Welcome to the Onshape forum! Ask questions and join in the discussions about everything Onshape.
First time visiting? Here are some places to start:- Looking for a certain topic? Check out the categories filter or use Search (upper right).
- Need support? Ask a question to our Community Support category.
- Please submit support tickets for bugs but you can request improvements in the Product Feedback category.
- Be respectful, on topic and if you see a problem, Flag it.
If you would like to contact our Community Manager personally, feel free to send a private message or an email.
how does cloud compute work with onshape?
JPC204
Member, Developers Posts: 8 ✭
loved the webinar that jon did last week regarding the history of cad and vision of onshape.
one thing i'm trying to wrap my head around is how cloud compute helps onshape. we hear all the time that most cad tool feature are single threaded and can only be processed by a single core. has onshape been written to do this differently and process features using multiple cores for all types of features/actions or is there something inherent about vCores or cloud compute that let a single thread be distributed among multiple cores?
thanks!
one thing i'm trying to wrap my head around is how cloud compute helps onshape. we hear all the time that most cad tool feature are single threaded and can only be processed by a single core. has onshape been written to do this differently and process features using multiple cores for all types of features/actions or is there something inherent about vCores or cloud compute that let a single thread be distributed among multiple cores?
thanks!
0
Answers
Sometimes Onshape does have to do a slow, single-threaded, full regeneration of a model. That regeneration "in the cloud" will perform comparably to a desktop with a high end CPU. We work hard to keep the number of such regenerations to a minimum.
but i hugely agree that it opens the doors to more users and opportunities for improved workflows.
and the future of cloud based simulation or rendering vs end-point based.
If you look at the system end to end, there's a lot going on at once. Although regenerating a single part studio from a feature list has a large serial component, we can regenerate thousands of part studios at the same time. If your assembly imports several part studios, we can regenerate them simultaneously instead of sequentially. A lot of the modeling server, including database access, is parallel.
After we send the result back to your browser, the server is still working on others' models. Meanwhile, your browser can rotate, zoom, and select entities for the next operation.
How often is your car moving, and how often is it parked? Odds are, it's parked 90% of the time. Similarly, a fancy computer for CAD design is not being stressed 90% of the time. Your bought it for peak performance.
The cloud is like a car-sharing service. We don't need 100 computers to support 100 designers, because they will not all be active at the same time. We probably don't even need 10.
Today people have multiple computers and often designers 'own' one of them to cad purpose with 3d mouse and other features. It would be nice to know where to put the money if you wan't to optimize pc for OnShape.
What I have understood in other threads - you can easily spend $1500 to pc not knowing that you could get same performance for OnS with $500 and maybe use some of left overs to speed up connection to net.
These do not have any of the admin challenges and incompatible software issues, nor do they suffer from the computational dilutions and complications and distractions, of the desktop CPU.
Furthermore, they offer hot (or at least warm) swapping of a thread across to a new resource if the current CPU encounters a problem.
This is presumably why, even at this beta phase, it's almost unheard of to lose more than the last feature (at most) when there's a glitch.