[gambit-list] parallel GC

Marc Feeley feeley at iro.umontreal.ca
Mon Oct 31 16:25:51 EDT 2016


The CPU time seems correct to me. With multiple threaded VMs the GC takes 25 seconds more CPU time than with single threaded VMs (where it takes 13 seconds).  This is probably because there is little parallelism that the GC can exploit (perhaps a small number of very large bignums?).

xactlog’s user interface is rather old-school (the program was originally written in the early 90’s!).  Here’s a quick summary of the commands, which you can get with the -? option:

COMMANDS:

keyboard: q -- quit
          . -- redraw window
          e -- generate EPS file <program_name>.eps
          0..9 -- pass through smoothing filter 'n' times
          t -- toggle title on/off
          s -- toggle state logs on/off
          a -- toggle activity profile on/off
          p -- toggle percent profile on/off
          h -- toggle state duration histograms on/off
          l -- toggle state labels on/off
          f -- toggle first state only on/off
          i -- revert to initial display
          z -- zoom out by factor of 5

mouse: to zoom in, use click&drag to select section of interest
       to change state ordering, use click&drag on state names to swap states

By the way, your screendump is strange… the middle part, the “percent profile”, should be a summation (put on a scale of 100%) of the activity bars above (one per processor).  But the red areas (for the GC) don’t line up…  Could you try to run xactlog with the -us or -ms option (to scale to microseconds or milliseconds)?

Marc


> On Oct 31, 2016, at 4:11 PM, Bradley Lucier <lucier at math.purdue.edu> wrote:
> 
> On 10/25/2016 02:41 PM, Marc Feeley wrote:
>> Gambit has reached a milestone today with the implementation of the
>> parallel garbage collector.  This is an important step towards truly
>> concurrent Gambit threads.
>> 
>> When Gambit is configured with --enable-multiple-threaded-vms the
>> available processors will cooperate to do the garbage collection in
>> parallel.
> 
> Can we trust what the "time" macro reports?  The reported CPU time seems too big.  For an example, I have the following example, after
> 
> > (##gc-report-set! #t)
> 
> I get for one routine (refine t) with one thread:
> 
> *** GC: 1281 ms, 18.1G alloc, 5.87G heap, 2.58G live (44% 2591854144+175273048)
> *** GC: 2983 ms, 21.0G alloc, 12.3G heap, 5.37G live (44% 5593837200+175273048)
> *** GC: 4395 ms, 26.9G alloc, 16.6G heap, 7.27G live (44% 7500855696+309556416)
> *** GC: 4676 ms, 35.0G alloc, 19.9G heap, 8.73G live (44% 8996104560+376665328)
> (time (refine t))
>    30711 ms real time
>    30692 ms cpu time (28784 user, 1908 system)
>    4 collections accounting for 13334 ms real time (13020 user, 304 system)
>    26621750736 bytes allocated
>    2909823 minor faults
>    no major faults
> 
> and with multiple vms:
> 
> *** GC: 405 ms, 18.0G alloc, 3.70G heap, 1.62G live (44% 1565082624+175273520)
> *** GC: 728 ms, 19.8G alloc, 7.61G heap, 3.33G live (44% 3403627936+175273520)
> *** GC: 1503 ms, 23.6G alloc, 16.1G heap, 7.04G live (44% 7382849808+175273520)
> *** GC: 1815 ms, 31.3G alloc, 17.6G heap, 7.71G live (44% 7903774784+376665800)
> *** GC: 1831 ms, 39.8G alloc, 17.9G heap, 7.83G live (44% 8026441792+376665800)
> (time (refine t))
>    26083 ms real time
>    59196 ms cpu time (56960 user, 2236 system)
>    5 collections accounting for 6281 ms real time (38660 user, 556 system)
>    26621756528 bytes allocated
>    2993944 minor faults
>    no major faults
> 
> The whole run is listed below.
> 
> BTW, what is xactlog supposed to do?  It generates a pretty picture for me, but the app doesn't seem to be aware of high resolution screens (at least not on Ubuntu 16.10).  And it has a user interface for something?
> 
> You may be interested in a screenshot for this job, which I attach.  I think the big GCs are for (refine t).
> 
> Brad
> 
> Multiple VMs, default p (=8?)
