technical requirements for Creo 2.0

[ptc-4970910]

Hi,

I am about buying new laptop.

I would like to find out what are technical requirements for Creo 2.0 that I should pay attention to?

Is there any available technical specification what computer should have?

I am thinking about two models

1) Asus N550JK-CN208H which has

Intel Core i7-4700HQ

8GB RAM

GeForce GTX850M 2GB

FullHD

2) Lenovo Y510P

Core i7-4700HQ

RAM 16GB

GeFroce GT750 2GB

FullHD

Cheers,

M.

[jsarkar]

I am also placcing to buy Asus X550LC-XX223D Laptop.

• Memory (Ram) : 8 GB

Dedicated Graphics Memory : 2GB NVIDIA GeForce GT 720M Graphics

Not sure will it be ok with PTC software or not?

[JonathanHodgson]

Jayanta Sarkar wrote:

Dedicated Graphics Memory : 2GB NVIDIA GeForce GT 720M Graphics

Not sure will it be ok with PTC software or not?

Apart from the usual questions about "certified hardware", the 720M is a pretty low-end card:

http://www.notebookcheck.net/NVIDIA-GeForce-GT-720M.90247.0.html

In comparison to the competition, the 720M offers a performance similar to the HD Graphics 4400/4600 or Radeon HD 8570M. Overall, the card is placed in the lower performance segment.

[ptc-4970910]

Thanks for advise, although they are out of my budget.

I am considering pay up to 1000euro/1300USD.

My types:

#1

Lenovo Y510P (59-407442-128SSD) za 4000zł

i7-4700MQ

RAM 8GB

HDD 1TB

SDD 128 GB

Full HD 15"

GeForce GT755M 2GB GDDR5

win 8

#2

MSI GE60 20E (045XPL-3900) za 3900zł

i7-4700MQ

RAM 8GB

HDD 1TB

Full HD 15"

GeForce GTX765M 2GB GDDR5

win 8

#3

Acer V3-772G (NX.M8SEP.02) za 3900zł

i7-4702MQ

RAM 12GB

HDD 1TB

Full HD 17"

GeForce GTX760M 2GB GDDR5

win 8

[JonathanHodgson]

Hello Mich,

Ignoring "certification", some general comments:

• The 4700MQ is a reasonably quick CPU. The 4702 is slightly slower, but still decent.
  http://www.cpubenchmark.net/singleThread.html
• 16GB is probably overkill, unless you do FEA / CFD type analysis, or load really big assemblies. 8GB should be enough for most work, I would think.
• An SSD is definitely nice for general system responsiveness.
• The 765M appears to be noticeably quicker than either the 755M or 760M
  http://www.videocardbenchmark.net/gpu_list.php
  but I suspect that any of them should be OK for most normal Creo use (based purely on performance - not compatibility, as above). In general I haven't found that Creo uses the graphics card heavily - it's mostly CPU-limited unless you're doing rendering, or turn the anti-aliasing way up.

So, the Lenovo you list has the SSD; the MSI has a faster GPU; and the Acer has a bigger screen...

HTH!

[jsarkar]

Thanks Jonathan for your valuable reply.

Regards,

Jayanta Sarkar

[BrianMartin]

Hi Jayanta...

There a few things I would suggest:

• Get the 16GB of RAM. Regardless of what you're doing (unless it's small assemblies or piece parts), you're better off with 16GB. It's not that Creo will necessarily require all of that RAM, but standard business applications like Microsoft Outlook are now chewing up ever increasing amounts of RAM. On workstations at my 'company', we routinely see Outlook eating up 3GB of RAM. By the time Creo gets to the table, it's dealing with far less than the 8GB you think you've allotted. Go with 16GB... it's cheap... and you won't regret the investment.
• Lenovo... in my personal opinion, stay away from them. There are reasons government agencies will not use these in secure environments (google it). My 'employer' will not use their hardware. We did accidentally procure a few workstations for training purposes but they were positively loaded with pre-installed adware that was very difficult to remove. Give me a Dell or an HP any day.
• Solid State Drive - You will notice a difference upon boot and launching applications. It really does make things 'snappy'. However, also get a backup secondary traditional drive. The SSD's are typically not very large (unless you want to spend quite a bit of money) so give yourself enough storage with a secondary drive.
• There is a point of diminishing returns for all computers. You do not need the top of the line processor... the second or third from the top is a better value. Get a good graphics card (2GB of RAM would be considered very good). Some people go for 4GB or more... in my experience supporting dozens of customers, we've actually seen degraded performance for cards over 4GB (not sure why yet). I'd postulate that perhaps those cards aren't tested as rigorously by PTC (not many people can afford them) therefore problems slip by? Just a theory...

