time comparison NX4 vs NX6 1

[w7663]

A few months ago, I switched to NX6 from NX4, and I noticed that a few functions took a lot more time in NX6 than I would have thought, so I switched back to NX4. My PC is a 64 bit machine, and I have dual processors working in NX4 and NX6.
I have just done a time comparison on the same part in NX4 and NX6. The solid had 1426 surfaces.

I extracted the surfaces from the solid:
- NX4 time to extract was 3 min 47 sec.
- NX6 time to extract was 19 min 02 sec.

Deleting the parameters from those extracted surfaces:
- NX4 time = 1 min 11 sec.
- NX6 time = 1 min 54 sec.

Copying those extracted surfaces to another layer:
- NX4 time = 1 min 11 sec.
- NX6 time = 7 min 08 sec.

So, is there some setting I need to make adjust to make NX6 faster.

Any help would be appreciated.
Thanks

 

[mgmug]

This is my experience:
- NX5 and NX6 slow in geometric progression with increasing number of objects to manage
- I work with parties that have over 10,000 articles: NX4 is much faster
- I have tried every possible setting and used the best hardware but have not yet been satisfactorily
- I know many users for the best performance continues to remain with NX4

Even I seek aid!

Ciao.

 

[hudson888]

The test that I always perform is a simple roll-back test. I take a model with a reasonable number of features of different types, sewn surfaces, trimmed bodies, booleans, primitives, extrusions, edge and face blends and simple holes. I roll it back to the first feature and then I put a stopwatch on how fast it can get back to the last feature. In this case an automotive panel with 475 features was used and compared in several different versions using and XP/32 Dell Laptop.

NX-3 00:47.35

NX-4 00:53.15

NX-5 00:44.62

NX-6 00:31.55

NX-7 00:33.62

Next I used the same component to perform the first of your tests, enough to be relevant with 1541 faces.

NX-3 00:57.79

NX-4 00:38.44

NX-5 03:29.21

NX-6 02:59.31

NX-7 02:38.03

So what that really shows is that if your criteria for success is to manipulate existing data rather than creating new models then you have the advantage in NX-4. I think the modelling roll-back test is way more relevant.

This may be a fair test and a fair statement but adding to the mix that later versions will take advantage of newer computing technology there will be some further improvement in their favour when NX-3 is compared with NX-7 running on a 64Bit processor.

That said why anything EVER has to slow down between versions is always a puzzle for users to understand and a question worth asking.

P.S. I haven't received my disks for NX-7.5 yet so that test will have to wait.

Best Regards

Hudson

www.jamb.com.au

Nil Desperandum illegitimi non carborundum