Puget Systems print logo


Read this article at https://www.pugetsystems.com/guides/1347
Article Thumbnail

SOLIDWORKS 2019 Intel CPU Performance

Written on February 8, 2019 by William George


Dassault Systèmes launched the initial version of SOLIDWORKS 2019 late last year, but with the recent release of SP1, we expect that customers will soon be using it in production environments. In preparation for that, we have tested the field of current Intel Core and Core X series processors to see how they compare across a wide variety of tasks within SOLIDWORKS. For this round of testing, we stuck with Intel chips only, since AMD Ryzen and Threadripper testing in past years didn't look great, though we may re-test with their 32-core another time.

Test Hardware and Methodology

To see how these different CPUs perform in SOLIDWORKS 2019, we used the following configurations:

The tests conducted on these systems were originally developed by my colleague here at Puget Systems: Matt Bach. He put together a series of AutoIt scripts that run through testing a variety of the capabilities in SOLIDWORKS, so rather than reinvent the wheel I used his. Only minor updates were needed to bring the scripts up to speed for SW 2019, but we did take this opportunity to add a more demanding flow simulation test provided by a contact at Dassault. He described it as a "Conjugate Heat Transfer Airflow" model, but I'm just calling it our "benchmark simulation". It is run at three different mesh sizes, to see if that has any impact on performance scaling.

Each test was run three times on each CPU, with the fastest (lowest time) of the three used for this article. We didn't have any significant outlier results and saw very little variance between runs, so we opted for this method over an average of scores. The results are broken up into their own graphs below, along with a chart showing the raw data, and followed by our analysis and conclusion.


Here are graphs for each of the tests we conducted - click to enlarge - along with a chart of the raw result times for each processor:

SOLIDWORKS 2019 Intel CPU Performance Test - File Open

SOLIDWORKS 2019 Intel CPU Performance Test - Rebuild Assembly

SOLIDWORKS 2019 Intel CPU Performance Test - Irradiance Pre-Pass

SOLIDWORKS 2019 Intel CPU Performance Test - Rendering

SOLIDWORKS 2019 Intel CPU Performance Test - Motion Study

SOLIDWORKS 2019 Intel CPU Performance Test - Stress Simulation

SOLIDWORKS 2019 Intel CPU Performance Test - Airflow Simulation

SOLIDWORKS 2019 Intel CPU Performance Test - Thermal Simulation

SOLIDWORKS 2019 Intel CPU Performance Test - Conjugate Heat Transfer Airflow Simulation at 500K Mesh Size


SOLIDWORKS 2019 Intel CPU Performance Test - Conjugate Heat Transfer Airflow Simulation at 1000K Mesh Size

SOLIDWORKS 2019 Intel CPU Performance Test - Conjugate Heat Transfer Airflow Simulation at 2000K Mesh Size


SOLIDWORKS 2019 Intel CPU Performance Test Results Chart


There are two fairly clear trends in the chart and graphs above. The features we have tested in SOLIDWORKS 2019 fall into one of two categories: those that benefit primarily from clock speed and those which instead scale with core count. File opening, rebuilding, motion study, and some of the simulations (stress and thermal) fell into this category. Rendering is much faster with more cores, though, as were the rest of the flow simulation tests (both normal airflow and conjugate heat transfer airflow models).

It is worth noting that we also tested assembly rotation performance across these CPUs, but didn't bother presenting graphs because they all performed within a fairly close spread - with no more than about 10% difference across various files and different view settings. That isn't surprising since rotation is mostly single-threaded and these CPUs all have max turbo speeds in the 4.3 to 5.0GHz range.


It is tricky to show overall relative performance when different aspects of an application behave differently, but if we break things down into features which seem to depend on clock speed versus those which scale well with core count, it can shed some light on things:

Summary of SOLIDWORKS 2019 CPU Performance Relative to Core i7 8700K

Based on that summary, and the full data shown in previous charts and graphs, it looks like the best CPU for general SOLIDWORKS usage (outside of rendering and some types of simulation) is the Core i9 9900K. No surprise there, I suppose, since it is the fastest mainstream processor Intel has made yet. The i7 9700K isn't far behind and is viable if you need to save some money.