> 
> > (load "final-tests")
> *** GC: 5 ms, 2.04G alloc, 68.3M heap, 23.2M live (34% 24079744+293688)
> *** GC: 7 ms, 2.06G alloc, 110M heap, 41.9M live (38% 43639200+297880)
> *** GC: 8 ms, 2.11G alloc, 121M heap, 46.6M live (39% 47804496+1088560)
> *** GC: 13 ms, 2.16G alloc, 164M heap, 65.3M live (40% 67411712+1096848)
> *** GC: 28 ms, 2.23G alloc, 303M heap, 126M live (42% 130383664+1625288)
> *** GC: 27 ms, 2.37G alloc, 325M heap, 136M live (42% 138275568+4250920)
> *** GC: 26 ms, 2.52G alloc, 274M heap, 114M live (41% 114838272+4267400)
> *** GC: 53 ms, 2.65G alloc, 552M heap, 235M live (43% 242228992+4267400)
> *** GC: 101 ms, 2.91G alloc, 1.02G heap, 450M live (43% 465008880+6372800)
> *** GC: 104 ms, 3.40G alloc, 1.08G heap, 480M live (43% 486840896+16866848)
> *** GC: 106 ms, 3.92G alloc, 1.11G heap, 493M live (43% 499677392+16866848)
> *** GC: 162 ms, 4.46G alloc, 1.68G heap, 745M live (43% 764363072+16899712)
> *** GC: 402 ms, 5.28G alloc, 3.52G heap, 1.54G live (44% 1632889712+16899712)
> *** GC: 389 ms, 6.96G alloc, 4.11G heap, 1.80G live (44% 1861163296+67264280)
> *** GC: 421 ms, 8.97G alloc, 4.59G heap, 2.00G live (44% 2085367712+67264280)
> (time (Triangulation->refined-triangulations coarse-square-triangulation number-of-refinements))
>    7986 ms real time
>    17880 ms cpu time (17200 user, 680 system)
>    15 collections accounting for 1852 ms real time (11484 user, 192 system)
>    8734058104 bytes allocated
>    939736 minor faults
>    no major faults
> *** GC: 337 ms, 11.2G alloc, 3.12G heap, 1.36G live (44% 1387891648+72953880)
> (time (triangulations->linear-finite-element-spaces triangulations))
>    1601 ms real time
>    3232 ms cpu time (3168 user, 64 system)
>    1 collection accounting for 337 ms real time (1956 user, 0 system)
>    2705639712 bytes allocated
>    125528 minor faults
>    no major faults
> *** GC: 400 ms, 12.7G alloc, 3.35G heap, 1.47G live (44% 1483446784+90354056)
> *** GC: 345 ms, 14.3G alloc, 3.45G heap, 1.51G live (44% 1496393344+127273872)
> (time (linear-finite-element-spaces->projectors spaces))
>    1647 ms real time
>    5216 ms cpu time (5212 user, 4 system)
>    2 collections accounting for 745 ms real time (4304 user, 4 system)
>    2643920832 bytes allocated
>    8446 minor faults
>    no major faults
> *** GC: 415 ms, 16.0G alloc, 4.31G heap, 1.89G live (44% 1872364048+155321840)
> (time (projectors->injectors projectors))
>    862 ms real time
>    2888 ms cpu time (2884 user, 4 system)
>    1 collection accounting for 415 ms real time (2436 user, 4 system)
>    1924367344 bytes allocated
>    9515 minor faults
>    no major faults
> (time (Triangulation-consistent? t))
>    1163 ms real time
>    1176 ms cpu time (1164 user, 12 system)
>    no collections
>    1107886128 bytes allocated
>    12769 minor faults
>    no major faults
> *** GC: 405 ms, 18.0G alloc, 3.70G heap, 1.62G live (44% 1565082624+175273520)
> *** GC: 728 ms, 19.8G alloc, 7.61G heap, 3.33G live (44% 3403627936+175273520)
> *** GC: 1503 ms, 23.6G alloc, 16.1G heap, 7.04G live (44% 7382849808+175273520)
> *** GC: 1815 ms, 31.3G alloc, 17.6G heap, 7.71G live (44% 7903774784+376665800)
> *** GC: 1831 ms, 39.8G alloc, 17.9G heap, 7.83G live (44% 8026441792+376665800)
> (time (refine t))
>    26083 ms real time
>    59196 ms cpu time (56960 user, 2236 system)
>    5 collections accounting for 6281 ms real time (38660 user, 556 system)
>    26621756528 bytes allocated
>    2993944 minor faults
>    no major faults
> *** GC: 1761 ms, 48.5G alloc, 13.9G heap, 6.09G live (44% 6160831792+376796968)
> (time (Triangulation->Linear-finite-element-space fine-t))
>    5257 ms real time
>    13948 ms cpu time (13644 user, 304 system)
>    1 collection accounting for 1761 ms real time (10272 user, 140 system)
>    8124499456 bytes allocated
>    218671 minor faults
>    no major faults
> (time (->Linear-finite-element-operator fine-fes))
>    1383 ms real time
>    1400 ms cpu time (1360 user, 40 system)
>    no collections
>    3961212352 bytes allocated
>    31399 minor faults
>    no major faults
> *** GC: 1434 ms, 55.0G alloc, 14.2G heap, 6.25G live (44% 5881756160+830027848)
> (time (add-mass-terms! L2-inner-product-operator (lambda (p) 1.)))
