{"id":2080,"date":"2024-03-16T09:41:04","date_gmt":"2024-03-16T09:41:04","guid":{"rendered":"https:\/\/mvermeulen.org\/perf\/?page_id=2080"},"modified":"2024-03-17T19:14:10","modified_gmt":"2024-03-17T19:14:10","slug":"draco","status":"publish","type":"page","link":"https:\/\/mvermeulen.org\/perf\/workloads\/phoronix\/draco\/","title":{"rendered":"draco"},"content":{"rendered":"\n<p>Testing the draco library which compresses\/decompresses 3d geometric meshes and point clouds. There are two workloads. The benchmark looks to be single-threaded and runs quickly.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1280\" height=\"960\" src=\"https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/systemtime-31.png\" alt=\"\" class=\"wp-image-2089\" srcset=\"https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/systemtime-31.png 1280w, https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/systemtime-31-1024x768.png 1024w, https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/systemtime-31-768x576.png 768w\" sizes=\"auto, (max-width: 1280px) 100vw, 1280px\" \/><\/figure>\n\n\n\n<p>Topdown profile is sparse but has a few more backend stalls than other stalls.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1280\" height=\"960\" src=\"https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/amdtopdown-31.png\" alt=\"\" class=\"wp-image-2091\" srcset=\"https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/amdtopdown-31.png 1280w, https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/amdtopdown-31-1024x768.png 1024w, https:\/\/mvermeulen.org\/perf\/wp-content\/uploads\/sites\/7\/2024\/03\/amdtopdown-31-768x576.png 768w\" sizes=\"auto, (max-width: 1280px) 100vw, 1280px\" \/><\/figure>\n\n\n\n<p>AMD metrics show this as memory-bound with a moderate L2 access but high L2 miss rate. Opcache  and icache misses are lower as is the overall amount of frontend stalls. There is a high number of page faults suggesting a larger image.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>elapsed              77.962\non_cpu               0.036          # 0.57 \/ 16 cores\nutime                40.788\nstime                4.022\nnvcsw                2043           # 83.73%\nnivcsw               397            # 16.27%\ninblock              104            # 1.33\/sec\nonblock              80032          # 1026.55\/sec\ncpu-clock            44838944612    # 44.839 seconds\ntask-clock           44842022604    # 44.842 seconds\npage faults          1795046        # 40030.442\/sec\ncontext switches     2648           # 59.052\/sec\ncpu migrations       255            # 5.687\/sec\nmajor page faults    2              # 0.045\/sec\nminor page faults    1795044        # 40030.398\/sec\nalignment faults     0              # 0.000\/sec\nemulation faults     0              # 0.000\/sec\nbranches             33523335325    # 148.823 branches per 1000 inst\nbranch misses        732131582      # 2.18% branch miss\nconditional          23698905001    # 105.209 conditional branches per 1000 inst\nindirect             1555350390     # 6.905 indirect branches per 1000 inst\ncpu-cycles           196511308061   # 0.16 GHz\ninstructions         224485305110   # 1.14 IPC\nslots                395880545292   #\nretiring             75548253138    # 19.1% (19.1%)\n-- ucode             94125257       #     0.0%\n-- fastpath          75454127881    #    19.1%\nfrontend             37847893235    #  9.6% ( 9.6%)\n-- latency           26504030406    #     6.7%\n-- bandwidth         11343862829    #     2.9%\nbackend              259262254091   # 65.5% (65.5%)\n-- cpu               22063845225    #     5.6%\n-- memory            237198408866   #    59.9%\nspeculation          23140243395    #  5.8% ( 5.8%)\n-- branch mispredict 22566641443    #     5.7%\n-- pipeline restart  573601952      #     0.1%\nsmt-contention       81622036       #  0.0% ( 0.0%)\ncpu-cycles           196873203387   # 0.15 GHz\ninstructions         223740459677   # 1.14 IPC\ninstructions         74815515355    # 16.773 l2 access per 1000 inst\nl2 hit from l1       867622158      # 54.20% l2 miss\nl2 miss from l1      385079902      #\nl2 hit from l2 pf    92216461       #\nl3 hit from l2 pf    84717755       #\nl3 miss from l2 pf   210321243      #\ninstructions         74852913626    # 97.722 float per 1000 inst\nfloat 512            65             # 0.000 AVX-512 per 1000 inst\nfloat 256            