On the other hand, if flow simulations or rendering are a bigger part of your workflow, then you may find it worthwhile to invest in a Core X series processor instead. Nothing below the Core i9 9900X will be useful, though, and even that processor is only a little bit faster than the 9900K for twice the price. The higher core count models are where performance really pulls ahead of the more mainstream chips.

If you are looking for a workstation for SOLIDWORKS, we have recommended systems for both general usage and rendering below.

Tags: Dassault, Systemes, Solidworks, 2019, CPU, Processor, Performance, Intel, Core, i5, i7, i9, Core X
Alan SW

Hi William thanks for this and the GPU article really nice to see these frequently to keep reccomendaitons up to date as hardware and software evolves. I just have some queries on the Simulation side. to start with i think i am right all of your "Benchmark simulations" at different mesh size are actually using SOLIDWORKS Flow Simulaiton (cfd) just thought i would mention since in SOLIDWORKS product names SOLIDWORKS Simulaiton is FEA and Flow Simulation is CFD might be worth just dropping the Flow word in there to make sense to SOLIDWORKS Users we typically observed the same high fluid cell count studies scale better with more cores.

Also what solver was used for the FEA study? typically with our own testing of the Direct Sparse solver we tended to see improvements of more cores for static simulation tending to level off around 8-12 cores. does depend on the simulation complexity, solver but usually seen more scaling than you did in your results. May be worth trying some other examples from the Simulaiton Training files if you haven't already to confirm the results https://www.solidworks.com/...

Posted on 2019-02-12 17:32:41

I will try to better note that info on the simulations we test in the future, but to answer your question: yes, the "benchmark" simulation uses SOLIDWORKS Flow Simulation... as do the "airflow" and "thermal" simulations. Only the "stress" simulation is FEA.

As for which solver the "stress" test uses, it is FFEPlus. I would be open to adding another test to future reviews, using a different solver, if someone could provide the files in a fully ready-to-run state.

Posted on 2019-02-12 19:46:47
Alan SW

Hi William cool just wanted to check I had read that correctly. Also makes sense re scaling with FEA as the FFEPlus solver does not scale well with more cores. I'll see what I can dig out from the testing we did in the past if you can let me know where to send it.
It may still be that you see less effect than when we last testedn as then the max you could get on latest generation CPUs was 4 cores i7700k atathr time you had to go to Skylake X (or Xeon equivalent in our case) for 6 and 8 cores+ Now that an 8 core relatively inexpensive cpu can be specfiied and still give the best single/ low threaded performance there may still not be as much benefit to Skylake X CPUs even when using the direct sparse solver.
Cheers Alan

Posted on 2019-02-12 20:50:25

Feel free to reach out to me at william {at} pugetsystems.com with any questions, suggestions, etc :)

Posted on 2019-02-12 21:01:23

Any way you guys could test the EPYC 7371 processor for solidworks and rendering and compare it to the intel counterparts? The EPYC 7371 has a much higher clock.

Posted on 2019-04-01 12:23:37

I believe we've asked AMD about EPYC processors, but they've not been able to provide sample CPUs or motherboards for us. With that said, though, I don't think there is any situation in which a single EPYC would be better than a Threadripper - the advantage with EPYC (as I understand it) is that it can be used in pairs, like the Xeon Scalable line. For example, the specific model you mentioned has the same core count as the Threadripper 1950X and 2950X processors, but lower clock speeds:

EPYC 7371: 3.1GHz base / 3.8GHz max boost (https://www.amd.com/en/prod...
TR 1950X: 3.4GHz base / 4.0GHz max boost (https://www.amd.com/en/prod...
TR 2950X: 3.5GHz base / 4.2GHz max boost (https://www.amd.com/en/prod...

However, as you may also have noticed, I didn't include any AMD chips in this particular article. If I have a chance I'd like to go back and do that at some point, but my priority was Intel processors for SOLIDWORKS because of how many parts of this application are single-threaded. Intel still holds a strong lead in that area, though I hear that AMD is hoping to close the gap in their next generation of CPUs. I guess we'll see :)

For rendering, though, AMD's Threadripper models do have very strong performance. We've got articles covering that by looking at V-Ray and Cinebench, if you're interested in more info.

Posted on 2019-04-01 16:14:29