Amazon EC2 Instance Types

文章推薦指數: 80 %
投票人數:10人

Amazon EC2 Instance Types ; General Purpose Instances · m6g.medium, 4 GiB of Memory, 1 vCPU, EBS only, 64-bit Arm platform; m6g.large, 8 GiB of Memory, 2 vCPUs, ... Products AmazonEC2 AmazonEC2InstanceTypes AmazonEC2InstanceTypes GeneralPurposeInstances CurrentGeneration AmazonEC2M6ginstancesarepoweredbyArm-basedAmazonWebServicesGraviton2processors.Theydeliverupto40%betterprice/performanceovercurrentgenerationM5instancesandofferabalanceofcompute,memory,andnetworkingresourcesforabroadsetofworkloads. m6g.medium,4GiBofMemory,1vCPU,EBSonly,64-bitArmplatform m6g.large,8GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform m6g.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bit Armplatform m6g.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bit Armplatform m6g.4xlarge,64GiBofMemory,16vCPUs,EBSonly,64-bit Armplatform m6g.8xlarge,128GiBofMemory,32vCPUs,EBSonly,64-bitArmplatform m6g.12xlarge,192GiBofMemory,48vCPUs,EBSonly,64-bit Armplatform m6g.16xlarge,256GiBofMemory,64vCPUs,EBSonly,64-bit Armplatform m6g.metal,256GiBofmemory,64vCPUs,EBSonly, 64-bit Armplatform AmazonEC2T4ginstancesarepoweredbyArm-basedAmazonGraviton2processorsanddeliverupto40%betterpriceperformanceoverT3instancesforabroadsetofburstablegeneralpurposeworkloads.UntilJune30th,2021,allnewandexistingAmazonWebServicescustomerscantrythet4g.microinstancesfreeforupto750hourspermonth. t4g.nano,0.5GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform t4g.micro,1GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform t4g.small,2GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform t4g.medium,4GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform t4g.large,8GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform t4g.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitArmplatform t4g.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitArmplatform AmazonEC2M5Instances arepoweredbyIntelXeon®Platinum81753.1GHzprocessors.WithM5dinstances,localNVMe-basedSSDsarephysicallyconnectedtothehostserverandprovideblock-levelstoragethatiscoupledtothelifetimeoftheM5instance. m5.large,8GiBofMemory,2vCPUs,EBSonly,64-bitplatform m5.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitplatform m5.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitplatform m5.4xlarge,64GiBofMemory,16vCPUs,EBSonly,64-bitplatform m5.8xlarge,128GiBofMemory,32vCPUs,EBSonly,64-bitplatform m5.12xlarge,192GiBofMemory,48vCPUs,EBSonly,64-bitplatform m5.16xlarge,256GiBofMemory,64vCPUs,EBSonly,64-bitplatform m5.24xlarge,384GiBofMemory,96vCPUs,EBSonly,64-bitplatform m5.metal:384GiBofMemory,96vCPUs,EBSonly,64-bitplatform m5d.metal:384GiBofMemory,96vCPUs,4x900NVMeSSD,64-bitplatform   m5a.large,8GiBofMemory,2vCPUs,EBSonly,64-bitplatform m5a.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitplatform m5a.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitplatform m5a.4xlarge,64GiBofMemory,16vCPUs,EBSonly,64-bitplatform m5a.8xlarge,128GiBofMemory,32vCPUs,EBSonly,64-bitplatform m5a.12xlarge,192GiBofMemory,48vCPUs,EBSonly,64-bitplatform m5a.16xlarge,256GiBofMemory,64vCPUs,EBSonly,64-bitplatform m5a.24xlarge,384GiBofMemory,96vCPUs,EBSonly,64-bitplatform   m5d.large,8GiBofMemory,2vCPUs,1x75NVMeSSD,64-bitplatform m5d.xlarge,16GiBofMemory,4vCPUs,1x150NVMeSSD,64-bitplatform m5d.2xlarge,32GiBofMemory,8vCPUs,1x300NVMeSSD,64-bitplatform m5d.4xlarge,64GiBofMemory,16vCPUs,2x300NVMeSSD,64-bitplatform m5d.8xlarge,128GiBofMemory,32vCPUs,2x600NVMeSSD,64-bitplatform m5d.12xlarge,192GiBofMemory,48vCPUs,2x900NVMeSSD,64-bitplatform m5d.16xlarge,256GiBofMemory,64vCPUs,4x600NVMeSSD,64-bitplatform m5d.24xlarge,384GiBofMemory,96vCPUs,4x900NVMeSSD,64-bitplatform PreviousGenerations m4.large:8GiBofmemory,2vCPU,EBS-only,64-bitplatform m4.xlarge:16GiBofmemory,4vCPUs,EBS-only,64-bitplatform m4.2xlarge:32GiBofmemory,8vCPUs,EBS-only,64-bitplatform m4.4xlarge:64GiBofmemory,16vCPUs,EBS-only,64-bitplatform m4.10xlarge:160GiBofmemory,40vCPUs,EBS-only,64-bitplatform m4.16xlarge:256GiBofmemory,64vCPUs,EBS-only,64-bitplatform m3.medium:3.75GiBofmemory,1vCPU,4GBofSSD-basedlocalinstancestorage,64-bitplatform m3.large:7.5GiBofmemory,2VCPUs,32GBofSSD-basedlocalinstancestorage,64-bitplatform m3.xlarge:15GiBofmemory,4vCPUs,80GBofSSD-basedlocalinstancestorage,64-bitplatform m3.2xlarge:30GiBofmemory,8vCPUs,160GBofSSD-basedlocalinstancestorage,64-bitplatform m1.small:1.7GiBofmemory,1EC2ComputeUnit(1virtualcorewith1EC2ComputeUnit),160GBoflocalinstancestorage,32-bitor64-bitplatform T3instancesarethenextgenerationlowcostburstablegeneral-purposeinstancetypethatprovideabaselinelevelofCPUperformancewiththeabilitytoburstCPUusageatanytimeforaslongasrequired.T3instancesaredesignedforapplicationswithmoderateCPUusagethatexperiencetemporaryspikesinuse. t3.nano,0.5GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3.micro,1GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3.small,2GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3.medium,4GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3.large,8GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitplatform t3.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitplatform   t3a.nano,0.5GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3a.micro,1GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3a.small,2GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3a.medium,4GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3a.large,8GiBofMemory,2vCPUs,EBSonly,64-bitplatform t3a.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitplatform t3a.