In upFront.eZine #740, we have have for your reading pleasure...
1. Benchmarking Rendering: Desktop vs Cloud
- Desktop Rendering Speed
- First Cloud Rendering Attempt
- Second Cloud Rendering Attempt
2. How Much is a Cloud Unit Worth?
3. Out of the Inbox, and the other regular columns.
You can read all about the business of CAD at http://www.upfrontezine.com/2012/upf-740.htm
Interesting Article, Ralph. Thanks for looking into it. For us ACA users who pay extra but can't use the service (it doesn't render aec objects!), it just seems another slap in the face but perhaps they want to work out the bugs for Autocad first. But what else is there in the latest version. Oh yes the intro says corner windows (didn't that come in on 12?).
I also think your comparison is off. Your render doesn't compare in quality with the online one (no sun & shadows) and also there is no way an online service is going to compete if your render only takes seconds at home with all the form filling out, uploading etc. But high end renders with lots of lights & reflections could see a return, not to mention the idea you can get on with your other work rather than tying up the machine.
The other issue I am concerned about is whether they will support RPC (Archvision) objects as part of the render.
But your main point is that Adesk need to get it together working smoothly and simply otherwise it's just another side show when they still can't fix old bugs in Acad/ACA.
Posted by: Nathan Ellery | Jun 20, 2012 at 10:10 AM
I got a lot of response to my article on cloud rendering, some of it quite passioned. No surprise, there were two camps: those who see that cloud-based CAD work still has its challenges, and hasn't quite reached Dropbox-simplicity; and those who felt my review was unfair.
Of those in the unfair camp, all insisted that my 2.3MB test model was not challenging enough; many said I should have used a more complex model, specifically from Revit. Which makes me wonder....
... is Revit really slow at rendering?
... is cloud rendering not meant for AutoCAD users?
... how do users know where the cut-off is between models simple enough to be rendered faster on their desktop computer. and those complex enough to pay extra to have them rendered on the cloud?
I know that benchmarking is controversial, and so I clearly laid out ALL the specifications for the benchmarking I performed so that readers could take the results at face value. I look forward to others doing the same in the near future so that we can compare results.
Thank you for your input!
Posted by: Ralph Grabowski | Jun 20, 2012 at 10:13 AM
Ralph:
It wasn't just that your model wasn't challenging. Your method was flawed. The model you used had no lights or materials. Those are the elements that increase rendering time. Your test basically compared the time it takes to regenerate the model to the time it takes to upload the file to the cloud, process the request, regenerate the model, and send an email notifying you that it's ready. You basically compared how long it would take you to turn off the lights in your office if you were already there versus the time it would take you to call your wife from another room and ask her to go into your office and turn off the light and concluded that your wife is slower at turning off lights than you are.
Posted by: David Cohn | Jun 30, 2012 at 09:54 AM