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.

[mmaes]

Hi Guys, Looking for some advice. 

Currently I am working in a model with 130,000+ parts and as you can imagine I am seeing some lag.  The computer has a 6700K overclocked to 4.9, 64gb ram, and a 6gb GTX1060.  One of our other computers here has a 6700 (non K), a 12gb Titan X, and 64gb ram.  Basically my question is simple, do you think I will see any gains by swapping the TitanX into my computer?  As my computer sits right now the CPU is around 30-50% and the ram is up in the 48 range while I am working.

As a side note, I do not typically work in such large assemblies but I am building a large model that will later be rendered.

Thanks in advance

[Lucas.Rapp]

I can not believe what I read 130 000

🙂 🙂 🙂

[jpms24]

Titan X

Sounds clear to me because of the 12gb ram but does inventor takes advantage of that?

I'm having issues managing 50,000 parts could you share your inventor settings?

[tom_vierling]

TL;DR: No

I believe @Neil_Cross has done testing of this with both NVIDIA and AMD, looking at multiple generations of GPU's and gaming vs workstations and concluded for the most part above a certain point (around the 1060/ RX480 mark if I recall), GPU's are underutilized in inventor in contrast to how CPU intensive it is. The benchmark seems to agree, even though it's in its early stages and will be balanced in the future. Therefore, it's safe to say that moving to a titan would give you no tangible benefit in an inventor scenario. Neil might have a video (or is working on one) going into more detail about this.

[Neil_Cross]

Hopefully this eliminates any doubt over what a graphics card upgrade will do to Inventor.  Feel free to point anyone towards this if... when the question is asked in the future. 

[testdastuff]

So everyone here is going on about their GPUs. I haven't done much in Inventor yet, just starting out and wanted to encourage people without GPUs.. it works as well. Specs of my custom "built for Photoshop and Lightroom" machine:

i5 6600k @4.0GHz 

16gb RAM 2400, single slot

Intel HD Graphics 530 onboard igpu

SanDisk Ultra II 256GB SSD (somehow the benchmark detected my pen drives as main Drives 😄 )

Probably it'll slow down if you have big complex stuff, do stress simulations and so on but for now, works for me.

Cheers, and thanks for the great benchmark tool @Raider_71.

Janik

[proj964]

I built this system back in Sept 2016 and didn't get around to adding it to the forum.

It's yet another data point in the Inventor Performance curve...

Better late-than-never? -- only you can tell.

PC Parts Picker list is attached as a .pdf.

[mmaes]

Hi I posted at the top of this thread page about the large assembly I am working with and opted not to swap to the Titan.  That being said I am currently building four new computers all of which will have the I7-7700K, one with a GTX1080 and the other three with GTX1060s. 

I have a couple quick questions.  Has anyone done any bench tests with the 7700k?  if so were the gains over the 6700k notable?

My last question is, has anyone tested an I7-6850K (maybe there is a better option than this CPU?)?  I've been toying around with building a computer with one so I can have 128gb ram.  The more I continue to add on and work on this large assembly I've mentioned, I end up running out of my current 64gb and wanted to see if bumping up to 128gb will help me at all.

Any advice is appreciated. 

Thank you

[Neil_Cross]

I don't think anyone has posted a test of the 7700K yet, but based on synthetic tests it's not really a significant improvement over the 6700K more of a standard incremental gain.  It's not even significant enough to make me want to change out my 4790K which would need a new mobo & RAM, a significant cost for not a lot of gain.  The only thing of note is future Intel Optane support which again wouldn't really benefit Inventor directly, more a system wide thing.

I haven't tested a 6850K but I've got a close Xeon equivalent in a workstation (1650v4) over here.  It's as you would expect, a good all rounder, but won't win at this test due to the lower clock speed.  

If you're needing to pile up the RAM for massive assemblies, have you considered the Xeon range? 

[tom_vierling]

I'll be sticking with my 4790k at home as well. I probably wont upgrade until Icelake/Tigerlake (2018/9)

[OSULemon]

Thanks, Neil. Looking back at my laptop workstation results, I wouldn't be surprised if the CPU was one of the primary causes of my performance issues. 

[mmaes]

will be testing the 7700k with three different graphics cards to see if there are any noticeable differences

[josh]

Just to back you up with the graphics card thing...

https://www.youtube.com/watch?v=LC_sx6A5Wko

[brotherkennyh]

Hi Neil,

I just watched your video. Thanks for the great review.

I have been saying this for years. I noticed GPU-Z barely went over 1GB video memory. Interesting as you even had a large assembly there. What a waste my K2200 is in this workstation. I ran GPU-Z myself while moving a 3000+ part model and my graphics card doesn't break a sweat, very similar results to your own. The CPU still has to calculate component positions before the graphics card can render the image. My 11 year old laptop still does ok with 200-300 part models and it only has an AMD x1400 graphics card.

If only Autodesk would spend some time on performance optimisation so we could make more use of our graphics cards. I would be happy if all they did in the next release was improve performance if it did remove the CPU bottle neck.

It actually annoys me that high end Quadro cards are touted as the best for Inventor as this is clearly nonsense. The only reason these may see stability here is that they never do enough work to run hot or be maxing out the GPU.

My desktop PC has a 6790 graphics card, which is 5-6 years old now I think. When I built my desktop I put in a FX-9590 and the Samsung SM961 512GB NVMe. I used the 6790 from my old PC to save on budget up front. This PC is awesome for Inventor, even with the old graphics card in it. I have never measured the FPS, but you can just tell it is good by how smoother the experience is.

The Inventor bench mark tool gave my be quite a low score. Around 7 I think(from memory), which I attributed to the old graphics card. I think the tool weights score too much to graphics performance.

