How fast is your Inventor PC really?

[Raider_71]

Hi guys,

We have had to do some testing on a bunch of Inventor PC's recently to determine which of the PC's needs to be replaced. Obviously we needed to find out which of the PC's are the worst eprformers as there was only budget to replace 50% of the design PC's. So we thought the Darwin theory will come in handy right... 🙂

Anyways I started searching on the net for toppics on how to benchmark an Inventor PC. Then I thought whats the point of using gaming benchmark tools because Inventor is not a game and there are more aspects than just graphics performance when it comes to percieved performance on an Inventor PC right.

So we decided to create our own Inventor benchmark tool which tests various aspects of an Inventor PC to give us an overview of our PC's performance. This then helped us make a decision as to whcih pc's to replace.

We have made the tool available free of charge to anyone interested in checking how their PC stacks up to their peers or friends. 🙂

Please download it here and post your results here as well if you want. Would be interesting to see what beast workstations are out there.

I would like to say thanks to Kirk #karthur1, for helping in testing the app.

Please feel free to send any suggestions our way. There is an email link in the app.

Download and Install

The application will work with Inventor 2014 to 2016 only.

IMPORTANT: After installation there will be an Inventor Bench icon on your desktop that looks like this: 

My resluts:

HP Elitebook 8560w with an SSD upgrade.

[Neil_Cross]

This has just gone live if anyone is interested, related to this thread.  I did these tests over Christmas before I found this bench tool, the results would have probably been more surprising if it wasn't for this bench tool! 

Linky

[mcgyvr]

---

@AlexFielder wrote:

 I just tried this system with 1920x1080p resolution and ::drumroll:: the result is almost identical:

Despite the resolution difference between this and the last result I submitted the category totals are almost identical between this and the last run.

::victorydance::

🙂

---

@AlexFielder  Nice benchmark..

Now turn off hyperthreading and try the test again..

For me the INV benchmark was a good bit better with HT turned off.. Having HT turned off got me .7 better on the IPI score and faster across the board for everything else..

[Neil_Cross]

---

@mcgyvr wrote:

---

@AlexFielder wrote:

 I just tried this system with 1920x1080p resolution and ::drumroll:: the result is almost identical:

Despite the resolution difference between this and the last result I submitted the category totals are almost identical between this and the last run.

::victorydance::

🙂

---

@AlexFielder  Nice benchmark..

Now turn off hyperthreading and try the test again..

For me the INV benchmark was a good bit better with HT turned off.. Having HT turned off got me .7 better on the IPI score and faster across the board for everything else..

---

That's very interesting.  The Dell Precision I benched in that video, I also ran all of those tests with HT off and the real world benchmarks with HT off are worse than with HT on, for example creating the two drawing views with HT off took 5 seconds longer to do.  Not significantly slower but I was expecting not significantly faster.

I therefore excluded showing the 'HT off' videos as they didn't improve on the HT on times.

[proj964]

@Neil_Cross

I have watched your latest video on the e5-2687wv3 vs i7-4790k several times.
Thanks for doing this...I think. Maybe it makes things even more confusing.

It would be nice if you could run the e5 tests again with the cpuid hwmonitor
running so you could see the performance of all ten cores.

Could it be that the processor is being thermally throttled because it doesn't
have adequate cooling for the workload?

"Due to varying power characteristics, some parts with Intel® Turbo Boost Technology 2.0 may not achieve maximum turbo frequencies when running heavy workloads and using multiple cores concurrently."
Intel Turbo Boost

HWMonitor is a hardware monitoring program that reads PC systems main health sensors:
CPUID HWMonitor

PS Are either running overclocked?

[Cris-Ideas]

Thanks Neil, I have found this benchmark tool thanks to your other video.

I had run it on my DELL XPS 17 laptop I use. It is an quite old i5 machine, way over warranty period. Still I am quit happy with it.

Here are results:

DELL XPS 17 laptop

Cris.

[Neil_Cross]

Yah you need a monster of a laptop to get high scores from it, my laptop has an i7 in and I think the score was still low around 6.7.  

Do you do work on that? It doesn't seem to have a mobile GPU in, according to the test results you're using the graphics module on the CPU which is as bad as it gets for CAD! 

[Cris-Ideas]

Yes I work on it. It is basically my work and only computer.

I use inventor a lot and either I am used to how fast or slow it is, or what I suspect is the case, it is fast enough for me. Occasionally with some models I am experiencing some operations to take long but it is not always caused by hardware. More often I struggle with software bugs.

So my overall experience is that I can work (model, make drawings, etc.) in a pace that is limited with "clicking rate of my finger".

