Hi all,
for a large-scale graphic project I am considering to buy new hardware to speed up my work routine.
I am using referenced closed curves from Rhino, fill them with hatches (CREATEHATCH), explode the hatches into single curves (HATCHEXPLODE), wrap these curves around a sphere (SURFACEMORPH), and then project them back into a plane with MAKE2D.
Eventually, I will have to deal with up to 400.000 singular objects (curves) at once.
Before I invest into new hardware, I would like to request advice from the community which hardware and system factors determine the best workflow for this particular task.
the closed curves initially come from Illustrator. I import them into Rhino and then reference them into Grasshoppper.
The rest is a definition in grasshopper.
(That is, currently I bake the morphed lines into Rhino and do the make2D in Rhino. But I was told that with Rhino6+GH includes a make2D component which ideally could be run before baking.)
In my opinion, you are already using a fairly high performance system. How does your 16GB memory work out for your typical task? Are you close to using it all? That might suggest increasing to 24 or 32GB.
As to the CPU, you could improve to more cores and higher clock speed, but I don’t think the average user on the forum (including myself) could advise on whether that will speed things up more than nominally for your workload. Same for your video card. I think this advice would need to come from developers with a deep knowledge of how Rhino does things internally. @nathanletwory, @stevebaer, @jeff, @dale
The single most important and best recommendation I can give is this:
Buy the best graphics card you can afford (try to push your budget) with as much GPU memory as possible… The very minimum would be 4GB of video memory, but I recommend 8GB… and if you plan on using multiple monitors, then double the memory amount.
I realize that your K2200 already has 4GB, but even it is now somewhat outdated and there are lots of better choices out there now. It won’t matter how fast your CPUs are if your display is the bottleneck and you’re doing any frame updates in your calculations…but I guess it also depends on what kind of feedback you want/expect during whatever it is you’re doing.
Also, if you plan on doing any kind of renderings using the new Raytraced mode, then NVidia and CUDA core counts should definitely be on your list.
I have no qualms with Win7…but driver support is diminishing, as is OEM and MS support. The driver model is different between Win10 and Win7, which means driver developers need to maintain two different code branches…and chances are, they’re not all that focused on older versions of the OS, which we’ve seen in many different cases.
If you’re going to be getting latest and greatest hardware, then I highly suggest that you put the latest and greatest software on it. Putting a 10 year old OS on new hardware is a recipe for potential problems…mainly because it was written long before any of the hardware even existed. In fact, certain motherboards today can’t support Windows 7 (I have one here), and you get a message at install time that you will not be able to install Windows 7 on your current hardware.
Lastly, unless you plan on building your own system and installing the OS yourself, I don’t think you can even buy a computer today with Windows 7 pre-installed. And again, if you plan on installing everything yourself, you might run into a situation (sooner than later) where there are no Win7 drivers for a specific piece of hardware or device.
So, in my opinion, putting Windows 7 on today’s latest hardware is a roll of the dice at this point in time.
Thanks Jeff et al for your profound and very helpful feedback! This is very much appreciated.
If we buy the new machine, I can run a comparison/benchmark test for the operation and post it here.