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.

[nutral]

Okay, let me start by saying i didn't want to insinuate anything. I just noticed different behaviour in inventor 2020 compared to inventor 2019. Which makes things more logical to me.

Your statement is, Inventor doesn't use GPU.

If that is true, how come my computer gets the same score on 1080p as on 4k, while there are 4 times more pixels in 4k? That proves that the inventor does use the GPU to some extent.

Yes, in your pictures, the 2080Ti has lower framerates than the GTX580. Does that mean that inventor doesn't use the GPU ? No, it just means that in the test, the GPU doesn't have much to do. The CPU has to deliver information for every frame, if it can't keep up with a graphics card, then it becomes CPU limited.

When we get inventor 2020 in the office i'm going to test some really large models in framerate between 2 computers, 1 having a much better graphics card but the same cpu, at 4k. 

Maybe everything what i'm saying is wrong and the only thing different in inventor 2020, is a small change in the graphics engine, increasing the framerate on higher resolutions (which is great).

[nutral]

To add to that, i've just checked a large model on my personal pc. The performance on Inventor 2020 is so much better on really large models (28000 parts) on a 4k screen. That is definitely something that has been improved in inventor 2020.

When i downclock the GPU with 1000mhz to 885mhz. the resulting load definitely increases.

Still though, comparing these numbers an GTX1070 is about 60% as fast as a RTX2080, so only with 28000 parts, on a 4k screen would you saturate a GTX1070. Which makes it a great choice still..

[Neil_Cross]

This isn't a topic which is up for debate, I lectured a class at Autodesk University 2018 on this exact subject, the source material was collaborated with the Inventor development team who sat in the audience and watched, it's an objective fact that Inventor isn't GPU accelerated, at any resolution, it isn't a mystery which needs solved!

You are of course welcome to test yourself, but that's been done already.  See below.  The same FPS test, same PC, same environment, but 5 vastly different GPUs, all produced exactly the frame rates in Inventor... this was one of several tests which yielded the exact the end result.

I appreciate your enthusiasm but you're wading into an area which has already been extensively researched, right here, and the results are all published and verified.  This has all been cross referenced with the Inventor devs, there's no ambiguity.  It doesn't change at 4K either, granted 2020 may be better optimised at 4K but the core application is the same as it always has been.  You need enough VRAM for textures, that's relevant to 4K, but beyond that whether you've got a Quadro RTX 6000 or a M2000 you'll get the same frame rate in Inventor.

[Neil_Cross]

By the way @nutral certainly please do your own research and benchmarks, I wasn't in any way trying to discourage that in fact it's always best that everyone does that if they have the opportunity. 

Just make sure the tests are fairly accurate and fair if you're going to post the results online... for example the results above are as water tight as it could possibly be.  It was the exact same system, same dataset, same operating system, same display, literally the only variable from test to test was the GPU. 

[Anonymous]

Dell Precision M6800

Does this result look right for that workstation?

[Jason.kc]

I did that Graphics card force change (I'm using a HP) and have gone from just under 8 to 8.5-ish on the same machine.

It's certainly throttling for heat load though!

[leowarren34]

Looks about right given the lower cpu clock speed (2.8GHz), ideally need to be at a base of atleast 3.5gHz.

Did you do the benchmark plugged in?

[phlyx]

Ran some benchmark tests (CrystalDiskMark and ATTO) on the Plextor SSD drive that we're using and getting the slow Model Save Time and HDD Test scores for running Inventor Bench.  Not sure how to interpret if there is any issues with the scores these benchmarks show but thought maybe someone here could.

[leowarren34]

@phlyx 

Crystal Disk mark looks more than fine, anything other 1.5GB read and 1GB is excellent and more than enough for Inventor.

Atto also seems fine as well.

You're Crystal mark speeds are about the same as mine, if not better.

[phlyx]

@ leowarren34 thanks but that makes it more of a mystery why we're testing around 10 on Model Save Time with Inventor Bench when other folks are testing around 3-4.  Any clue as to why we'd have such slow save times when we have screaming SSD's?

[leowarren34]

@phlyx 

I was looking back at where your original data was and also noticed you mentioned your user benchmark for SSD had improved...

Have you ran an extensive test in IV bench (like 10 iterations)? As that should minimize variability.

I'd also suggest running the benchmark with intel XTU in the background just to see how thermals are.

If your results are consistently getting worse over time then thermals could be limiting Turbo speeds.

