is to add >malloc_madvise which further reduces footpring and fix ipa-reference. >Hopefully Martin will do a bit about ipa-icf. > >I will dig into what inliner does but it produces a lot of clones so I >think it is mostly clone and summary duplication. Perhaps we can avoid >copying some of summaries for inline clones. > >In TOP I see about 900MB instead of 1.4GB before WPA streaming starts >with both ggc_trim and madvise. > >Note that I also tried to hack ggc_free to recognize free pages but at >least in simple implementation it is a loss since it makes ggc_alloc >more expensive (it needs to bring pages back and add into freelists) >which hurts stream-in performance. > >I think sweeping once per WPA is no problem, it is definitly less than >1% of WPA time. > >Bootstrapped/regtested x86_64-linux, OK? Ok. Richard. > * ggc-page.c (release_pages): Output statistics when !quiet_flag. > (ggc_collect): Dump later to not interfere with release_page dump. > (ggc_trim): New function. > * ggc-none.c (ggc_trim): New. > > * lto.c (lto_wpa_write_files): Call ggc_trim. >Index: ggc-page.c >=================================================================== >--- ggc-page.c (revision 276707) >+++ ggc-page.c (working copy) >@@ -529,7 +529,6 @@ static void clear_page_group_in_use (pag > #endif > static struct page_entry * alloc_page (unsigned); > static void free_page (struct page_entry *); >-static void release_pages (void); > static void clear_marks (void); > static void sweep_pages (void); > static void ggc_recalculate_in_use_p (page_entry *); >@@ -1016,6 +1015,8 @@ free_page (page_entry *entry) > static void > release_pages (void) > { >+ size_t n1 = 0; >+ size_t n2 = 0; > #ifdef USING_MADVISE > page_entry *p, *start_p; > char *start; >@@ -1061,6 +1062,7 @@ release_pages (void) > else > G.free_pages = p; > G.bytes_mapped -= mapped_len; >+ n1 += len; > continue; > } > prev = newprev; >@@ -1092,6 +1094,7 @@ release_pages (void) >/* Don't count those pages as mapped to not touch the garbage collector > unnecessarily. */ > G.bytes_mapped -= len; >+ n2 += len; > while (start_p != p) > { > start_p->discarded = true; >@@ -1124,6 +1127,7 @@ release_pages (void) > } > > munmap (start, len); >+ n1 += len; > G.bytes_mapped -= len; > } > >@@ -1152,10 +1156,20 @@ release_pages (void) > *gp = g->next; > G.bytes_mapped -= g->alloc_size; > free (g->allocation); >+ n1 += g->alloc_size; > } > else > gp = &g->next; > #endif >+ if (!quiet_flag && (n1 || n2)) >+ { >+ fprintf (stderr, " {GC"); >+ if (n1) >+ fprintf (stderr, " released %luk", (unsigned long)(n1 / 1024)); >+ if (n2) >+ fprintf (stderr, " madv_dontneed %luk", (unsigned long)(n2 / 1024)); >+ fprintf (stderr, "}"); >+ } > } > > /* This table provides a fast way to determine ceil(log_2(size)) for >@@ -2178,19 +2192,22 @@ ggc_collect (void) > return; > > timevar_push (TV_GC); >- if (!quiet_flag) >- fprintf (stderr, " {GC %luk -> ", (unsigned long) G.allocated / >1024); > if (GGC_DEBUG_LEVEL >= 2) > fprintf (G.debug_file, "BEGIN COLLECTING\n"); > > /* Zero the total allocated bytes. This will be recalculated in the > sweep phase. */ >+ size_t allocated = G.allocated; > G.allocated = 0; > > /* Release the pages we freed the last time we collected, but didn't > reuse in the interim. */ > release_pages (); > >+ /* Output this later so we do not interfere with release_pages. */ >+ if (!quiet_flag) >+ fprintf (stderr, " {GC %luk -> ", (unsigned long) allocated / >1024); >+ > /* Indicate that we've seen collections at this context depth. */ >G.context_depth_collections = ((unsigned long)1 << (G.context_depth + >1)) - 1; > >@@ -2221,9 +2238,25 @@ ggc_collect (void) > fprintf (G.debug_file, "END COLLECTING\n"); > } > >-/* Assume that all GGC memory is reachable and grow the limits for >next collection. >- With checking, trigger GGC so -Q compilation outputs how much of >memory really is >- reachable. */ >+/* Return free pages to the system. */ >+ >+void >+ggc_trim () >+{ >+ timevar_push (TV_GC); >+ G.allocated = 0; >+ sweep_pages (); >+ release_pages (); >+ if (!quiet_flag) >+ fprintf (stderr, " {GC trimmed to %luk, %luk mapped}", >+ (unsigned long) G.allocated / 1024, >+ (unsigned long) G.bytes_mapped / 1024); >+ timevar_pop (TV_GC); >+} >+ >+/* Assume that all GGC memory is reachable and grow the limits for >next >+ collection. With checking, trigger GGC so -Q compilation outputs >how much >+ of memory really is reachable. */ > > void > ggc_grow (void) >Index: ggc-none.c >=================================================================== >--- ggc-none.c (revision 276707) >+++ ggc-none.c (working copy) >@@ -72,3 +72,8 @@ void > ggc_grow (void) > { > } >+ >+void >+ggc_trim (void) >+{ >+} >Index: lto/lto.c >=================================================================== >--- lto/lto.c (revision 276707) >+++ lto/lto.c (working copy) >@@ -304,6 +306,7 @@ lto_wpa_write_files (void) > > timevar_push (TV_WHOPR_WPA_IO); > >+ ggc_trim (); > /* Generate a prefix for the LTRANS unit files. */ > blen = strlen (ltrans_output_list); > temp_filename = (char *) xmalloc (blen + sizeof ("2147483648.o"));
On Fri, Oct 11, 2019 at 09:03:53AM +0200, Jan Hubicka wrote: > Bootstrapped/regtested x86_64-linux, OK? > > * ggc-page.c (release_pages): Output statistics when !quiet_flag. > (ggc_collect): Dump later to not interfere with release_page dump. > (ggc_trim): New function. > * ggc-none.c (ggc_trim): New. > > * lto.c (lto_wpa_write_files): Call ggc_trim. > @@ -1152,10 +1156,20 @@ release_pages (void) > *gp = g->next; > G.bytes_mapped -= g->alloc_size; > free (g->allocation); > + n1 += g->alloc_size; > } > else > gp = &g->next; > #endif This broke !defined(USING_MMAP) support, the second g->alloc_size read is after the memory containing *g is freed. Fixed thusly, tested with #undef USING_MMAP in the file (without the patch self-test ICEs, with it succeeds), committed to trunk as obvious. 2019-10-18 Jakub Jelinek <jakub@redhat.com> PR middle-end/92153 * ggc-page.c (release_pages): Read g->alloc_size before free rather than after it. --- gcc/ggc-page.c.jj 2019-10-11 14:10:44.987386981 +0200 +++ gcc/ggc-page.c 2019-10-18 19:13:59.458085610 +0200 @@ -1155,8 +1155,8 @@ release_pages (void) { *gp = g->next; G.bytes_mapped -= g->alloc_size; - free (g->allocation); n1 += g->alloc_size; + free (g->allocation); } else gp = &g->next; Jakub