[Neil_Cross]

Thanks man.
Although the results won't change much, I need to redo all those tests for a formal log for comparison. A lot of the GPU activity and VRAM consumption you saw on GPU-Z was actually the screen capture software XSplit which itself knocked about 20% off the performance of Inventor because it's heavily CPU weighted itself as well as using GPU. So I need to redo all the tests without Xsplit running which won't change the message, just corrects/tweaks the numbers.
Also regarding the historical Quadro recommendations and Quadro/FirePro being the only certified cards, off the back of my video all I can say is watch this space, I can't really say any more than that but the issue is firmly on the agenda for discussion now.

[June89]

@Neil_Cross Hello Neil,

Thank you for all your videos, they have been of great help especially Vault ones.

From the test it is clear that I need to upgrade my RAM only 16GB and move to SSD instead of HDD.

Also,I want to build a CAD workstation for both Inventor(70%) & Solidworks(30%), after going through this discussion and your videos here is what I'm thinking.

CPU : 7700K @ 4.2 GHz

RAM: 32GB

SSD: M.2 vs PCIe ?? (Your suggestions)

GPU: AMD Radeon RX 480 vs GeForce GTX 1070 ??  (Your suggestions)

Thanks,

Arjun

[Neil_Cross]

Hi @June89 you're welcome, glad they've helped.

Regarding your questions, the 1070 is by far more powerful than a RX 480 but you'll not see any gains in Inventor, I'm not sure if Solidworks supports those cards though.  Doesn't it disable features if there's no professional card installed? The 1070 is also double the price of the RX 480 but it is a better card, that's what I have.

For the SSD, you might want to research that a little further.  M.2 refers to the physical form of the SSD, whereas PCIe refers to how the SSD communicates with the CPU.  So it's normal for a M.2 drive to also be PCIe based.  But you can also have a M.2 drive which is SATA based.  Normally PCIe means a faster drive but you need to then check the type of M.2 slot on your motherboard, is it a SATA M.2 slot, is it PCIe2 or PCIe3, does it have one at all?

[brotherkennyh]

Hi @June89,

Wow, you must have a great life you can join in the Inventor vs Solidworks debate on both sides of the argument, depending on who has ticked you off that day.

You are basically building my ideal Inventor PC there. The 7700K is at the top of the single core performance charts and fair well in the overall charts. Perfect Inventor CPU until the devs work out decent performance optimisation (when pigs fly). Personally I think the graphics card is overkill and if you watch Neil's video further up the post you will see why. Still if you use any other graphics hungry applications it may be worth it if you have the budget.

I echo Neil's comments about the SSD. I believe PCIe is the faster, at least I have not seen any that use SATA at the top of the performance benchmarks. I have a PCIe SSD and it rocks. It also uses NVMe, which is a protocol rather than an interface, which also means it is faster. My drive is an M.2 drive, it has an adapter card to plug it into the PCIe slot.

Be aware that you will need a PCIe 3.0 slot. The drive will work in most PCIe 2.0 slots, but you will limit the performance. I would be fair to say that even in a PCIe 2.0 slot it will spank a normal SSD. Also be aware of the number of lanes it requires, x4, x8, x16 etc. and how many lanes are available to each slot on your motherboard when you have multiple cards installed. When I first installed my m.2 SSD I inadvertently installed it into a PCIe 2.0 port with not many lanes available (I forget how many it was) and it had similar performance to a normal SSD. I then moved it to another slot and it more than tripled the performance.

You should also be aware that on some motherboards with a SATA m.2 slot a SATA port can be disabled when the m.2 port is used. I don't know if the same is true with boards having a m.2 PCIe slot.

Cheers

Kenny

[Raider_71]

Hi guys,

Ok I finally got my Intel 7700K system going and I thought I must share my findings here.

Some of the components in this system:

CPU: Intel 7700K

Mobo: GIGABYTE Z270-HD3P

VGA: Varies

Mem: CORSAIR 32GB DDR4; 2400MHZ

Drive: SAMSUNG MZ-V6E250BW 960 EVO

I have decided to test three different graphics cards (all I had lying around) to see the effect on overall performance score.

These were the cards I have tested in this system: Quadro 600, Quadro K620, GeForce 1070.

Now, I have mentioned before that in my opinion the current test sequence of InventorBench is a little too much CPU focussed i.e. there are too many CPU based tasks contributing to the IPI score obviously making the IPI a bit CPU biased and I feel that this may have created the impression that Inventor is “very” CPU dependant. Inventor is CPU dependant but if you isolate the graphics results (see images below) and look at the effect of changing your graphics card out you can see that it does have, in my opinion, a substantial effect.

The next version of InventorBench will be more balanced but will also include some more advanced tests like larger assembly modeling and some more graphics related tests as well.

I have to also mention that the system was left “standard” with no overclocking. The only thing I did do was enable the XMP memory profile at bios level to ensure correct memory timings for the Corsair memory I put in. So this is a stock system and identical except for the graphics cards which were changed and obviously drivers updated for each card.

Here are the test results:

Quadro 600:

Quadro K620:

GeForce 1070:

Below is a comparison between the effect the cards had on the Inventor graphics performance VS their individual PassMark score as per www.videocardbenchmark.net

There is a definite correlation between the results I would say?

Anyways I just thought I’d share but as always, please feel free to comment. 

Cheers

Pieter

[Neil_Cross]

Interesting.  In my video using real world live FPS monitoring I got the complete opposite results:

You can actually see these results happening live in the video with each card shown installed.

I'm still of the opinion that your test didn't create the impression of CPU reliance, it just emphasised it.

For example, the PC used in the image above has a Xeon 2.6Ghz, with the GTX1070 installed.

I then put the very same 1070 into a different PC with a 4790K @4.6GHz and this happened:

Double the real time live FPS using the same card in a different PC.  Again this is shown here: