Home
%3CLINGO-SUB%20id%3D%22lingo-sub-393262%22%20slang%3D%22en-US%22%3ETuesdays%20with%20Corey%3A%20New%20specialized%20HPC%20VM%20sizes%20available%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-393262%22%20slang%3D%22en-US%22%3E%3CP%3EHave%20you%20even%20thought%20about%20what%20happens%20when%20you%20select%20a%20%E2%80%9Cmachine%20size%E2%80%9D%20when%20deploying%20a%20VM%20in%20Azure%3F%20You%20are%20selecting%20a%20set%20of%20resources%20(CPU%2C%20Cores%2C%20RAM%2C%20Networking%20speeds%2C%20Storage)%20configured%20together%20into%20a%20VM%20that%20will%20run%20your%20workload.%20As%20a%20hyperscale%20public%20cloud%20provider%20%E2%80%93%20Azure%20must%20forecast%20customer%20demand%20for%20specific%20configurations%20to%20ensure%20we%20have%20the%20available%20technologies%20and%20resources%20deployed%20in%20different%20Regions.%3C%2FP%3E%0A%3CP%3E%26nbsp%3B%3C%2FP%3E%0A%3CP%3EWhat%20if%20your%20workloads%20needed%20to%20leverage%20non%20standard%20resources%20for%20those%20workloads%20like%20a%20faster%20network%20between%20VM%20nodes%20or%20specialized%20processors%20and%20GPUs%3F%26nbsp%3B%20Don%E2%80%99t%20worry%20%E2%80%93%20we%20got%20you%20covered%20there%20as%20well.%20Certain%20regions%20have%20clusters%20of%20specialized%20hardware%20that%20can%20be%20included%20in%20your%20workload%20configurations%2C%20just%20by%20choosing%20the%20right%20machine%20type%20when%20you%20deploy.%20This%20includes%20(but%20not%20limited%20to)%3A%3C%2FP%3E%0A%3CUL%3E%0A%3CLI%3EDedicated%20GPU%20for%20computational%20heavy%20or%20visual%20rendering%20acceleration%3C%2FLI%3E%0A%3CLI%3EHigh%20speed%20networking%20interconnects%3C%2FLI%3E%0A%3CLI%3EHigh%20speed%20local%20storage%20for%20extremely%20low%20latency%20and%20high%20IO%20workloads%3C%2FLI%3E%0A%3C%2FUL%3E%0A%3CP%3EWhat%20about%20super%20specialized%20environments%20for%20fluid%20dynamics%20or%20structural%20dynamic%20simulations%3F%26nbsp%3B%20These%20workloads%20typically%20would%20be%20%22locked%22%20on%20premises%20due%20to%20the%20specialized%20networking%20or%20hardware%20components.%20Azure%20released%20a%20set%20of%20new%20HPC%20specific%20VM%20sizes%20%E2%80%93%20the%20H-Series%20machines%20Hb%20(memory%20bandwidth)%20and%20Hc%20(dense%20compute).%20I%20sat%20down%20with%20my%20good%20friend%20Corey%20Sanders%20(corporate%20VP%20of%20Azure%20Compute)%20and%20special%20guest%20Evan%20Burness%20(Principal%20PM%20in%20Azure%20Specialized%20Compute)%20to%20record%20their%20chat%20explaining%20the%20nuances%20of%20these%20two%20new%20machine%20sizes%20and%20when%20to%20use%20them.%3C%2FP%3E%0A%3CP%3E%26nbsp%3B%3C%2FP%3E%0A%3CP%3E%3C%2FP%3E%3CDIV%20class%3D%22video-embed-center%20video-embed%22%3E%3CIFRAME%20class%3D%22embedly-embed%22%20src%3D%22https%3A%2F%2Fcdn.embedly.com%2Fwidgets%2Fmedia.html%3Fsrc%3Dhttps%253A%252F%252Fwww.youtube.com%252Fembed%252F-GgNQ9eFI3I%253Ffeature%253Doembed%26amp%3Burl%3Dhttp%253A%252F%252Fwww.youtube.com%252Fwatch%253Fv%253D-GgNQ9eFI3I%26amp%3Bimage%3Dhttps%253A%252F%252Fi.ytimg.com%252Fvi%252F-GgNQ9eFI3I%252Fhqdefault.jpg%26amp%3Bkey%3Db0d40caa4f094c68be7c29880b16f56e%26amp%3Btype%3Dtext%252Fhtml%26amp%3Bschema%3Dyoutube%22%20width%3D%22600%22%20height%3D%22337%22%20scrolling%3D%22no%22%20frameborder%3D%220%22%20allow%3D%22autoplay%3B%20fullscreen%22%20allowfullscreen%3D%22true%22%20title%3D%22Video%22%3E%3C%2FIFRAME%3E%3C%2FDIV%3E%3CP%3E%3C%2FP%3E%0A%3CP%3E%26nbsp%3B%3C%2FP%3E%0A%3CP%3ECool%20points%20I%20got%20from%20this%20chat%3F%3C%2FP%3E%0A%3CUL%3E%0A%3CLI%3EThese%20machines%20are%20available%20now%20in%20select%20regions%3C%2FLI%3E%0A%3CLI%3EIndustry%20leading%201.5x%20to%202x%20bandwidth%20from%20RAM%20to%20CPU%3C%2FLI%3E%0A%3CLI%3E1000%2B%20node%20deployments%3C%2FLI%3E%0A%3CLI%3E100%20gigabit%20network%20interconnects%20between%20nodes%3C%2FLI%3E%0A%3CLI%3EMore%20new%20technology%20deployments%20year%20over%20year%20%E2%80%93%20more%20to%20come%3C%2FLI%3E%0A%3C%2FUL%3E%0A%3CP%3EThis%20video%20is%20part%20of%20a%20series%20I%20do%20called%20%3CSTRONG%3ETuesdays%20with%20Corey%3C%2FSTRONG%3E.%20Because%20of%20various%20complications%20%E2%80%93%20I%E2%80%99ve%20not%20had%20the%20chance%20to%20get%20more%20recording%20slots%20lined%20up%20as%20frequently%20as%20I%E2%80%99d%20like.%20Corey%20has%20committed%20to%20make%20this%20work%20and%20keep%20the%20show%20going.%20Are%20there%20any%20topics%20you%E2%80%99d%20like%20to%20see%20covered%20in%20a%20future%20episode%3F%20Comment%20here%20on%20the%20post%20to%20let%20us%20know%20what%20you%E2%80%99d%20like%20to%20see%20in%20future%20episodes%20and%20I%E2%80%99ll%20do%20my%20best%20to%20track%20down%20whoever%20can%20come%20on%20and%20talk%20about%20your%20requested%20topic!%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-TEASER%20id%3D%22lingo-teaser-393262%22%20slang%3D%22en-US%22%3E%3CP%3ECorey%20Sanders%20breaks%20down%20some%20very%20cool%20announcements%20around%20new%20specialized%20HPC%20VM%20Size%20offerings.%3C%2FP%3E%3C%2FLINGO-TEASER%3E%3CLINGO-LABS%20id%3D%22lingo-labs-393262%22%20slang%3D%22en-US%22%3E%3CLINGO-LABEL%3EAzure%3C%2FLINGO-LABEL%3E%3C%2FLINGO-LABS%3E
Microsoft