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitplatform t2.nano,0.5GiBofMemory,1vCPUs,EBSonly,64-bitplatform t2.micro,1GiBofMemory,1vCPUs,EBSonly,64-bitplatform t2.small,2GiBofMemory,1vCPUs,EBSonly,64-bitplatform t2.medium,4GiBofMemory,2vCPUs,EBSonly,64-bitplatform t2.large,8GiBofMemory,2vCPUs,EBSonly,64-bitplatform t2.xlarge,16GiBofMemory,4vCPUs,EBSonly,64-bitplatform t2.2xlarge,32GiBofMemory,8vCPUs,EBSonly,64-bitplatform ComputeOptimizedInstances AmazonEC2C6ginstancesarepoweredbyArm-basedAmazonWebservicesGraviton2processors.Theydeliverupto40%betterpriceperformanceovercurrentgenerationC5instancesforcompute-intensiveapplications. c6g.medium,2GiBofMemory,1vCPU,EBSonly,64-bit Armplatform c6g.large,4GiBofMemory,2vCPUs,EBSonly,64-bitArmplatform c6g.xlarge,8GiBofMemory,4vCPUs,EBSonly,64-bit Armplatform c6g.2xlarge,16GiBofMemory,8vCPUs,EBSonly,64-bit Armplatform c6g.4xlarge,32GiBofMemory,16vCPUs,EBSonly,64-bit Armplatform c6g.8xlarge,64GiBofMemory,32vCPUs,EBSonly,64-bit Armplatform c6g.12xlarge,96GiBofMemory,48vCPUs,EBSonly,64-bit Armplatform c6g.16xlarge,128GiBofMemory,64vCPUs,EBSonly,64-bit Armplatform c6g.metal,128GiBofMemory,64vCPUs,EBSonly,64-bitArmplatform C5Instancesareoptimizedforcompute-intensiveworkloadsanddelivercost-effectivehighperformanceatalowpricepercomputeratio.WithC5dinstances,localNVMe-basedSSDsarephysicallyconnectedtothehostserverandprovideblock-levelstoragethatiscoupledtothelifetimeoftheC5instance. c5.large:4GiBofmemory,2vCPUs,64-bitplatform c5.xlarge:8GiBofmemory,4vCPUs,64-bitplatform c5.2xlarge:16GiBofmemory,8vCPUs,64-bitplatform c5.4xlarge:32GiBofmemory,16vCPUs,64-bitplatform c5.9xlarge:72GiBofmemory,36vCPUs,64-bitplatform c5.18xlarge:144GiBofmemory,72vCPUs,64-bitplatform c5.metal:96vCPUs,192GiBofMemory,EBSonly,64-bitplatform c5d.metal:96vCPUs,192GiBofMemory,4x900NVMeSSD,64-bitplatform   c5d.large:4GiBofmemory,2vCPUs,1x50NVMeSSD,64-bitplatform c5d.xlarge:8GiBofmemory,4vCPUs,1x100NVMeSSD,64-bitplatform c5d.2xlarge:16GiBofmemory,8vCPUs,1x200NVMeSSD,64-bitplatform c5d.4xlarge:32GiBofmemory,16vCPUs,1x400NVMeSSD,64-bitplatform c5d.9xlarge:72GiBofmemory,36vCPUs,1x900NVMeSSD,64-bitplatform c5d.18xlarge:144GiBofmemory,72vCPUs,2x900NVMeSSD,64-bitplatform   c5a.large,4GiBofMemory,2vCPUs,EBSonly,64-bitplatform c5a.xlarge,8GiBofMemory,4vCPUs,EBSonly,64-bitplatform c5a.2xlarge,16GiBofMemory,8vCPUs,EBSonly,64-bitplatform c5a.4xlarge,32GiBofMemory,16vCPUs,EBSonly,64-bitplatform c5a.8xlarge,64GiBofMemory,32vCPUs,EBSonly,64-bitplatform c5a.12xlarge,96GiBofMemory,48vCPUs,EBSonly,64-bitplatform c5a.16xlarge,128GiBofMemory,64vCPUs,EBSonly,64-bitplatform c5a.24xlarge,192GiBofMemory,96vCPUs,EBSonly,64-bitplatform PreviousGenerations c4.large:3.75GiBofmemory,2vCPUs,64-bitplatform c4.xlarge:7.5GiBofmemory,4vCPUs,64-bitplatform c4.2xlarge:15GiBofmemory,8vCPUs,64-bitplatform c4.4xlarge:30GiBofmemory,16vCPUs,64-bitplatform c4.8xlarge:60GiBofmemory,36vCPUs,64-bitplatform c3.large:3.75GiBofmemory,2vCPUs,32GBofSSD-basedlocalinstancestorage,64-bitplatform c3.xlarge:7.5GiBofmemory,4vCPUs,80GBofSSD-basedlocalinstancestorage,64-bitplatform