596            # 0.000 AVX-256 per 1000 inst\nfloat 128            7314742792     # 97.722 AVX-128 per 1000 inst\nfloat MMX            0              # 0.000 MMX per 1000 inst\nfloat scalar         0              # 0.000 scalar per 1000 inst\ninstructions         224508389933   #\nopcache              46644408269    # 207.762 opcache per 1000 inst\nopcache miss         2316394861     #  5.0% opcache miss rate\nl1 dTLB miss         901995710      # 4.018 L1 dTLB per 1000 inst\nl2 dTLB miss         428419600      # 1.908 L2 dTLB per 1000 inst\ninstructions         224518696767   #\nicache               5188957924     # 23.111 icache per 1000 inst\nicache miss          247155351      #  4.8% icache miss rate\nl1 iTLB miss         8246787        # 0.037 L1 iTLB per 1000 inst\nl2 iTLB miss         0              # 0.000 L2 iTLB per 1000 inst\ntlb flush            17624          # 0.000 TLB flush per 1000 inst\n<\/code><\/pre>\n\n\n\n<p>Intel metrics confirm this as a dram bound application.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>elapsed              78.459\non_cpu               0.035          # 0.56 \/ 16 cores\nutime                41.446\nstime                2.806\nnvcsw                2144           # 87.51%\nnivcsw               306            # 12.49%\ninblock              381824         # 4866.55\/sec\nonblock              68792          # 876.79\/sec\ncpu-clock            44260949968    # 44.261 seconds\ntask-clock           44264376235    # 44.264 seconds\npage faults          1785230        # 40331.078\/sec\ncontext switches     2664           # 60.184\/sec\ncpu migrations       286            # 6.461\/sec\nmajor page faults    43             # 0.971\/sec\nminor page faults    1785187        # 40330.106\/sec\nalignment faults     0              # 0.000\/sec\nemulation faults     0              # 0.000\/sec\nbranches             32716277551    # 146.506 branches per 1000 inst\nbranch misses        461064056      # 1.41% branch miss\nconditional          32716289647    # 146.506 conditional branches per 1000 inst\nindirect             1555729043     # 6.967 indirect branches per 1000 inst\nslots                860942765204   #\nretiring             217875645299   # 25.3% (25.3%)\n-- ucode             14864431684    #     1.7%\n-- fastpath          203011213615   #    23.6%\nfrontend             47874051035    #  5.6% ( 5.6%)\n-- latency           22289417139    #     2.6%\n-- bandwidth         25584633896    #     3.0%\nbackend              493844325597   # 57.4% (57.4%)\n-- cpu               134378018395   #    15.6%\n-- memory            359466307202   #    41.8%\nspeculation          104189435651   # 12.1% (12.1%) high\n-- branch mispredict 101968937864   #    11.8%\n-- pipeline restart  2220497787     #     0.3%\nsmt-contention       0              #  0.0% ( 0.0%)\ncpu-cycles           143952942248   # 0.11 GHz\ninstructions         222454692992   # 1.55 IPC\nl2 access            6186401409     # 27.816 l2 access per 1000 inst\nl2 miss              4300814427     # 69.52% l2 miss\ncpu-cycles           143931637736   # 46.2% memory latency\nload stalls          64714460372    #  0.0% l1 bound\nl1 miss              77055639785    # 13.1% l2 bound\nl2 miss              58150653301    #  5.4% l3 bound\nl3 miss              50425863737    # 35.0% dram bound\nstore_stalls         1809094793     #  1.3% store bound\n<\/code><\/pre>\n\n\n\n<p>Process information shows draco_encoder as the primary user application but also a higher percentage of test overhead.<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>362 processes\n\t  6 draco_encoder           39.71     3.01\n\t 68 clinfo                  16.71     5.99\n\t 38 vulkaninfo               1.72     0.95\n\t  4 vulkani:disk$0           0.18     0.10\n\t  6 glxinfo:gdrv0            0.14     0.02\n\t  6 glxinfo:gl0              0.14     0.02\n\t  2 llvmpipe-0               0.09     0.05\n\t  2 llvmpipe-1               0.09     0.05\n\t  2 llvmpipe-10              0.09     0.05\n\t  2 llvmpipe-11              0.09     0.05\n\t  2 llvmpipe-12              0.09     0.05\n\t  2 llvmpipe-13              0.09     0.05\n\t  2 llvmpipe-14              0.09     0.05\n\t  2 llvmpipe-15              0.09     0.05\n\t  2 llvmpipe-2               0.09     0.05\n\t  2 llvmpipe-3               0.09     0.05\n\t  2 llvmpipe-4               0.09     0.05\n\t  2 llvmpipe-5               0.09     0.05\n\t  2 