Have you even thought about what happens when you select a “machine size” when deploying a VM in Azure? You are selecting a set of resources (CPU, Cores, RAM, Networking speeds, Storage) configured together into a VM that will run your workload. As a hyperscale public cloud provider – Azure must forecast customer demand for specific configurations to ensure we have the available technologies and resources deployed in different Regions.

 

What if your workloads needed to leverage non standard resources for those workloads like a faster network between VM nodes or specialized processors and GPUs?  Don’t worry – we got you covered there as well. Certain regions have clusters of specialized hardware that can be included in your workload configurations, just by choosing the right machine type when you deploy. This includes (but not limited to):

  • Dedicated GPU for computational heavy or visual rendering acceleration
  • High speed networking interconnects
  • High speed local storage for extremely low latency and high IO workloads

What about super specialized environments for fluid dynamics or structural dynamic simulations?  These workloads typically would be "locked" on premises due to the specialized networking or hardware components. Azure released a set of new HPC specific VM sizes – the H-Series machines Hb (memory bandwidth) and Hc (dense compute). I sat down with my good friend Corey Sanders (corporate VP of Azure Compute) and special guest Evan Burness (Principal PM in Azure Specialized Compute) to record their chat explaining the nuances of these two new machine sizes and when to use them.

 

 

Cool points I got from this chat?

  • These machines are available now in select regions
  • Industry leading 1.5x to 2x bandwidth from RAM to CPU
  • 1000+ node deployments
  • 100 gigabit network interconnects between nodes
  • More new technology deployments year over year – more to come

This video is part of a series I do called Tuesdays with Corey. Because of various complications – I’ve not had the chance to get more recording slots lined up as frequently as I’d like. Corey has committed to make this work and keep the show going. Are there any topics you’d like to see covered in a future episode? Comment here on the post to let us know what you’d like to see in future episodes and I’ll do my best to track down whoever can come on and talk about your requested topic!