>    9280 ms real time
>    16284 ms cpu time (16244 user, 40 system)
>    1 collection accounting for 1434 ms real time (8356 user, 12 system)
>    8187287360 bytes allocated
>    24626 minor faults
>    no major faults
> *** GC: 1512 ms, 61.9G alloc, 14.8G heap, 6.53G live (44% 6178961760+830027848)
> (time (->Linear-finite-element-operator fine-fes))
>    2842 ms real time
>    9856 ms cpu time (9828 user, 28 system)
>    1 collection accounting for 1512 ms real time (8512 user, 0 system)
>    3961218320 bytes allocated
>    5676 minor faults
>    no major faults
> *** GC: 1462 ms, 68.9G alloc, 15.0G heap, 6.61G live (44% 5881756992+1216084216)
> *** GC: 1428 ms, 76.2G alloc, 15.0G heap, 6.61G live (44% 5881756816+1216084216)
> (time (add-stiffness-terms! H1-inner-product-operator (lambda (p) 1.)))
>    11172 ms real time
>    25156 ms cpu time (25100 user, 56 system)
>    2 collections accounting for 2890 ms real time (16756 user, 32 system)
>    16106139008 bytes allocated
>    52699 minor faults
>    no major faults
> *** GC: 1507 ms, 83.5G alloc, 15.0G heap, 6.61G live (44% 5881756640+1216084216)
> (time (add-mass-terms! H1-inner-product-operator (lambda (p) 1.)))
>    9234 ms real time
>    16784 ms cpu time (16760 user, 24 system)
>    1 collection accounting for 1507 ms real time (8968 user, 20 system)
>    8187287264 bytes allocated
>    15808 minor faults
>    no major faults
> *** GC: 1731 ms, 90.7G alloc, 17.9G heap, 7.89G live (44% 7230882624+1245501928)
> *** GC: 1757 ms, 99.2G alloc, 19.0G heap, 8.39G live (44% 7596419264+1413405384)
> (time (Operator-adjoint (coarse<-fine fes fine-fes)))
>    7716 ms real time
>    25084 ms cpu time (24828 user, 256 system)
>    2 collections accounting for 3489 ms real time (20772 user, 32 system)
>    13733230480 bytes allocated
>    332422 minor faults
>    no major faults
> 
> Multiple threads, p=4:
> 
> heine:~/WWW/615-2016/software> gsc -:d2,p4
> *** GC: 0 ms, 393K alloc, 14.0M heap, 66.5K live (0% 66448+1600)
> Gambit v4.8.5
> 
> > (load "_meroon")
> [ Meroon V3 Paques2001+1 $Revision: 1.2 $ ]
> "/home/lucier/WWW/615-2016/software/_meroon.o4"
> > (compile-file "all")
> *** GC: 1 ms, 4.84M alloc, 13.0M heap, 1.91M live (15% 1986848+20496)
> *** GC: 1 ms, 7.94M alloc, 17.0M heap, 3.72M live (22% 3885008+20496)
> *** GC: 1 ms, 13.2M alloc, 25.0M heap, 6.91M live (28% 7228144+20496)
> *** GC: 2 ms, 22.9M alloc, 38.0M heap, 12.9M live (34% 13471632+20496)
> *** GC: 4 ms, 38.4M alloc, 60.0M heap, 22.2M live (37% 23251200+20496)
> *** GC: 8 ms, 65.9M alloc, 92.0M heap, 36.2M live (39% 37927760+20496)
> *** GC: 12 ms, 108M alloc, 119M heap, 48.3M live (41% 50594016+20496)
> *** GC: 14 ms, 163M alloc, 146M heap, 59.8M live (41% 62633312+20496)
> *** GC: 16 ms, 232M alloc, 161M heap, 66.6M live (41% 69838688+32880)
> *** GC: 21 ms, 308M alloc, 221M heap, 92.8M live (42% 97220336+69776)
> *** GC: 21 ms, 412M alloc, 253M heap, 107M live (42% 111656528+514392)
> *** GC: 19 ms, 534M alloc, 253M heap, 107M live (42% 111655840+514392)