c3.2xlarge:15GiBofmemory,8vCPUs,160GBofSSD-basedlocalinstancestorage,64-bitplatform c3.4xlarge:30GiBofmemory,16vCPUs,320GBofSSD-basedlocalinstancestorage,64-bitplatform c3.8xlarge:60GiBofmemory,32vCPUs,640GBofSSD-basedlocalinstancestorage,64-bitplatform MemoryOptimizedInstances AmazonEC2R6ginstancesarepoweredbyArm-basedAmazonWebServicesGraviton2processors.Theydeliverupto40%betterpriceperformanceovercurrentgenerationR5instancesformemory-intensiveapplications. r6g.medium,8GiBofMemory,1vCPU,EBSonly,64-bit Armplatform r6g.large,16GiBofMemory,2vCPUs,EBSonly,64-bit Armplatform r6g.xlarge,32GiBofMemory,4vCPUs,EBSonly,64-bit Armplatform r6g.2xlarge,64GiBofMemory,8vCPUs,EBSonly,64-bit Armplatform r6g.4xlarge,128GiBofMemory,16vCPUs,EBSonly,64-bit Armplatform r6g.8xlarge,256GiBofMemory,32vCPUs,EBSonly,64-bit Armplatform r6g.12xlarge,384GiBofMemory,48vCPUs,EBSonly,64-bit Armplatform r6g.16xlarge,512GiBofMemory,64vCPUs,EBSonly,64-bit Armplatform r6g.metal,512GiBofMemory,64vCPUs,EBSonly,64-bitArmplatform X1Instancesareoptimizedforlarge-scale,enterprise-class,in-memoryapplicationsandhavethelowestpriceperGiBofRAMamongAmazonEC2instancetypes. x1.16xlarge:976GiBofmemory,64vCPUs,1x1,920GBofSSD-basedinstancestorage,64-bitplatform,10GigabitEthernet x1.32xlarge:1,952GiBofmemory,128vCPUs,2x1,920GBofSSD-basedinstancestorage,64-bitplatform,20GigabitEthernet R5instances deliver5%additionalmemorypervCPUthanR4andthelargestsizeprovides768GiBofmemory.Inaddition,R5instancesdelivera10%priceperGiBimprovementanda~20%increasedCPUperformanceoverR4.WithR5dinstances,localNVMe-basedSSDsarephysicallyconnectedtothehostserverandprovideblock-levelstoragethatiscoupledtothelifetimeoftheR5instance. r5.large:16GiBofmemory,2vCPUs,64-bitplatform r5.xlarge:32GiBofmemory,4vCPUs,64-bitplatform r5.2xlarge:64GiBofmemory,8vCPUs,64-bitplatform r5.4xlarge:128GiBofmemory,16vCPUs,64-bitplatform r5.8xlarge:256GiBofmemory,32vCPUs,64-bitplatform r5.12xlarge:384GiBofmemory,48vCPUs,64-bitplatform r5.16xlarge:512GiBofmemory,64vCPUs,64-bitplatform r5.24xlarge:768GiBofmemory,96vCPUs,64-bitplatform r5.metal:768GiBofmemory,96vCPUs,64-bitplatform r5d.metal:768GiBofmemory,96vCPUs,4x900NVMeSSD,64-bitplatform   r5a.large,16GiBofMemory,2vCPUs,EBS-Only,64-bitplatform r5a.xlarge,32GiBofMemory,4vCPUs,EBS-Only,64-bitplatform r5a.2xlarge,64GiBofMemory,8vCPUs,EBS-Only,64-bitplatform r5a.4xlarge,128GiBofMemory,16vCPUs,EBS-Only,64-bitplatform r5a.8xlarge,256GiBofMemory,32vCPUs,EBS-Only,64-bitplatform r5a.12xlarge,384GiBofMemory,48vCPUs,EBS-Only,64-bitplatform r5a.16xlarge,512GiBofMemory,64vCPUs,EBS-Only,64-bitplatform r5a.24xlarge,768GiBofMemory,96vCPUs,EBS-Only,64-bitplatform   r5d.large:16GiBofmemory,2vCPUs,1x75NVMeSSD,64-bitplatform r5d.xlarge:32GiBofmemory,4vCPUs,1x150NVMeSSD,64-bitplatform r5d.2xlarge:64GiBofmemory,8vCPUs,1x300NVMeSSD,64-bitplatform r5d.4xlarge:128GiBofmemory,16vCPUs,2x300NVMeSSD,64-bitplatform r5d.8xlarge:256GiBofmemory,32vCPUs,2x600NVMeSSD,64-bitplatform