{"id":818,"date":"2024-01-22T11:02:16","date_gmt":"2024-01-22T11:02:16","guid":{"rendered":"https:\/\/mvermeulen.org\/perf\/?page_id=818"},"modified":"2024-01-22T19:15:49","modified_gmt":"2024-01-22T19:15:49","slug":"stress-ng","status":"publish","type":"page","link":"https:\/\/mvermeulen.org\/perf\/workloads\/phoronix\/stress-ng\/","title":{"rendered":"stress-ng"},"content":{"rendered":"\n

A stress engine of various Linux operations with 39 workload tests around various system calls. As might be expected there is a somewhat differing set of workloads for different system calls and similarly at least one that tries having a large number of runnable processes and some that test interrupt behavior.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Topdown overview shows similar behavior all over the map. To dive deeper, makes sense to separate out the workloads.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Composite is shown in AMD topdown metrics. Overall keeping most cores busy. Some floating point and a reasonable number of branches.<\/p>\n\n\n\n

elapsed              5084.757\non_cpu               0.773          # 12.37 \/ 16 cores\nutime                32335.341\nstime                30543.925\nnvcsw                798829363      # 24.59%\nnivcsw               2449890031     # 75.41%\ninblock              6920           # 1.36\/sec\nonblock              22937248       # 4510.98\/sec\ncpu-clock            63338126482931 # 63338.126 seconds\ntask-clock           63370492886426 # 63370.493 seconds\npage faults          252282205      # 3981.067\/sec\ncontext switches     3247234343     # 51242.056\/sec\ncpu migrations       40594272       # 640.586\/sec\nmajor page faults    3885           # 0.061\/sec\nminor page faults    244989996      # 3865.995\/sec\nalignment faults     0              # 0.000\/sec\nemulation faults     0              # 0.000\/sec\nbranches             33287849051067 # 134.232 branches per 1000 inst\nbranch misses        2040381135392  # 6.13% branch miss\nconditional          21738326981131 # 87.659 conditional branches per 1000 inst\nindirect             1043706930671  # 4.209 indirect branches per 1000 inst\ncpu-cycles           246020356645369 # 3.02 GHz\ninstructions         247522262682045 # 1.01 IPC\nslots                494262022605330 #\nretiring             91346010020862 # 18.5% (23.6%)\n-- ucode             458614951576   #     0.1%\n-- fastpath          90887395069286 #    18.4%\nfrontend             175331383055920 # 35.5% (45.4%)\n-- latency           138264249623772 #    28.0%\n-- bandwidth         37067133432148 #     7.5%\nbackend              116225434441614 # 23.5% (30.1%)\n-- cpu               36027639621668 #     7.3%\n-- memory            80197794819946 #    16.2%\nspeculation          3425655658401  #  0.7% ( 0.9%)\n-- branch mispredict 3407508332244  #     0.7%\n-- pipeline restart  18147326157    #     0.0%\nsmt-contention       107922040366956 # 21.8% ( 0.0%)\ncpu-cycles           237754882893348 # 3.02 GHz\ninstructions         245483193483317 # 1.03 IPC\ninstructions         82870905019605 # 29.832 l2 access per 1000 inst\nl2 hit from l1       1876625583798  # 11.86% l2 miss\nl2 miss from l1      121920128803   #\nl2 hit from l2 pf    424202227051   #\nl3 hit from l2 pf    109437666801   #\nl3 miss from l2 pf   61904909671    #\ninstructions         82821378194925 # 95.294 float per 1000 inst\nfloat 512            180            # 0.000 AVX-512 per 1000 inst\nfloat 256            3512658061     # 0.042 AVX-256 per 1000 inst\nfloat 128            7888690609108  # 95.249 AVX-128 per 1000 inst\nfloat MMX            0              # 0.000 MMX per 1000 inst\nfloat scalar         141197153      # 0.002 scalar per 1000 inst\n<\/code><\/pre>\n\n\n\n
Intel metrics<\/p>\n\n\n\n
elapsed              6757.530\non_cpu               0.785          # 12.56 \/ 16 cores\nutime                53318.572\nstime                31555.783\nnvcsw                2058548634     # 26.09%\nnivcsw               5832029530     # 73.91%\ninblock              637688         # 94.37\/sec\nonblock              13248472       # 1960.55\/sec\ncpu-clock            85035232641948 # 85035.233 seconds\ntask-clock           85064973177120 # 85064.973 seconds\npage faults          371711301      # 4369.734\/sec\ncontext switches     7887342199     # 92721.386\/sec\ncpu migrations       68032834       # 799.775\/sec\nmajor page faults    12513          # 0.147\/sec\nminor page faults    361797738      # 4253.193\/sec\nalignment faults     0              # 0.000\/sec\nemulation faults     0              # 0.000\/sec\nbranches             46012903714132 # 129.178 branches per 1000 inst\nbranch misses        283085258308   # 0.62% branch miss\nconditional          46013293141460 # 129.180 conditional branches per 1000 inst\nindirect             14151704748033 # 39.730 indirect branches per 1000 inst\nslots                517436504170508 #\nretiring             227053908476335 # 43.9% (43.9%)\n-- ucode             27953978000842 #     5.4%\n-- fastpath          199099930475493 #    38.5%\nfrontend             133851851528919 # 25.9% (25.9%)\n-- latency           66874983273868 #    12.9%\n-- bandwidth         66976868255051 #    12.9%\nbackend              141357582855630 # 27.3% (27.3%)\n-- cpu               29262841118363 #     5.7%\n-- memory            112094741737267 #    21.7%\nspeculation          14979447431469 #  2.9% ( 2.9%)\n-- branch mispredict 13354893130669 #     2.6%\n-- pipeline restart  1624554300800  #     0.3%\nsmt-contention       0              #  0.0% ( 0.0%)\ncpu-cycles           194962391661297 # 1.65 GHz\ninstructions         292457779985404 # 1.50 IPC\nl2 access            4054293120327  # 17.627 l2 access per 1000 inst\nl2 miss              729690506741   # 18.00% l2 miss\n<\/code><\/pre>\n\n\n\n
Process summary, dies after 23 tests, there are more than five million processes forked – so this is also a good stress test for the tools.<\/p>\n","protected":false},"excerpt":{"rendered":"
A stress engine of various Linux operations with 39 workload tests around various system calls. As might be expected there is a somewhat differing set of workloads for different system calls and similarly at least one that tries having a …<\/span> Continue reading →<\/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-818","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/818","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=818"}],"version-history":[{"count":2,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/818\/revisions"}],"predecessor-version":[{"id":831,"href":"https:\/\/mvermeulen.org\/perf\/wp-json\/wp\/v2\/pages\/818\/revisions\/831"}],"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=818"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}