> *** GC: 18 ms, 656M alloc, 253M heap, 107M live (42% 111657504+514392)
> *** GC: 18 ms, 778M alloc, 253M heap, 107M live (42% 111657280+514392)
> *** GC: 20 ms, 900M alloc, 253M heap, 107M live (42% 111656736+514392)
> "/home/lucier/WWW/615-2016/software/all.o4"
> > (load "all")
> "/home/lucier/WWW/615-2016/software/all.o4"
> > (compile-file "final-tests")
> *** GC: 19 ms, 1020M alloc, 191M heap, 79.6M live (42% 83380752+37608)
> *** GC: 5 ms, 1.08G alloc, 71.2M heap, 27.0M live (38% 28153712+177832)
> *** GC: 4 ms, 1.12G alloc, 71.2M heap, 27.0M live (38% 28154064+177832)
> "/home/lucier/WWW/615-2016/software/final-tests.o3"
> > (load "final-tests")
> *** GC: 4 ms, 1.15G alloc, 54.3M heap, 19.7M live (36% 20391840+297712)
> *** GC: 10 ms, 1.17G alloc, 107M heap, 43.1M live (40% 44923136+297712)
> *** GC: 9 ms, 1.22G alloc, 116M heap, 46.9M live (40% 48133072+1088392)
> *** GC: 15 ms, 1.27G alloc, 166M heap, 68.7M live (41% 70946992+1096680)
> *** GC: 29 ms, 1.35G alloc, 294M heap, 124M live (42% 128814192+1625120)
> *** GC: 28 ms, 1.48G alloc, 324M heap, 138M live (43% 140675600+4250752)
> *** GC: 26 ms, 1.63G alloc, 278M heap, 118M live (42% 119649344+4267232)
> *** GC: 55 ms, 1.76G alloc, 564M heap, 243M live (43% 250798928+4267232)
> *** GC: 113 ms, 2.03G alloc, 1.03G heap, 458M live (43% 473688240+6372632)
> *** GC: 106 ms, 2.52G alloc, 1.05G heap, 469M live (43% 474954128+16866680)
> *** GC: 102 ms, 3.03G alloc, 1.11G heap, 494M live (44% 500901968+16866680)
> *** GC: 177 ms, 3.57G alloc, 1.69G heap, 754M live (44% 773812656+16899544)
> *** GC: 388 ms, 4.40G alloc, 3.56G heap, 1.56G live (44% 1653302480+16899544)
> *** GC: 450 ms, 6.11G alloc, 4.12G heap, 1.80G live (44% 1868049120+67264112)
> *** GC: 418 ms, 8.12G alloc, 4.60G heap, 2.01G live (44% 2091404448+67264112)
> (time (Triangulation->refined-triangulations coarse-square-triangulation number-of-refinements))
>    7497 ms real time
>    13172 ms cpu time (12512 user, 660 system)
>    15 collections accounting for 1931 ms real time (7404 user, 172 system)
>    8734066368 bytes allocated
>    945119 minor faults
>    no major faults
> *** GC: 349 ms, 10.3G alloc, 3.13G heap, 1.37G live (44% 1395876048+72953712)
> (time (triangulations->linear-finite-element-spaces triangulations))
>    1587 ms real time
>    2548 ms cpu time (2452 user, 96 system)
>    1 collection accounting for 349 ms real time (1304 user, 0 system)
>    2705638608 bytes allocated
>    122104 minor faults
>    no major faults
> *** GC: 366 ms, 11.9G alloc, 3.39G heap, 1.49G live (44% 1504691264+90936280)
> *** GC: 355 ms, 13.5G alloc, 3.61G heap, 1.58G live (44% 1573948112+127273704)
> (time (linear-finite-element-spaces->projectors spaces))
>    1616 ms real time
>    3612 ms cpu time (3612 user, 0 system)
>    2 collections accounting for 721 ms real time (2716 user, 0 system)
>    2643920928 bytes allocated
>    9262 minor faults
>    no major faults
> *** GC: 444 ms, 15.2G alloc, 4.25G heap, 1.86G live (44% 1854603312+144811224)