r5d.12xlarge:384GiBofmemory,48vCPUs,2x900NVMeSSD,64-bitplatform r5d.16xlarge:512GiBofmemory,64vCPUs,4x600NVMeSSD,64-bitplatform r5d.24xlarge:768GiBofmemory,96vCPUs,4x900NVMeSSD,64-bitplatform AmazonEC2z1dinstancesofferbothhighcomputecapacityandahighmemoryfootprint.Highfrequencyz1dinstancesdeliverasustainedallcorefrequencyofupto4.0GHz,thefastestofanycloudinstance.Withz1dinstances,localNVMe-basedSSDsarephysicallyconnectedtothehostserverandprovideblock-levelstoragethatiscoupledtothelifetimeofthez1dinstance. z1d.large:16GiBofMemory,2vCPUs,1x75NVMeSSD,64-bitplatform z1d.xlarge:32GiBofMemory,4vCPUs,1x150NVMeSSD,64-bitplatform z1d.2xlarge:64GiBofMemory,8vCPUs,1x300NVMeSSD,64-bitplatform z1d.3xlarge:96GiBofMemory,12vCPUs,1x450NVMeSSD,64-bitplatform z1d.6xlarge:192GiBofMemory,24vCPUs,1x900NVMeSSD,64-bitplatform z1d.12xlarge:384GiBofMemory,48vCPUs,2x900NVMeSSD,64-bitplatform z1d.metal:384GiBofMemory,48vCPUs,2x900NVMeSSD,64-bitplatform PreviousGenerations r4.large:15.25GiBofmemory,2vCPUs,64-bitplatform r4.xlarge:30.5GiBofmemory,4vCPUs,64-bitplatform r4.2xlarge:61GiBofmemory,8vCPUs,64-bitplatform r4.4xlarge:122GiBofmemory,16vCPUs,64-bitplatform r4.8xlarge:244GiBofmemory,32vCPUs,64-bitplatform r4.16xlarge:488GiBofmemory,64vCPUs,64-bitplatform r3.large:15.25GiBofmemory,2vCPUs,1x32GBofSSD-basedinstancestorage,64-bitplatform r3.xlarge:30.5GiBofmemory,4vCPUs,1x80GBofSSD-basedinstancestorage,64-bitplatform r3.2xlarge:61GiBofmemory,8vCPUs,1x160GBofSSD-basedinstancestorage,64-bitplatform r3.4xlarge:122GiBofmemory,16vCPUs,1x320GBofSSD-basedinstancestorage,64-bitplatform r3.8xlarge:244GiBofmemory,32vCPUs,2x320GBofSSD-basedinstancestorage,64-bitplatform,10GigabitEthernet StorageOptimizedInstances InstancesofthisfamilyprovideveryhighdiskI/Operformanceorproportionallyhigherstoragedensityperinstance,andareideallysuitedforapplicationsthatbenefitfromhighsequentialI/Operformanceacrossverylargedatasets.Storage-optimizedinstancesalsoprovidehighlevelsofCPU,memoryandnetworkperformance. i3.large:15.25GiBofmemory,2vCPUs,1x0.475NVMeSSD,64-bitplatform i3.xlarge:30.5GiBofmemory,4vCPUs,1x0.95NVMeSSD,64-bitplatform i3.2xlarge:61GiBofmemory,8vCPUs,1x1.9NVMeSSD,64-bitplatform i3.4xlarge:122GiBofmemory,16vCPUs,2x1.9NVMeSSD,64-bitplatform i3.8xlarge:244GiBofmemory,32vCPUs,4x1.9NVMeSSD,64-bitplatform i3.16xlarge:488GiBofmemory,64vCPUs,8x1.9NVMeSSD,64-bitplatform i3en.large:16GiBofmemory,2vCPUs,1x1.25TBNVMeSSD,64-bitplatform i3en.xlarge:32GiBofmemory,4vCPUs,1x2.5TBNVMeSSD,64-bitplatform i3en.2xlarge:64GiBofmemory,8vCPUs,2x2.5TBNVMeSSD,64-bitplatform i3en.3xlarge:96GiBofmemory,12vCPUs,1x7.5TBNVMeSSD,64-bitplatform i3en.6xlarge:192GiBofmemory,24vCPUs,2x7.5TBNVMeSSD,64-bitplatform i3en.12xlarge:384GiBofmemory,48vCPUs,4x7.5TBNVMeSSD,64-bitplatform i3en.24xlarge:768GiBofmemory,96vCPUs,8x7.5TBNVMeSSD,64-bitplatform I3en.metal:768GiBofmemory,96vCPUs,8x7.5TBNVMeSSD,64-bitplatform