As others have suggested, always research the Platform Support and System Requirements sections for whichever version of Creo you'll be running on the ptc.com website.

Good luck...

-Brian

[Dale_Rosema]

Brian,

I wish I would have seen your post about Lenovo before someone in our family bought one. I agree with the all the adware that is difficult to remove.

Thanks, Dale

[BrianMartin]

Just to make sure I'm not unfairly maligning Lenovo... I should clarify what I meant by "adware". On our training systems, whenever you opened a browser or even the DOS command window, a small "assistant" would pop up along with it. This was in the form of a small avatar of a 'customer service' rep with a headset waiting to 'help' you. If you clicked this little rep picture, you were provided will all sorts of helpful assistance about how to buy software products, receive support, etc. The thing is, you couldn't get this junk to turn off. I tried repeatedly to uninstall this tool and it refused to uninstall citing some odd error.

So... while it's possible only our set our training computers was pre-loaded with this terribly intrusive software, my personal experience would disincline me to purchase hardware from this vendor in the future. I'm sure the people at Lenovo are good to their mothers, pay their taxes, and are truly the salt of the earth... but I've had a bad experience with their hardware and I've never had a similar experience with other hardware vendors.

[jsarkar]

Thanks brian, really helpful and valuable information your provided, I am now looking for 16 GB ram laptops.

Regards,

Jayanta Sarkar

[ptc-4970910]

Thanks for your concrete reply.

I take it into concideration.

[Inoram]

Read how it is written in the PDF Vlad attached. It's what I call "gray area". But really PTC isn't going to "support" any hardware. I think their wording is a little misleading.

[VladimirN]

Hi,

Please find in attachment "Creo 2.0 Hardware Notes".

Hardware Notes - Creo 2.0

Parametric, Direct, Layout, Schematics, Options Modeler, Simulate

Table of Content

Last updated: April 10, 2014

 Platform Support

 System Requirements

 Graphics Information

 Certified and Supported Graphics Cards

 Desktop Virtualization Environment Support

 Supported Peripherals and Accessories

 Supported MCAD Systems

 Supported Finite Element Solvers

 Platform Support for Data Exchange

[dgschaefer]

Generally, GeForce cards are gaming cards, not well suited for CAD work. Creo will do OK with a few windows open but open too many it'll come to a crawl. I'm thinking more complex geometry models will choke a gamer card too, but I'm not certain.

You really want a workstation class graphics card, the Nvidia Quadro line is the one most prefer.

My understanding si that Creo doesn't much care about video memory like video games do, that's why more video memory doesn't help performance.

[Inoram]

I remember back in the day and just googled it again, people are still hacking GeForce cards to make them Quadro cards by changing some resistors. So I don't know how much the hardware is really that different. When I had them side by side a few years ago the only difference I could see was the Quadro seemed like wireframes looked smoother. I'd like to see what a comparison of the same gen Geforce to Quadro, but I doubt that will ever happen for fear of hurting the Quadro branding.

If someone has more experience with the same gen cards side by side, let me know though. I'd like know what others have seen.

[JonathanHodgson]

I'm sure I read somewhere that gaming cards are optimised for DirectX/Direct3D whereas workstation cards are optimised for OpenGL. Not having any knowledge of the difference between these, I have no idea whether that's true or even plausible...

[Inoram]

http://www.xbitlabs.com/articles/graphics/display/nvidia-quadro-k5000_7.html#sect0

[JonathanHodgson]

Matt Griswold wrote:

http://www.xbitlabs.com/articles/graphics/display/nvidia-quadro-k5000_7.html#sect0

Interesting - and also interesting to compare with the synthetic OpenGL benchmark in which the k5000 destroys the GTX 680:

http://www.xbitlabs.com/articles/graphics/display/nvidia-quadro-k5000_4.html#sect1

I wonder what Pro/E benchmark they used - and which version of Pro/E...