llvmpipe-6               0.09     0.05\n\t  2 llvmpipe-7               0.09     0.05\n\t  2 llvmpipe-8               0.09     0.05\n\t  2 llvmpipe-9               0.09     0.05\n\t  6 php                      0.07     0.08\n\t  2 glxinfo                  0.06     0.02\n\t  2 glxinfo:cs0              0.06     0.02\n\t  2 glxinfo:disk$0           0.06     0.02\n\t  2 glxinfo:sh0              0.06     0.02\n\t  2 glxinfo:shlo0            0.06     0.02\n\t  6 clang                    0.03     0.09\n\t  3 rocminfo                 0.03     0.00\n\t  1 lspci                    0.00     0.02\n\t 84 sh                       0.00     0.00\n\t 13 gcc                      0.00     0.00\n\t 13 gsettings                0.00     0.00\n\t  8 stat                     0.00     0.00\n\t  8 systemd-detect-          0.00     0.00\n\t  6 draco                    0.00     0.00\n\t  6 llvm-link                0.00     0.00\n\t  5 phoronix-test-s          0.00     0.00\n\t  2 cc                       0.00     0.00\n\t  2 gmain                    0.00     0.00\n\t  2 lscpu                    0.00     0.00\n\t  2 uname                    0.00     0.00\n\t  2 which                    0.00     0.00\n\t  2 xset                     0.00     0.00\n\t  1 date                     0.00     0.00\n\t  1 dconf worker             0.00     0.00\n\t  1 dirname                  0.00     0.00\n\t  1 dmesg                    0.00     0.00\n\t  1 dmidecode                0.00     0.00\n\t  1 grep                     0.00     0.00\n\t  1 ifconfig                 0.00     0.00\n\t  1 ip                       0.00     0.00\n\t  1 lsmod                    0.00     0.00\n\t  1 mktemp                   0.00     0.00\n\t  1 ps                       0.00     0.00\n\t  1 qdbus                    0.00     0.00\n\t  1 readlink                 0.00     0.00\n\t  1 realpath                 0.00     0.00\n\t  1 sed                      0.00     0.00\n\t  1 sort                     0.00     0.00\n\t  1 stty                     0.00     0.00\n\t  1 systemctl                0.00     0.00\n\t  1 template.sh              0.00     0.00\n\t  1 wc                       0.00     0.00\n\t  1 xrandr                   0.00     0.00\n0 processes running\n47 maximum processes\n<\/code><\/pre>\n\n\n\n<p>Computation structure<\/p>\n\n\n\n<pre class=\"wp-block-code\"><code>      59852) draco            cpu=14 start=5.59  finish=11.96\n        59853) draco_encoder    cpu=4 start=5.59  finish=11.95\n      59856) draco            cpu=11 start=15.96 finish=22.34\n        59857) draco_encoder    cpu=12 start=15.97 finish=22.32\n      59858) draco            cpu=11 start=26.34 finish=32.71\n        59859) draco_encoder    cpu=4 start=26.34 finish=32.70\n      59860) sh               cpu=11 start=32.71 finish=32.71\n        59861) sh               cpu=13 start=32.71 finish=32.71\n      59862) draco            cpu=11 start=42.89 finish=50.82\n        59863) draco_encoder    cpu=4 start=42.90 finish=50.81\n      59864) draco            cpu=11 start=54.83 finish=62.73\n        59865) draco_encoder    cpu=4 start=54.83 finish=62.72\n      59866) draco            cpu=11 start=66.73 finish=74.64\n        59867) draco_encoder    cpu=4 start=66.74 finish=74.63\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"<p>Testing the draco library which compresses\/decompresses 3d geometric meshes and point clouds. There are two workloads. The benchmark looks to be single-threaded and runs quickly. Topdown profile is sparse but has a few more backend stalls than other stalls. AMD <span class=\"excerpt-dots\">&hellip;<\/span> <a class=\"more-link\" href=\"https:\/\/mvermeulen.org\/perf\/workloads\/phoronix\/draco\/\"><span class=\"more-msg\">Continue reading &rarr;<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":58,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-2080","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/2080","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/comments?post=2080"}],"version-history":[{"count":2,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/2080\/revisions"}],"predecessor-version":[{"id":2092,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/2080\/revisions\/2092"}],"up":[{"embeddable":true,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/58"}],"wp:attachment":[{"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/media?parent=2080"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}