What takes long and I need to wait is: Updating some drawings, FEA with non linear effects.

As for a graphic card, this is a kind of mystery. Test results show Intel but I also have Nvidia GeForce GT 555M. So I suppose this one should be indicated.

Cris.

[Anonymous]

---

@Cris-Ideas wrote:

Yes I work on it. It is basically my work and only computer.

I use inventor a lot and either I am used to how fast or slow it is, or what I suspect is the case, it is fast enough for me. Occasionally with some models I am experiencing some operations to take long but it is not always caused by hardware. More often I struggle with software bugs.

So my overall experience is that I can work (model, make drawings, etc.) in a pace that is limited with "clicking rate of my finger".

What takes long and I need to wait is: Updating some drawings, FEA with non linear effects.

As for a graphic card, this is a kind of mystery. Test results show Intel but I also have Nvidia GeForce GT 555M. So I suppose this one should be indicated.

Cris.

---

It might be faster for you to VMWare into a computer and use it remotely at that rate.

[Cris-Ideas]

---

@Anonymous wrote:

---

---

It might be faster for you to VMWare into a computer and use it remotely at that rate.

---

Are you making fun of me, or what?

I cannot afford to change my computer now, I am happy to use it, I am happy not having to relay on cloud service, internet connection, and so on.

For what I do it is obviously good enough.

Cris.

[Anonymous]

@Cris-Ideas

I am running a laptop similar to yours but with an I7-3630qm processor (2.40ghz).  The overall scores are similar (1-4.4) to yours (processing a little faster, Graphics a lot slower 28.40 to 47.40). Mine is running the intel hd4000 on board graphics so I would say your laptop must be running the the geforce graphics but not reporting it.

I would rerun the test on my machine and post the results but I am rendering a short video (short to play...long to render) with 3D studio now which brings me to a couple of questions about rendering and Graphics cards -

@Neil_Cross 

Should a quadro card render faster than a Geforce card?  The benchmark doesn't contain any rendering test and I don't know if rendering takes place in the GPU or CPU or how it compares to display frame rates.  Or how rendering in inventor compares to rendering in 3DS.

I have an old desktop that has a Quadro 600 in it.  So I am thinking to myself I will just upgrade the processor and motherboard and have a pretty good machine...then I ran the benchmark on it. It's results (with the quadro 600) on the graphics portion of the benchmark were not quite as good as the laptop running hd4000 graphics?  Sooooo do I scrap the quadro card and just get a gaming card will rendering swing the balance back in favor of a quadro (but perhaps not a 600)?

@Raider_71

I kind of hate to ask because you have developed a very cool benchmark (Many Thanks) but......Would it be possible to add a render to the benchmark?

Thanks guys,

Chuck

[Neil_Cross]

@Anonymous I can't speak for 3DS Max as I don't know the software, but inventor is 100% CPU based for any worthwhile rendering.  Same for Showcase, that uses GPU for its main real time stuff but then Ray tracing uses CPU.  If you check out my video linked above, you can see the difference a beast of a CPU makes in comparison to a consumer chip for rendering.  @-niels- knows a fair bit about Max though, he could chip in.

regarding your Quadro, currently no feature at all within Inventor makes any use of any features unique to Quadro. There is literally no functionality benefit to having a Quadro, the only gain you get is peace of mind that the cards are certified, built by NVIDIA themselves and not a third party, and the drivers are tested by Autodesk.  Your Quadro 600 should be replaced pretty quickly if you can, and if it's your own money I'd recommend a gaming card. 

[Raider_71]

Hi @Anonymous

Thanks for your suggestion about getting another test type inside InventorBench. We have been looking at that already and will include some sort of rendering test mainly because its probably the only part of Inventor that's 100% multithreaded (will use all the CPU cores).

And like @Neil_Cross said, rendering inside Inventor is a CPU based task and not its not performed by the GPU at all.

I am also investigating including a bench test on a larger assembly model.

Regarding your graphics performance and the fact that InventorBench is only picking up the on-chip graphics, please try and set your graphics card settings, using the Nvidia Control panel, to use the Nvidia card for Inventor. Your PC uses Nvidia Optimus which does the graphics card selection automatically (onboard or nvidia) depending on what application is started. You can select Inventor from the list and force it to use the Geforce on next startup.

You could also specify that ONLY the Geforce is used for any application and that the onboard card is never used. Now remember the whole idea of Optimus is to assist in getting your PC to be as power efficient as possible by utilising the correct graphics hardware for the correct application so by setting this your PC will be using more battery power than before.

Also try updating your system and graphics drivers. The newer Optimus tech does a better job at selecting the VGA hardware.

Anyways its worth a try just to see if your PC was not perhaps selecting the wrong VGA card for Inventor. It certainly would affect your graphics scores.

Let me know what you find.

Here is the setting I am talking about:

[-niels-]

---

@Neil_Cross wrote:

@Anonymous I can't speak for 3DS Max as I don't know the software, but inventor is 100% CPU based for any worthwhile rendering.  Same for Showcase, that uses GPU for its main real time stuff but then Ray tracing uses CPU.  If you check out my video linked above, you can see the difference a beast of a CPU makes in comparison to a consumer chip for rendering.  @-niels- knows a fair bit about Max though, he could chip in.

---

I'll chip in then, though i don't know that much.

From what i've observed with 3DS Max it looks like it's also mainly using the CPU to render, based on when i render images it displays render boxes in an equal amount to the CPU cores.

I do know that there are options available to set it to use the GPU depending on which render engine you use, but i've never really tried that.

So, not that much information i guess...

[Raider_71]

Hi @-niels-

Yes you are quite correct the default 3DS Max rendering engine uses the CPU. The nice thing is that I think since 3DS Max 2013 you have the option of selecting a GPU rendering engine as well. So 3DS has options but with Inventor there are no such options and currently the rendering engine is still CPU based.

[Anonymous]

---

@Cris-Ideas wrote:

---

@Anonymous wrote:

---

---

It might be faster for you to VMWare into a computer and use it remotely at that rate.

---

Are you making fun of me, or what?

I cannot afford to change my computer now, I am happy to use it, I am happy not having to relay on cloud service, internet connection, and so on.

For what I do it is obviously good enough.

Cris.

---

No. I am completely serious. You might get better scores using a remote session on a beefy computer.

[Cris-Ideas]

@Anonymous

In such case I thank you for suggestion. Currently I am happy with what I have. But in the future if I have problems with my hardware and will find it not sufficient I will have this option.

Thanks

Cris

[mmaes]

So I am looking at upgrading some components on my desktop and found Neil's video and this thread and decided to run the test on my current desktop setup as well as my laptop. 

My laptop is a Dell M6700

i7-3840QM @2.80GHZ

Quadro k5000m

32Gb 1600MHZ RAM

Crucial SSD

Then I have my old desktop that I built in 2011

Asus P8Z68-V MOBO

i7-2600k CPU Overclocked to 4.4GHZ

GeForce GTX 580 GPU Overclocked to 950MHZ

16Gb G.Skill Ram 1600MHZ

Raidmax 1000 Watt PSU

I did notice differnt frame speeds at different settings (performance vs quality, one color vs gradient background, and 720 vs 1080).  My fastest modeling test time was in the 11.8x seconds and my slowest was 12.3x seconds.  Hardly noteworthy but still some varience. 

All that said, I recently ordered new RAM, G.Skill Trident X 1600 (will be overclocked), another GPU, GTX 580 to run in SLI, and a SSD, 500Gb Samsung 850 EVO.  As soon as new components are installed I will rerun the test (my prediction is very small changes).  In real world modeling and usage (especially in rendering) I expect to see some bigger changes.

The goal for me is to do more rendering with the desktop.  The models we are rendering are very large (15,000 parts in some cases) and will max out my CPU and RAM for hours while Ray Tracing (The image below is from my laptop).

All I can say is it is very discouraging to see some of these benchmark scores.  Knowng the kind of money we have spent on laptops throughout the company and knowing that the performance of much less expensive machines far surpasses our expensive laptops.

[Neil_Cross]

It's not really fair to compare laptops to desktops, even older desktops to newer laptops, but I would have expected better results from that one.  That HDD save time on a SSD definitely doesn't weigh up.

Also, my desktop has 32GB RAM and during Ray Tracing a big assembly of equal size to yours, I didn't get anywhere near using 32GB RAM.  What's your paging file/virtual memory set to on that laptop? For CAD PCs they recommend the virtual memory be set to double your system RAM, so in your case 64GB of virtual memory.

I've never seen an Inventor session alone max out a 32GB RAM system.

[Neil_Cross]

There's definitely something not right with your laptop though.  The laptop I'm on now is inferior to yours in every respect, my CPU is worse, RAM, GPU, it's a cheap £800 Dell Inspiron 17R 7720 SE with an average and old 650M gaming chip.  And it beat your score, just.  It really has no right to be a higher score than your system.

[mmaes]

@Neil_Cross I agree 100% it is not fair to compare laptops to desktops.  My thought was from a cost standpoint, not performance.

As for my Vram, I just checked and it is set to 32Gb.  What do you suggest for these two fields?  This is something I have never modified in the past.  Thank you in advance for any suggestions.