> (time (projectors->injectors projectors))
>    892 ms real time
>    2144 ms cpu time (2136 user, 8 system)
>    1 collection accounting for 444 ms real time (1692 user, 0 system)
>    1924367392 bytes allocated
>    11027 minor faults
>    no major faults
> (time (Triangulation-consistent? t))
>    1192 ms real time
>    1196 ms cpu time (1184 user, 12 system)
>    no collections
>    1107886128 bytes allocated
>    11379 minor faults
>    no major faults
> *** GC: 434 ms, 17.3G alloc, 3.99G heap, 1.75G live (44% 1704991024+175273352)
> *** GC: 867 ms, 19.2G alloc, 8.19G heap, 3.59G live (44% 3679600192+175273352)
> *** GC: 1737 ms, 23.2G alloc, 17.0G heap, 7.44G live (44% 7811269968+175273352)
> *** GC: 1738 ms, 31.4G alloc, 17.9G heap, 7.86G live (44% 8062752528+376665632)
> *** GC: 1745 ms, 40.1G alloc, 17.8G heap, 7.80G live (44% 7997456944+376665632)
> (time (refine t))
>    24133 ms real time
>    43372 ms cpu time (41316 user, 2056 system)
>    5 collections accounting for 6520 ms real time (25328 user, 352 system)
>    26621756432 bytes allocated
>    3128283 minor faults
>    no major faults
> *** GC: 1661 ms, 48.8G alloc, 14.3G heap, 6.26G live (44% 6341580544+376796800)
> (time (Triangulation->Linear-finite-element-space fine-t))
>    5030 ms real time
>    9652 ms cpu time (9468 user, 184 system)
>    1 collection accounting for 1661 ms real time (6164 user, 100 system)
>    8124497792 bytes allocated
>    120281 minor faults
>    no major faults
> (time (->Linear-finite-element-operator fine-fes))
>    1346 ms real time
>    1352 ms cpu time (1308 user, 44 system)
>    no collections
>    3961212352 bytes allocated
>    24563 minor faults
>    no major faults
> *** GC: 1623 ms, 55.4G alloc, 14.2G heap, 6.25G live (44% 5881748768+830027680)
> (time (add-mass-terms! L2-inner-product-operator (lambda (p) 1.)))
>    9759 ms real time
>    13796 ms cpu time (13768 user, 28 system)
>    1 collection accounting for 1623 ms real time (5608 user, 16 system)
>    8187287360 bytes allocated
>    35581 minor faults
>    no major faults
> *** GC: 1583 ms, 62.4G alloc, 17.5G heap, 7.68G live (44% 7415091200+830027680)
> (time (->Linear-finite-element-operator fine-fes))
>    2941 ms real time
>    7188 ms cpu time (7160 user, 28 system)
>    1 collection accounting for 1583 ms real time (5804 user, 16 system)
>    3961218448 bytes allocated
>    13304 minor faults
>    no major faults
> *** GC: 1628 ms, 70.6G alloc, 15.0G heap, 6.61G live (44% 5881749040+1216084048)
> *** GC: 1536 ms, 77.9G alloc, 15.0G heap, 6.61G live (44% 5881749616+1216084048)
> (time (add-stiffness-terms! H1-inner-product-operator (lambda (p) 1.)))
>    11895 ms real time
>    19696 ms cpu time (19456 user, 240 system)
>    2 collections accounting for 3164 ms real time (10872 user, 52 system)
>    16106139008 bytes allocated
>    372094 minor faults
>    no major faults
> *** GC: 1523 ms, 85.2G alloc, 15.0G heap, 6.61G live (44% 5881749584+1216084048)
> (time (add-mass-terms! H1-inner-product-operator (lambda (p) 1.)))
