As a software developer, I’m required (long story) to run Windows Server as my primary OS. This is my regular desktop operating system, no different than Win10.
I’ve read every post regarding Rhino 7 (the regular desktop app, not a server/license install) on Windows Server.
Some posts claim it’s not supported, while two other posts claim that it is.
One post claims that newer versions of Rhino 7 now support Windows Server for use on AWS. Another post claims it can be used if licensing with a “Zoo” license (either cloud or local).
Is this true; if I license with “Zoo”, can I use Rhino 7 on my Win 2022 Server?
Is there some sort of trial license I can run to confirm this will work properly on Win Server, prior to paying for the full license?
I’m having the same issue.
We have a couple of different machines through VMware, some are on windows 10, but this means we max out at 4 cores, and need to run windows server (as our normal desktop OS in these virtual machines) to get more than 4 cores and then we can’t run Rhino
Just a note for people running Server on their desktop: by default, Server prioritizes background processes, which is different from client apps with a UI. So Server and client are not 100% identical.
That can be somewhat mitigated by changing the process prioritization setting in performance options, but it’s still not 100% the same thread calculus as the client SKU.
Rhino running on Windows Server is used for a different background use.
I think it’s Rhino Compute, but I’m not involved with that project so I’m not positive.
While foreground/background processes are a moot point in my instance (or anyone who has adequate/available processing power), Server far outperforms client versions of Windows, especially beginning late Win 10 and continuing to 11, including Enterprise, LTSC, and highly optimized releases.
Unrelated to that, there are key many reasons why Server is required in a development scenario.
Rhino allows a registry entry that allows Server installs; however, that’s only temporary.
It’s a relatively outdated and ineffective licensing limitation (not compatibility) that McNeel added, likely because they had concerns with the software being shared through Remote Desktop or similar.
I suppose they are not aware that RDP can be easily hacked in about two minutes to offer unlimited sessions on any Pro/Higher version of Windows. Therefore, someone that would intend to use it outside the legally licensed scope, would be just as likely to do so on the non-Server OS.
Do you have a cite for that performance claim? I’d like to see side x side comparable test results to back that up. You’re welcome to run whatever you want, of course, but I question the claim given that we have a single kernel but with different background threading prioritization.
On the same hardware, Server doesn’t provide any performance benefits in the pro audio / musician industry where background processes are taking away needed cycles from real-time audio work that can be traced back up to a root window. I haven’t done a test with a graphics-intensive program like Rhino, but I wouldn’t expect different results there.
It was common back around Windows NT/2000 for developers to run server to get access to multi-processor systems, but after everything fully unified on NT, there was no real reason to.
PS: The FG/BG prioritization is not a small difference.
PPS: It may be useful for some developers (folks use whatever is useful to them), but we all run the normal client here.
Yes, I have empirical data to support the claim or I wouldn’t have made it. As I explained, I’m a developer by trade and performance testing is something we do relatively frequently. I also worked for MS in the late 90’s and I am intimately familiar with the post-NT architecture.
To suggest that identical kernels should offer identical performance is to misunderstand modern Windows architecture completely. Particularly post-20H2 where architecture/policy-enforcement changes have severely bloated Windows and removing that bloat, contrary to past efforts, often makes things worse.
Rhino is not a relatively GPU-intensive software versus iterative rendering, etc. Any rendering app, or a photogrammetry app like Metashape is much easier to observe the delta. Better yet, load the Cinebench benchmark if you’re concerned with GPU-intense benchmarking.
The optimizations you’re referring to re:Server include thread-scheduling, I/O prioritization, and memory management that directly benefits high-iteration workloads such as rendering.
It’s not at all germane to the topic re: Rhino, and I have no interest in arguing the point. If you feel that way, by all means, believe whatever you’d like.
To return to the topic –
Rhino, by default, does not allow server installation, purely for licensing purposes. They offer a workaround; however, it expires with each new sub-version and this is something that should be made permanent.
It’s largely impractical in modern workflows to share sessions at scale in RDP; however, where it is practical, it’s just as easy to do this at scale in Win Pro/Enterprise, for those people who do not honor licensing ToS.
I also do not want to argue or get into a measuring contest, and you are correct that this is somewhat off topic. I had a much longer post drafted, but I’ll just leave it at this:
Windows has changed a lot in the past few decades, especially with the OneCore approach which unified kernel, HAL, core APIs, and more across all SKUs on all processor architectures. Differences are largely in drivers (CPU, chipset, peripheral) provided by companies supplying server-class processors, motherboards, and peripherals, and in configuration.
If folks are considering Server, they should get an eval and do the performance test themselves on the same hardware before committing and not just trust claims on the internet.
do the performance test themselves on the same hardware before committing and not just trust claims on the internet.
Right, as I explained, using a tool such as Cinebench.
I’ve made it clear that people shouldn’t use Internet opinion over empirical data, which is more reason why your post dissuading experts from running Server is unwarranted in the context of this post.
People aren’t paying exorbitant licensing fees for Windows Server simply because they read somewhere that it’s faster. They’re paying substantially higher fees because they are required to use it for a specific use case, which both myself and the recent reply have explained is their basis for the question.
I get it – you’d like to discuss an unrelated topic about performance between server and workstation. Please start a new thread to do that. This thread has absolutely nothing to do with performance, and everything to do with our own individual use-cases. Please don’t take us any further off-topic. Thank you.
