Latency for graphics heavy workloads

%3CLINGO-SUB%20id%3D%22lingo-sub-827240%22%20slang%3D%22en-US%22%3ELatency%20for%20graphics%20heavy%20workloads%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-827240%22%20slang%3D%22en-US%22%3E%3CP%3E%26nbsp%3BI%20am%20going%20to%20be%20spinning%20up%20a%20WVD%20deployment%20in%20the%20next%20week%20or%20so%20%26amp%3B%20I%20was%20wondering%20about%20the%20acceptable%20latency%20for%20graphics%20heavy%20workloads.%26nbsp%3B%20It%20makes%20a%20mention%20that%20150ms%20is%20what%20you%20should%20stay%20under%20for%20Windows%20WVD%2C%20but%20have%20you%20seen%20successful%20graphics%20heavy%20deployments%20with%20that%20high%20of%20latency%3F%26nbsp%3B%20We%20have%20tried%20other%20heavy%20graphics%20VDI%20deployments%20on%20cloud%20hosted%20services%20in%20the%20past%20and%20the%20only%20way%20to%20get%20acceptable%20performance%20up%20to%20that%20high%20of%20latency%20was%20to%20use%20Citrix%20HDX%26nbsp%3B%3CA%20href%3D%22https%3A%2F%2Ftechcommunity.microsoft.com%2Ft5%2Fuser%2Fviewprofilepage%2Fuser-id%2F37638%22%20target%3D%22_blank%22%3E%40Heather%20Poulsen%3C%2FA%3E%26nbsp%3B%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-SUB%20id%3D%22lingo-sub-827420%22%20slang%3D%22en-US%22%3ERe%3A%20Latency%20for%20graphics%20heavy%20workloads%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-827420%22%20slang%3D%22en-US%22%3E%3CP%3EHi%20%3CA%20href%3D%22https%3A%2F%2Ftechcommunity.microsoft.com%2Ft5%2Fuser%2Fviewprofilepage%2Fuser-id%2F213819%22%20target%3D%22_blank%22%3E%40Mike%20O'Shaughnessy%3C%2FA%3E%2C%26nbsp%3Bthat%20latency%20recommendation%20is%20generic%20and%20may%20not%20be%20appropriate%20for%20high--performance%20latency-sensitive%20workloads.%20We're%20working%20on%20more%20granularity%20in%20latency%20recommendations%2C%20but%20we%20would%20also%20love%20your%20feedback%20on%20your%20needs%20and%20target%20scenarios.%20Note%20as%20well%20that%20there%20will%20be%20much%20more%20than%20raw%20latency%20alone%20that%20defines%20quality%20of%20experience%20for%20graphics%20intensive%20workloads.%20Thank%20you!%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-SUB%20id%3D%22lingo-sub-827576%22%20slang%3D%22en-US%22%3ERe%3A%20Latency%20for%20graphics%20heavy%20workloads%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-827576%22%20slang%3D%22en-US%22%3E%3CP%3EThanks%2C%20we%20are%20pretty%20aware%20of%20the%20challenges%20other%20than%20latency%20in%20regard%20to%20graphics%20workloads%2C%20but%20the%20main%20failure%20of%20other%20solutions%20seems%20to%20be%20how%20they%20handle%20latency%20%26amp%3B%20the%20translation%20of%20this%20latency%20to%20click%20lag%20on%20the%20end%20user%20client%20side.%26nbsp%3B%20We%20are%20centralizing%20data%20and%20providing%20VDI%20to%20help%20with%20the%20local%20file%20access%20performance%20required%20by%20auto%20desk%20applications%20such%20as%20Revit%2C%20Civli%203D%2C%20Navisworks%2C%20and%20P%26amp%3BID.%26nbsp%3B%20If%20there%20are%20even%20a%20couple%20seconds%20of%20click%20lag%20on%20these%20applications%2C%20it%20can%20cause%20the%20solution%20to%20be%20unsuitable%20for%20daily%20use%20for%20the%20end%20user%20or%20cause%20them%20errors%20within%20the%20files%20they%20are%20working%20in%20as%20their%20clicks%20need%20to%20be%20very%20concise.%26nbsp%3B%3CA%20href%3D%22https%3A%2F%2Ftechcommunity.microsoft.com%2Ft5%2Fuser%2Fviewprofilepage%2Fuser-id%2F399732%22%20target%3D%22_blank%22%3E%40Neil_Hinnant%3C%2FA%3E%26nbsp%3B%3C%2FP%3E%3C%2FLINGO-BODY%3E
Highlighted
New Contributor

 I am going to be spinning up a WVD deployment in the next week or so & I was wondering about the acceptable latency for graphics heavy workloads.  It makes a mention that 150ms is what you should stay under for Windows WVD, but have you seen successful graphics heavy deployments with that high of latency?  We have tried other heavy graphics VDI deployments on cloud hosted services in the past and the only way to get acceptable performance up to that high of latency was to use Citrix HDX @Heather Poulsen 

2 Replies
Highlighted

Hi @Mike O'Shaughnessy, that latency recommendation is generic and may not be appropriate for high--performance latency-sensitive workloads. We're working on more granularity in latency recommendations, but we would also love your feedback on your needs and target scenarios. Note as well that there will be much more than raw latency alone that defines quality of experience for graphics intensive workloads. Thank you!

Highlighted

Thanks, we are pretty aware of the challenges other than latency in regard to graphics workloads, but the main failure of other solutions seems to be how they handle latency & the translation of this latency to click lag on the end user client side.  We are centralizing data and providing VDI to help with the local file access performance required by auto desk applications such as Revit, Civli 3D, Navisworks, and P&ID.  If there are even a couple seconds of click lag on these applications, it can cause the solution to be unsuitable for daily use for the end user or cause them errors within the files they are working in as their clicks need to be very concise. @Neil_Hinnant