AwAppp/benchmarks_8bit_batch_size10
0
1{2 "prefill": {3 "memory": {4 "unit": "MB",5 "max_ram": 3525.599232,6 "max_vram": 11714.691072,7 "max_reserved": 11221.860352,8 "max_allocated": 9991.7245449 },10 "latency": {11 "unit": "s",12 "mean": 0.3010050461713006,13 "stdev": 0.005007823178304763,14 "values": [15 0.32892837524414065,16 0.3053179016113281,17 0.3002685546875,18 0.2994657287597656,19 0.3013898315429687,20 0.29939508056640624,21 0.29945855712890623,22 0.3003228149414062,23 0.29954458618164065,24 0.29985791015625,25 0.3006095275878906,26 0.3026851806640625,27 0.2997637023925781,28 0.300037109375,29 0.2996305847167969,30 0.29964801025390625,31 0.3008983154296875,32 0.2997923889160156,33 0.2989916076660156,34 0.29947698974609377,35 0.2995968017578125,36 0.2994698181152344,37 0.29934283447265625,38 0.299536376953125,39 0.29968487548828127,40 0.29946783447265624,41 0.30125567626953126,42 0.30218035888671874,43 0.2998056945800781,44 0.29990194702148437,45 0.2995118103027344,46 0.29968487548828127,47 0.29941860961914063,48 0.29983129882812549 ]50 },51 "throughput": {52 "unit": "tokens/s",53 "value": 531.552550480980154 },55 "energy": null,56 "efficiency": null57 },58 "decode": {59 "memory": {60 "unit": "MB",61 "max_ram": 3536.162816,62 "max_vram": 12268.3392,63 "max_reserved": 11775.50848,64 "max_allocated": 10078.90227265 },66 "latency": {67 "unit": "s",68 "mean": 25.19557006835937,69 "stdev": 0,70 "values": [71 25.1955700683593772 ]73 },74 "throughput": {75 "unit": "tokens/s",76 "value": 39.29262157252172577 },78 "energy": null,79 "efficiency": null80 },81 "per_token": {82 "memory": null,83 "latency": {84 "unit": "s",85 "mean": 0.25450070776120576,86 "stdev": 0.0028173351708351006,87 "values": [88 0.2622341003417969,89 0.2604031982421875,90 0.2598922119140625,91 0.259346435546875,92 0.26023934936523435,93 0.259125244140625,94 0.259641357421875,95 0.2600816650390625,96 0.2582261657714844,97 0.2586593322753906,98 0.2587289733886719,99 0.2589450378417969,100 0.2583848876953125,101 0.2578052978515625,102 0.25718887329101564,103 0.2568939514160156,104 0.25893478393554686,105 0.2582753295898437,106 0.25697381591796875,107 0.2581626892089844,108 0.256964599609375,109 0.256321533203125,110 0.25642291259765626,111 0.25666763305664064,112 0.25534873962402344,113 0.25577267456054686,114 0.25576243591308595,115 0.25529037475585936,116 0.25610955810546876,117 0.2554173431396484,118 0.2555606994628906,119 0.25545318603515627,120 0.2541045684814453,121 0.2553528289794922,122 0.2549780426025391,123 0.2547394561767578,124 0.2549903411865234,125 0.2602659912109375,126 0.2590689392089844,127 0.25401242065429686,128 0.2539376678466797,129 0.254271484375,130 0.2540543975830078,131 0.25435136413574216,132 0.2539878387451172,133 0.25366937255859373,134 0.25358233642578126,135 0.25377484130859373,136 0.2529126434326172,137 0.2540585021972656,138 0.25315840148925783,139 0.2530897979736328,140 0.2536099853515625,141 0.2528061370849609,142 0.25310617065429686,143 0.25320346069335936,144 0.2535393218994141,145 0.2527958984375,146 0.253264892578125,147 0.25287986755371095,148 0.2530160675048828,149 0.2529689636230469,150 0.25187532043457034,151 0.25355877685546874,152 0.25151078796386717,153 0.2520494079589844,154 0.25327001953125,155 0.25154150390625,156 0.2519019470214844,157 0.25225625610351565,158 0.25124659729003906,159 0.25213848876953127,160 0.25257369995117185,161 0.2515240936279297,162 0.2524241943359375,163 0.2529085388183594,164 0.2525675506591797,165 0.25248051452636716,166 0.2522879943847656,167 0.2524682312011719,168 0.2532095947265625,169 0.25251429748535154,170 0.2515650634765625,171 0.25252351379394533,172 0.25155378723144534,173 0.2517022705078125,174 0.25210067749023435,175 0.25160902404785157,176 0.25191014099121095,177 0.25165721130371094,178 0.25126296997070313,179 0.25138584899902344,180 0.250998779296875,181 0.2516172790527344,182 0.2514114532470703,183 0.25093426513671874,184 0.25146060180664065,185 0.251789306640625,186 0.25099366760253905187 ]188 },189 "throughput": {190 "unit": "tokens/s",191 "value": 39.292621572521725192 },193 "energy": null,194 "efficiency": null195 }196}