>    9586 ms real time
>    13332 ms cpu time (13312 user, 20 system)
>    1 collection accounting for 1523 ms real time (5224 user, 12 system)
>    8187287264 bytes allocated
>    8 minor faults
>    no major faults
> *** GC: 1656 ms, 92.4G alloc, 17.4G heap, 7.69G live (44% 6935764784+1321048480)
> (time (Operator-adjoint (coarse<-fine fes fine-fes)))
>    5801 ms real time
>    10272 ms cpu time (10148 user, 124 system)
>    1 collection accounting for 1656 ms real time (6084 user, 24 system)
>    13733224704 bytes allocated
>    162321 minor faults
>    no major faults
> "/home/lucier/WWW/615-2016/software/final-tests.o3"
> 
> 
> With a single thread:
> 
> > (load "final-tests")
> *** GC: 18 ms, 1.14G alloc, 52.3M heap, 21.6M live (41% 22353728+293216)
> *** GC: 27 ms, 1.17G alloc, 94.3M heap, 39.8M live (42% 41461296+297408)
> *** GC: 27 ms, 1.21G alloc, 115M heap, 48.9M live (42% 50175616+1088088)
> *** GC: 33 ms, 1.26G alloc, 146M heap, 62.6M live (43% 64544192+1096376)
> *** GC: 69 ms, 1.33G alloc, 293M heap, 127M live (43% 131895792+1096376)
> *** GC: 67 ms, 1.47G alloc, 298M heap, 129M live (43% 131426272+4250448)
> *** GC: 48 ms, 1.61G alloc, 220M heap, 95.1M live (43% 95459424+4266928)
> *** GC: 104 ms, 1.71G alloc, 452M heap, 197M live (43% 201926432+4266928)
> *** GC: 232 ms, 1.93G alloc, 952M heap, 415M live (44% 431276816+4266928)
> *** GC: 259 ms, 2.38G alloc, 1.07G heap, 479M live (44% 485131008+16866376)
> *** GC: 241 ms, 2.89G alloc, 1.03G heap, 461M live (44% 466692528+16866376)
> *** GC: 347 ms, 3.40G alloc, 1.31G heap, 586M live (44% 597450368+16899240)
> *** GC: 679 ms, 4.04G alloc, 2.73G heap, 1.19G live (44% 1264461152+16899240)
> *** GC: 1114 ms, 5.37G alloc, 4.41G heap, 1.93G live (44% 2048148688+25304288)
> *** GC: 1046 ms, 7.50G alloc, 4.75G heap, 2.08G live (44% 2166182800+67263808)
> (time (Triangulation->refined-triangulations coarse-square-triangulation number-of-refinements))
>    9596 ms real time
>    9592 ms cpu time (8864 user, 728 system)
>    15 collections accounting for 4309 ms real time (4164 user, 136 system)
>    8734068864 bytes allocated
>    1044564 minor faults
>    no major faults
> *** GC: 950 ms, 9.80G alloc, 2.91G heap, 1.28G live (44% 1302125056+68742608)
> *** GC: 918 ms, 11.2G alloc, 3.42G heap, 1.50G live (44% 1535646544+72953408)
> (time (triangulations->linear-finite-element-spaces triangulations))
>    2985 ms real time
>    2984 ms cpu time (2936 user, 48 system)
>    2 collections accounting for 1869 ms real time (1856 user, 8 system)
>    2705643120 bytes allocated
>    76222 minor faults
>    no major faults
> *** GC: 902 ms, 12.9G alloc, 3.70G heap, 1.63G live (44% 1639180976+106268952)
> (time (linear-finite-element-spaces->projectors spaces))
>    1727 ms real time
>    1728 ms cpu time (1716 user, 12 system)
>    1 collection accounting for 902 ms real time (896 user, 8 system)
>    2643915456 bytes allocated
>    13273 minor faults
>    no major faults
> *** GC: 880 ms, 14.6G alloc, 3.70G heap, 1.63G live (44% 1606750224+141128560)
> (time (projectors->injectors projectors))
>    1346 ms real time
>    1344 ms cpu time (1332 user, 12 system)
>    1 collection accounting for 880 ms real time (880 user, 0 system)
>    1924367280 bytes allocated
>    6543 minor faults
>    no major faults
> *** GC: 828 ms, 16.4G alloc, 3.51G heap, 1.54G live (44% 1483272064+175273048)
> (time (Triangulation-consistent? t))
>    1713 ms real time
>    1712 ms cpu time (1708 user, 4 system)
>    1 collection accounting for 828 ms real time (824 user, 4 system)