PreviousGenerations i2.xlarge:30.5GiBofmemory,4vCPUs,800GBofSSD-basedinstancestorage,64-bitplatform i2.2xlarge:61GiBofmemory,8vCPUs,2x800GBofSSD-basedinstancestorage,64-bitplatform i2.4xlarge:122GiBofmemory,16vCPUs,4x800GBofSSD-basedinstancestorage,64-bitplatform i2.8xlarge:244GiBofmemory,32vCPUs,8x800GBofSSD-basedinstancestorage,64-bitplatform,10GigabitEthernet InstancesofthisfamilyprovidelowcoststorageandveryhighdiskthroughputandareideallysuitedforapplicationsthatbenefitfromhighsequentialI/Operformanceacrossverylargedatasetsonlocalstorage. d2.xlarge:30.5GiBofmemory,4vCPUs,3x2000GBofHDD-basedinstancestorage,64-bitplatform d2.2xlarge:61GiBofmemory,8vCPUs,6x2000GBofHDD-basedinstancestorage,64-bitplatform d2.4xlarge:122GiBofmemory,16vCPUs,12x2000GBofHDD-basedinstancestorage,64-bitplatform d2.8xlarge:244GiBofmemory,36vCPUs,24x2000GBofHDD-basedinstancestorage,64-bitplatform,10GigabitEthernet AcceleratedComputingInstances InstancesofthisfamilyprovideaccesstoworkloadacceleratorssuchasGPU.Theyareidealforapplicationssuchasmachinelearning,computationalfluiddynamics,computationalfinance,seismicanalysis,molecularmodeling,genomics,andotherhigh-performancecomputingworkloads. AmazonEC2P3instancesdeliverhighperformancecomputeinthecloudwithupto8NVIDIA®V100TensorCoreGPUsandupto100GbpsofnetworkingthroughputformachinelearningandHPCapplications.Theseinstancesdeliveruptoonepetaflopofmixed-precisionperformanceperinstancetosignificantlyacceleratemachinelearningandhighperformancecomputingapplications.AmazonEC2P3instanceshavebeenproventoreducemachinelearningtrainingtimesfromdaystominutes,aswellasincreasethenumberofsimulationscompletedforhighperformancecomputingby3-4x.  p3.2xlarge:1GPU,8vCPUs,61GiBofmemory,upto10Gbpsnetworkperformance p3.8xlarge,4GPU,32vCPUs,244GiBofmemory,10Gbpsnetworkperformance p3.16xlarge,8GPU,64vCPUs,488GiBofmemory,25Gbpsnetworkperformance AmazonEC2G4instancesdeliveracost-effectiveGPUinstancefordeployingmachinelearningmodelsinproductionandgraphics-intensiveapplications.G4instancesprovidethelatestgenerationNVIDIAT4GPUs,AmazonWebServicescustomIntelCascadeLakeCPUs,upto100Gbpsofnetworkingthroughput,andupto1.8TBoflocalNVMestorage.Theseinstancesdeliverupto65TFLOPsofFP16performancetoacceleratemachinelearninginferenceapplicationsandray-tracingcorestoaccelerategraphicsworkloadssuchasgraphicsworkstations,videotranscoding,andgamestreaminginthecloud. g4dn.xlarge,1GPU,4vCPUs,16GiBofmemory,125NVMeSSD,upto25Gbpsnetworkperformance g4dn.2xlarge,1GPU,8vCPUs,32GiBofmemory,225NVMeSSD,upto25Gbpsnetworkperformance g4dn.4xlarge,1GPU,16vCPUs,64GiBofmemory,225NVMeSSD,upto25Gbpsnetworkperformance g4dn.8xlarge,1GPU,32vCPUs,128GiBofmemory,1x900NVMeSSD,50Gbpsnetworkperformance g4dn.16xlarge,1GPU,64vCPUs,256GiBofmemory,1x900NVMeSSD,50Gbpsnetworkperformance g4dn.12xlarge,4GPUs,48vCPUs,192GiBofmemory,1x900NVMeSSD,50Gbpsnetworkperformance