Hi guys. Probably this question is common here, but I hve a slightly different doubt: In case of a large assembly on SE (5000-1000 individual parts) which component we should update more urgently, in case of iminent need ? And which one represents more gain about performance (CPU x RAM x OpenGL Graphics)
In my case, my system is a 4-cores Q9650 3GHz, 6 GB ram, 16 GB virtual memory, Video Geforce GTX 550 Ti.
* I only started to deal with large assemblies (and their respectives Drafts). Not CAE. And some rendering, but rarely.
I've seen some Xeon-based computers, even slower than mine Q9650, can offer some improvements due to cache memory and large multi-threading. Is this true for Solid Edge (assemblies, modeling, drafts) ? Or only pure speed (clocks) is more important than threads/cores ?
About graphics, what counts more to have a good performance while manipulating assemblies? Memory? Clock? Bandwidth ? (Considering Quadros and Firepros, obviously).
I remember years ago some of us made a 'benchmark' with a Draft file shared here in community. It can be considered for a initial parameter? Or Draft and Assemblies are treated differently by the system?
I'm still looking for more informaton about these points here on community. (Unfortunately some old posts are not so useful anymore, due to natural changes).
Thanks for any help/comment.
Your request is an ongoing and ever changing question. But the things that are always at the top is Memory and Graphics card. The only thing that's been added to it in the past few years, is switching from the classic Hard Drive to Solid State.
But as always the first thing I would say is Memory. Your 6GB is just barely enough for Windows 7 or 8 so that's screams out as your first upgrade. Secondly your graphics card is more of a consumer grade so hat would be my next upgrade.
So I'd at least double your memory to 12. 16GB would be better and if you can afford it 24+ would help with those 5,000 part Assemblies. Your existing 2X (16GB) virtual memory is set properly, but it's still dumping a lot workload (read/writing) on your Hard Drives, since that's were the VM sits, when working with those large Assemblies.
Then look at getting a professional graphics card. The nVidia Quadro K2000 card at $350 is a good choice.
Then get a couple of SSD drives and use one for the OS and programs, and the 2nd purely for Solid Edge data. This setup also helps with maintaining good backups. I recently upgraded from a 5 year old 128GB SSD due to it slowing downto a 254GB SSD. SSD drives need a lot of "headroom" to stay fast. So make sure you size yours to be at least 2X what you think you need. But the old 128GB was good as dead storage so I kept it in my system as a mirror backup by using a software product called "GoodSync". Great use for an older "stuffed" drive that's not performing daily read/writes.
As far as Xeons go, you won't see any real improvement in performance unless you have a Dual Xeon setup and you're doing FEA or are planning on doing heavy rendering with Keyshot. So your Q9650 or any i7 chips are very fast and will do just fine.
That's my opinion, but there may be some out there with different thoughts and suggestions.
As Bob suggested the the Video card is very important.
I have assembled two new workstation last week and I have selected the Nvidia Quadro K620, this one had the new "maxwell" architecture and should perform as well or even outperform the "older" K2000, price for this K620 video card is approx. $199,-
I have used these new "motherboard mounted" SSD's from Plextor (Plextor M.2 256GB )and one i7-4970K watercooled at 4.5 Ghz and one i7-5820K (with DDR4 memory) at 4.0 Ghz on Gigabyte UD Motherbords..
They run very sweet... no complaints from these users anymore :-)
As mentioned before 16GB is a minimum if you are working with large(r) assemblies.