>    1107891680 bytes allocated
>    2337 minor faults
>    no major faults
> *** GC: 1281 ms, 18.1G alloc, 5.87G heap, 2.58G live (44% 2591854144+175273048)
> *** GC: 2983 ms, 21.0G alloc, 12.3G heap, 5.37G live (44% 5593837200+175273048)
> *** GC: 4395 ms, 26.9G alloc, 16.6G heap, 7.27G live (44% 7500855696+309556416)
> *** GC: 4676 ms, 35.0G alloc, 19.9G heap, 8.73G live (44% 8996104560+376665328)
> (time (refine t))
>    30711 ms real time
>    30692 ms cpu time (28784 user, 1908 system)
>    4 collections accounting for 13334 ms real time (13020 user, 304 system)
>    26621750736 bytes allocated
>    2909823 minor faults
>    no major faults
> *** GC: 3722 ms, 44.6G alloc, 12.9G heap, 5.65G live (44% 5685403968+376796496)
> (time (Triangulation->Linear-finite-element-space fine-t))
>    7027 ms real time
>    7020 ms cpu time (6664 user, 356 system)
>    1 collection accounting for 3722 ms real time (3584 user, 136 system)
>    8124501472 bytes allocated
>    382531 minor faults
>    no major faults
> *** GC: 3857 ms, 50.9G alloc, 16.4G heap, 7.19G live (44% 7276603408+443971008)
> (time (->Linear-finite-element-operator fine-fes))
>    5247 ms real time
>    5244 ms cpu time (5188 user, 56 system)
>    1 collection accounting for 3857 ms real time (3844 user, 12 system)
>    3961218352 bytes allocated
>    15427 minor faults
>    no major faults
> *** GC: 3282 ms, 58.6G alloc, 14.2G heap, 6.25G live (44% 5881720288+830027376)
> (time (add-mass-terms! L2-inner-product-operator (lambda (p) 1.)))
>    11088 ms real time
>    11084 ms cpu time (10808 user, 276 system)
>    1 collection accounting for 3282 ms real time (3252 user, 28 system)
>    8187287360 bytes allocated
>    376237 minor faults
>    no major faults
> (time (->Linear-finite-element-operator fine-fes))
>    1380 ms real time
>    1380 ms cpu time (1348 user, 32 system)
>    no collections
>    3961212144 bytes allocated
>    10256 minor faults
>    no major faults
> *** GC: 2841 ms, 65.2G alloc, 15.0G heap, 6.61G live (44% 5881720848+1216083744)
> *** GC: 2844 ms, 72.5G alloc, 15.0G heap, 6.61G live (44% 5881721024+1216083744)
> (time (add-stiffness-terms! H1-inner-product-operator (lambda (p) 1.)))
>    13722 ms real time
>    13712 ms cpu time (13608 user, 104 system)
>    2 collections accounting for 5685 ms real time (5672 user, 8 system)
>    16106139008 bytes allocated
>    140337 minor faults
>    no major faults
> *** GC: 2831 ms, 79.8G alloc, 15.0G heap, 6.61G live (44% 5881720816+1216083744)
> (time (add-mass-terms! H1-inner-product-operator (lambda (p) 1.)))
>    10198 ms real time
>    10192 ms cpu time (10136 user, 56 system)
>    1 collection accounting for 2831 ms real time (2832 user, 0 system)
>    8187287264 bytes allocated
>    53093 minor faults
>    no major faults
> *** GC: 2932 ms, 87.1G alloc, 15.1G heap, 6.68G live (44% 5954874464+1216083744)
> *** GC: 3748 ms, 94.4G alloc, 18.1G heap, 8.01G live (44% 7242602944+1358837992)
> (time (Operator-adjoint (coarse<-fine fes fine-fes)))
>    10323 ms real time
>    10320 ms cpu time (10172 user, 148 system)
>    2 collections accounting for 6680 ms real time (6600 user, 80 system)
>    13733231104 bytes allocated
>    28553 minor faults
>    no major faults
> "/home/lucier/WWW/615-2016/software/final-tests.o2"
> >
> 
> <Screenshot from 2016-10-31 15-10-27.png>_______________________________________________
> Gambit-list mailing list
> Gambit-list at iro.umontreal.ca
> https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list




More information about the Gambit-list mailing list