AmazonEC2Inf1instancesdeliverupto30%higherthroughputandupto45%lowercostperinferencethanAmazonEC2G4instances,whichwerealreadyhighlycost-efficientformachinelearninginferenceinthecloud.Inf1instancesarebuiltfromthegrounduptosupportmachinelearninginferenceapplications.Theseinstancesfeatureupto16AmazonWebServicesInferentiachips,high-performancemachinelearninginferencechipsdesignedandbuiltbyAmazonWebServices.Additionally,Inf1instancesincludethelatest2ndgenerationIntel®Xeon®Scalableprocessorsandupto100Gbpsnetworkingtoenablehighthroughputinference.UsingInf1instances,customerscanrunlargescalemachinelearninginferenceapplicationssuchassearchrecommendation,computervision,speechrecognition,naturallanguageprocessing,personalization,andfrauddetection,atlowcostinthecloud. inf1.xlarge,1Inferentiachip,4vCPUs,8GiBofmemory,upto25Gbpsofnetworkperformance inf1.2xlarge,1Inferentiachip,8vCPUs,16GiBofmemory,upto25Gbpsofnetworkperformance inf1.6xlarge,4Inferentiachips,24vCPUs,48GiBofmemory,25Gbpsofnetworkperformance inf1.24xlarge,16Inferentiachips,96vCPUs,192GiBofmemory,100Gbpsofnetworkperformance PreviousGenerations BackedbytheNVIDIATeslaM60GPUs,G3instancesareidealforgraphicsworkloadssuchas3Drendering,3Dvisualizations,graphics-intensiveremoteworkstations,videoencoding,andvirtualrealityapplications.g3.4xlarge:1GPU,16vCPUs,122GiBofmemory,upto10Gbitnetworkperformance g3.4xlarge,1GPU,16vCPUs,122GiBofMemory,upto10Gbitnetworkperformance g3.8xlarge:2GPUs,32vCPUs,244GiBofmemory,10Gbitnetworkperformance g3.16xlarge:4GPUs,64vCPUs,488GiBofmemory,20Gbitnetworkperformance p2.xlarge:1GPU,4vCPUs,61GiBofmemory,highnetworkperformance p2.8xlarge:8GPU,32vCPUs,488GiBofmemory,10Gbitnetworkperformance p2.16xlarge:16GPU,64vCPUs,732GiBofmemory,20Gbitnetworkperformance MicroInstances Microinstances(t1.micro)provideasmallamountofconsistentCPUresourcesandallowyoutoincreaseCPUcapacityinshortburstswhenadditionalcyclesareavailable.Theyarewellsuitedforlowerthroughputapplicationsandwebsitesthatrequireadditionalcomputecyclesperiodically.YoucanlearnmoreabouthowyoucanuseMicroinstancesandappropriateapplicationsintheAmazonEC2documentation. t1.micro:(Default)613MiBofmemory,upto2ECUs(forshortperiodicbursts),EBSstorageonly,32-bitor64-bitplatform EC2ComputeUnit(ECU)–OneEC2ComputeUnit(ECU)providestheequivalentCPUcapacityofa1.0-1.2GHz2007Opteronor2007Xeonprocessor. OnlineLiveChatChatWithCloudComputingExpertsToAnswerYourQuestions ContactUsbyHotline10100766 AmazonWebServicesChina(Beijing)RegionOperatedBySinnet10100966 AmazonWebServicesChina(Ningxia)RegionOperatedByNWCD ContactSalesContact AmazonWebServicesexpertstolearnmoreabout  AmazonWebServices PageContent GeneralPurposeInstances ComputeOptimizedInstances MemoryOptimizedInstances StorageOptimizedInstances AcceleratedComputingInstances MicroInstances EndingSupportforInternetExplorer Gotit OursupportforInternetExplorerendson07/31/2022.SupportedbrowsersareChrome,Firefox,Edge,andSafari. Gotit



請為這篇文章評分?