Community
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.
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.
If anyone is interested, my budget system build in full along with commentary is linked in the image. This is YouTube so it's aimed at all the folk who don't know about this bench tool or this forum thread.
New toy, Alienware 17 R3 laptop.
Skylake 6820HK overclocked to 4.0GHz
NVIDIA GTX 980M (Bench tool reporting on the CPU graphics but it is running on the 980M)
32GB RAM
PCIe M.2 Samsung NVME 512GB SSD
4K Screen running at 1080p because Microsoft Windows and Autodesk don't know how to 4K.
@-niels- Yep got the graphics amplifier, and the AW Vindicator backpack 
Which is quite sexual but a bit tight for space.
Not sure about the non-overclocked score, I've ran it on a permanent overclock since I received it, but it wouldn't be a like-for-like with yours as off memory I think you got the 6700HQ CPU? I can do a bench test if you're interested in seeing it, but it would be the 6700HQ & 970M vs 6820HK & 980M.
I've briefly tested the FirePro W9100 in the amplifier but I'm not sure how it did, I don't think AMD have done the necessary drivers for it, I opened Showcase and on face value didn't notice too much of a difference between that and the 980M. I then got a mail from a guy who works at AMD who said they've just released their Xconnect external box, it isn't FirePro certified yet but it probably will be soon. To be honest though, the 980M is so good that I'm not sure if I have anything which will be able to max that out and then see any extra benefit from the W9100. But I enjoy trying!!
This is non-overclocked, stock speed at 2.7 with a 3.6GHz turbo. Amazing what 0.4Ghz can do!
The overclock on these is an Alienware default, they put the profiles in the BIOS and have 3 levels of overclocking, unfortunately I can only use level 1 & 2, if I use level 3 which is meant to boost it to 4.1GHz the system blue screens on the regular. So I can only go to 4.0GHz.
Also what's strange with these overclocks is the AW profile sets all 4 cores to different speeds, 3.7, 3.8, 3.9 & 4.0GHz. I'm sure there's a reason for that but I read in an OC guide that it's safer to set all cores to the same speed for increased stability.
Hi guys,
This whole thing about the app not reporting the graphics card correctly got me slightly worried and I was concerned that maybe, just maybe, Inventor was in fact not using the correct graphics card during testing. I thought let me build up a PC to do some testing. I wanted to confirm that even though InventorBench was reporting the on-board graphics, that the actual dedicated GPU was being used. I ran InventorBench on a system with a Quadro K620 installed and as reported by some of you as well the app reported the on-board graphics card. I recorded the test result and then rebooted the PC into BIOS mode and disabled the on-board graphics at hardware level. Booted back into Windows 10 and then started InventorBench again. This time the graphics was reported correctly as the Quadro K620. I ran an extensive test again and the score was within 1% compared to the previous score when the graphics card was reported to be the on-board one.
Ok so that has put my mind at ease about that topic. So for any of you having the same problem, just reboot and disable the on-board graphics in the Bios and you should not see it represented any longer.
Here are my results comparing the difference between the GeForce GTX 780 and Quadro K620 on the same system. As you can see the graphics card makes a huge difference right!?!? Compare the graphics scores only.
Anyway so, as a second point, this got me thinking some more about the overall perception I think many may have that the CPU has the biggest impact on Invetor’s performance. Even though this may or may not be true, I just want to point out something. I started thinking about the way InventorBench is using the results to create the overall IPI value and I feel that maybe the way it’s currently done is a bit biased towards the CPU. It’s not really an issue but its more about how the test results are calculated relative to each other. You see, the “Modeling Test” and the “Drawing Test” is actually CPU based tests and the faster your CPU the better your PC will do in these two areas right. Only the “Graphics Test” is aimed at testing the GPU. So currently the IPI is calculated on the total time of all these tests and because there are two CPU tests influencing this result, it may seem that changing the CPU for a faster one has the biggest effect on your PC’s performance if you know what I mean. I mean these are all legitimate tests where the modelling test will test the raw power of one CPU core and the drawing test will again show off the power of multiple cores so the more cores the better the result will be here.
Also in the near future I want to include a rendering test which again is a CPU based test and will also contribute to the overall IPI score of course. This will more so overshadow the contribution the graphics card has on the IPI score.
Anyways I am just mentioning this to see what you guys take is on this. I think a good idea is to develop the graphics based tests some more as well so that it contributes some more to the overall result and maybe implement some sort of weighting.
Cheers
Unless it's a laptop it's not possible for the system to switch to the onboard integrated graphics...? If a monitor is fed from the PCIe card then it can't possibly use the onboard graphics as that's fed only through the motherboard port. But I suppose it was worth clarifying the curiosity.
What I'd also say is I think your test has the perfect balance. Even for the graphics test, that's still heavily CPU based. There's a test result somewhere hear from a guy with a Quadro K6000, that card has 12GB RAM and 2880 frickin cuda cores amongst other big numbers and specs, and his Hz scores were around 112Hz. My £350 budget PC has a cheap AMD R9 380 in with 2GB RAM and 1792 cores along with smaller specs across the board, my Hz score was around 180-190Hz. If the Hz scores are mainly GPU dependent, we would surely see consistency with the bigger GPUs pulling the biggest Hz scores, but that's hasn't happened. If the big GPUs are in a system with a low clocked CPU, the Hz scores are way down.
I think another good test for graphics cards would be to try and manipulate the VRAM. When Inventor hangs for a while during a zoom in/out, that's the system reading and writing to the GPU VRAM, good GPUs with fast and plentiful VRAM and high memory bandwidth are going to score better at that than a poor card like the K620 or the likes of the old Quadro 2000's. Plentiful VRAM, GDDR5, high bandwitdh cards are going to be able to excel at that test whilst the lower end cards should be noticeably worse. You'd need to find a dataset which can ask those questions and consume a fair amount of VRAM, things like enabling the visual effects will test the VRAM as well as having multiple models and drawings open in a session. If possible, try and fabricate a scenario whereby the system demands over 1GB of VRAM, this isn't very difficult at all and will hammer those older cards such as the Quadro FX range which have less than 1GB of VRAM.
Forums Links
Can't find what you're looking for? Ask the community or share your knowledge.
Post to forums