[phlyx]

My benchmarks didn't improve, mine have stayed pretty much the same.  Was the engineer beside me who has the same workstation and his numbers for Model Save Time (HDD) were just a tiny bit little better than mine (mine 9.98 and his were 9.08-9.26) but still stable, but his overall results varied each time he took the test from 8.83 to 11.08 (CHAD1, 2 and 3).  The other test results (NORTH, BWATS, DESIGN, LAPTOP) were from others who posted their results in this thread. 

[e_paquette]

PC Index of 11.18 on a 3 cycle test.

[tgregory3G7FA]

I’d like to thank Neil and everyone for all the info in this thread. It’s been very informative.  I just built/upgraded a machine and it went great.

I’ve been using Inventor for 10 or so years. At my previous company all the hardware was controlled by the IT Dept. (who often got everything wrong). Every couple years they’d spend several thousand dollars each on new HP workstations… which never made Inventor any better.  I still have one as a hand-me-down as my home PC.  A 10yro Xeon 2.83ghz with 10GB ram.  It managed to score an Inventor Benchmark of 3.58 in Inventor 2018.

I came across Neil’s YouTube video for the workstation slayer.  And realized I could easily and cheaply upgrade my PC.  Right around this time, my present company was budgeting for new PC’s for this year. I mentioned that we could build PC’s that were cheaper and faster.  Plus we could pass on our ‘old’ PC’s to other office people.  With an i5, SSD and 32GB ram, these are way more than capable for regular office work. 

For my home upgrade I went with:

MSI B450-A Pro Max: $100

Sabrent Rocket 1TB NVMe PCIe M.2 2280: $100

G.SKILL 32GB Ripjaws 3200MHz: $130

AMD Ryzen 5 3600: $190

And I picked up a used Gigabyte RX570 8gb graphic card: $80

I did the upgrade in steps to prevent any compatibility/driver issues.  And I ran the Inventor Benchmark after each step to see what affects there were.

Original HP Workstation:

IPI: 3.58

Inventor Warm Start: 10.51

HDD Test: 17.45

Modeling Test: 29.72

Graphics Test: 36.49

Drawing Test: 55.87

New Motherboard, Ryzen 3600, & Ram (old mechanical HD & GT430 video):

IPI: 6.42

Inventor Warm Start: 6.55

HDD Test: 12.39

Modeling Test: 9.44

Graphics Test: 34.95

Drawing Test: 21.08

New SSD Drive:

IPI: 6.99

Inventor Warm Start: 5.96

HDD Test: 5.19

Modeling Test: 11.69

Graphics Test: 34.32

Drawing Test: 20.37

New Graphics Card:

IPI: 11.43

Inventor Warm Start: 6.07

HDD Test: 4.63

Modeling Test: 9.20

Graphics Test: 8.72

Drawing Test: 21.21

I’m really pleased and impressed.  For $600 this thing is super fast.

After this success, my company has given me the go ahead to build 4 new work PC’s.  These will be AMD Ryzen 9 3900X, 32GB ram and RX580 8GB graphics.  We'll see how it goes.

[phlyx]

Still would like to hear suggestions on why it seems everyone has model save times in the 4 second range and we have screaming fast Plextor SSD's yet our model save times are around 10 seconds.  🙄

[RichM1]

Do you have update physical properties on save checked or unchecked in the application options general tab? That made a 2.5 second difference for me. Not 6 seconds though.

[Mario-Villada]

---

@phlyx wrote:

@ leowarren34 thanks but that makes it more of a mystery why we're testing around 10 on Model Save Time with Inventor Bench when other folks are testing around 3-4.  Any clue as to why we'd have such slow save times when we have screaming SSD's?

---

You probably already know this but If you have another disk installed, make sure "my documents" folder is pointing to your fast disk. The test saves its file to wherever you have set up "my Documents" folder. So maybe, if you have a mechanical drive and your windows installation points to that drive i.e "D:/Users/phlyx/documents", then change it to C: Drive, i.e "C:/Users/phlyx/documents"

[Neil_Cross]

Far from the best of the best but not terrible for a laptop.

Really poor laptop though, seriously unimpressed with everything Lenovo have on offer at the moment.

Lenovo P53.

i9-9880H

RTX 4000

32GB

[machiel.veldkamp]

Is that laptop running a single 32GB ram stick? Single channel?

That may be why it's so slow?

[Neil_Cross]

Yes, they sent it over like that, no idea why.  I wouldn't say 12 is slow but that's not why the laptop is terrible, it's the thermals which are unacceptable.  I've made a video on it, showing the issue, it's nearly ready to go... but I'm conflicted over whether I should even release the video because it was a laptop sent to me through a demo pool for my day job.  I dunno what the legalities are over that.