[JonathanHodgson]

Trying to keep on-topic, though, it's probably not worth getting too hung up on getting the fastest graphics card possible.

Among the top ten (32-bit) scores on Olaf's benchmark:

http://www.proesite.com/cgi-bin/ocusb6.cgi

are several using a Quadro 600, which has just 96 CUDA cores compared to the K5000's 1536; and only 25.6 Gb/s memory bandwidth, which is pretty poor by gaming standards. Its Passmark score is 684 compared to 4022 for the K5000.

Therefore I would suggest that as long as the graphics card is good enough, it's not particularly important compared to the CPU. Having a CPU with fast single-thread performance is key.

[Inoram]

Yup. I love the Quadros, but it seems like there is a belief that Pro/Creo is severely crippled without one. And I just have never experienced that personally in many years of running both cards on many systems. Which goes back to the original post, if someone is looking for a budget laptop, Quadro is no where near that price range.

ALSO, at the end of the xbits review there is a comment about mini drivers for Creo2. And I have no idea what they are talking about (unless that is something built into the main drivers), Autocad had that years ago, but I have never seen that for Creo.

[neomechanikos]

I think they are referring to SPECapc it's just a paid benchmark that goes into more detail and looks like it wasn't updated at the time for the newest cards out there. So it doesn't take into account the professional level capabilities of the new cards....anti aliasing, floating point precision, etc. So in essence, I think it's the mini driver for their benchmark.

Where the quadros really shine is in large assemblies. I recently purchased a used Quadro 6000 6GB on ebay for cheap and the difference it made on some of my large assemblies (3000+ parts) was incredible compared to the Radeon gaming card that I was using previously....the difference "felt" about like what the difference in specviewperf scores are....about an order of magnitude. Of course, other hardware was updated as well which I'm sure helped.

[Inoram]

I assume they meant something like this:

http://www.nvidia.com/object/AutoCAD_PD_workstation.html

Grats on the 6000! (6000 not a K6000 though?).

Yeah it's hard to say what it did if you changed other hardware too although the 6000 should easily outperform the Radeon if was an older model.

[neomechanikos]

K6000 is an order of magnitude more $$$$ than what I got the 6000 for. CPU is a Xeon E5-1650 V2 3.6GHz OC'd to 4.2GHz, 32GB ECC RAM

[JonathanHodgson]

So, when you're spinning one of these 3000 part assys, what total CPU and GPU load do you see now?

(Looks like the E5-1650 is a 6-core so ~17% would be single-thread limited...)

[neomechanikos]

4413 part assembly.

CPU Load: 7%

GPU Load: 20% to 90%, .2 to 1.6 GB VRAM

[JonathanHodgson]

Have you got hyperthreading enabled? I'm wondering whether 7% is one thread (out of 12 possible)... otherwise it's limited by something else that's neither CPU nor GPU.

This does show that having 2 GB of VRAM may be helpful.

[neomechanikos]

Yes, hyperthreading was enabled at the time. I will probably be turning it off in the near future though. 7% was the average across all cores. If I recall correctly, the individual cores were sporadically loaded, max single core thread was about 20% but it jumped around through the cores and some cores shared some of the load.

[JonathanHodgson]

I use the "One graph, all CPUs" option to see the total load. Windows seems to share the load very actively across cores so it's impossible to read anything meaningful from the individual graphs - although just for an experiment, I suppose you could try temporarily setting affinity to just one core (to see if it maxes out) and then to two (to see if it helps).

[Inoram]

It seems like you guys need to come with something common if you want to figure out some kind of benchmark, like the borg-cube thing we were messing with a couple months ago.

[JonathanHodgson]

It's not so much about benchmarking, rather trying to work out whether an 'amazing' graphics card actually has any benefit over an 'OK' one; or whether the CPU is the limiting factor for all reasonable system specs (and therefore there's no point spending more money on the GPU).

We already have Olaf's benchmark (and results table):

http://www.proesite.com/cgi-bin/ocusb6.cgi?64bit=1

but unfortunately the top few systems in the (bigger) 64-bit benchmark all have very high-end GPUs as well as massively overclocked CPUs; and the first system to have a lower-spec GPU also has a CPU that's way below the top-scoring systems.