Index: ggc-page.c =================================================================== --- ggc-page.c (revision 276707) +++ ggc-page.c (working copy) @@ -529,7 +529,6 @@ static void clear_page_group_in_use (pag #endif static struct page_entry * alloc_page (unsigned); static void free_page (struct page_entry *); -static void release_pages (void); static void clear_marks (void); static void sweep_pages (void); static void ggc_recalculate_in_use_p (page_entry *); @@ -1016,6 +1015,8 @@ free_page (page_entry *entry) static void release_pages (void) { + size_t n1 = 0; + size_t n2 = 0; #ifdef USING_MADVISE page_entry *p, *start_p; char *start; @@ -1061,6 +1062,7 @@ release_pages (void) else G.free_pages = p; G.bytes_mapped -= mapped_len; + n1 += len; continue; } prev = newprev; @@ -1092,6 +1094,7 @@ release_pages (void) /* Don't count those pages as mapped to not touch the garbage collector unnecessarily. */ G.bytes_mapped -= len; + n2 += len; while (start_p != p) { start_p->discarded = true; @@ -1124,6 +1127,7 @@ release_pages (void) } munmap (start, len); + n1 += len; G.bytes_mapped -= len; } @@ -1152,10 +1156,20 @@ release_pages (void) *gp = g->next; G.bytes_mapped -= g->alloc_size; free (g->allocation); + n1 += g->alloc_size; } else gp = &g->next; #endif + if (!quiet_flag && (n1 || n2)) + { + fprintf (stderr, " {GC"); + if (n1) + fprintf (stderr, " released %luk", (unsigned long)(n1 / 1024)); + if (n2) + fprintf (stderr, " madv_dontneed %luk", (unsigned long)(n2 / 1024)); + fprintf (stderr, "}"); + } } /* This table provides a fast way to determine ceil(log_2(size)) for @@ -2178,19 +2192,22 @@ ggc_collect (void) return; timevar_push (TV_GC); - if (!quiet_flag) - fprintf (stderr, " {GC %luk -> ", (unsigned long) G.allocated / 1024); if (GGC_DEBUG_LEVEL >= 2) fprintf (G.debug_file, "BEGIN COLLECTING\n"); /* Zero the total allocated bytes. This will be recalculated in the sweep phase. */ + size_t allocated = G.allocated; G.allocated = 0; /* Release the pages we freed the last time we collected, but didn't reuse in the interim. */ release_pages (); + /* Output this later so we do not interfere with release_pages. */ + if (!quiet_flag) + fprintf (stderr, " {GC %luk -> ", (unsigned long) allocated / 1024); + /* Indicate that we've seen collections at this context depth. */ G.context_depth_collections = ((unsigned long)1 << (G.context_depth + 1)) - 1; @@ -2221,9 +2238,25 @@ ggc_collect (void) fprintf (G.debug_file, "END COLLECTING\n"); } -/* Assume that all GGC memory is reachable and grow the limits for next collection. - With checking, trigger GGC so -Q compilation outputs how much of memory really is - reachable. */ +/* Return free pages to the system. */ + +void +ggc_trim () +{ + timevar_push (TV_GC); + G.allocated = 0; + sweep_pages (); + release_pages (); + if (!quiet_flag) + fprintf (stderr, " {GC trimmed to %luk, %luk mapped}", + (unsigned long) G.allocated / 1024, + (unsigned long) G.bytes_mapped / 1024); + timevar_pop (TV_GC); +} + +/* Assume that all GGC memory is reachable and grow the limits for next + collection. With checking, trigger GGC so -Q compilation outputs how much + of memory really is reachable. */ void ggc_grow (void) Index: ggc-none.c =================================================================== --- ggc-none.c (revision 276707) +++ ggc-none.c (working copy) @@ -72,3 +72,8 @@ void ggc_grow (void) { } + +void +ggc_trim (void) +{ +} Index: lto/lto.c =================================================================== --- lto/lto.c (revision 276707) +++ lto/lto.c (working copy) @@ -304,6 +306,7 @@ lto_wpa_write_files (void) timevar_push (TV_WHOPR_WPA_IO); + ggc_trim (); /* Generate a prefix for the LTRANS unit files. */ blen = strlen (ltrans_output_list); temp_filename = (char *) xmalloc (blen + sizeof ("2147483648.o"));