Puget Systems print logo
https://www.pugetsystems.com
Read this article at https://www.pugetsystems.com/guides/2011
Article Thumbnail

SOLIDWORKS 2020 SP5 AMD Ryzen 5000 Series CPU Performance

Written on December 17, 2020 by William George
Share:

TL;DR: AMD Ryzen 5000 Series Performance in SOLIDWORKS

AMD's Ryzen 5000 Series processors are fantastic in SOLIDWORKS! These new chips continue to impress, and definitely take the performance lead in most aspects of this application. Intel's 10th Gen Core models aren't too far behind in general usage, and for a lot of folks doing strictly part and assembly design it would probably be hard to tell the difference. If you run a lot of simulations or do a bit of rendering, but not enough to justify the cost of a Threadripper, then the new Ryzen chips are definitely the way to go.

The one downside with these CPUs is that they have been in very short supply during the months following their launch in late 2020. We look forward to offering these AMD Ryzen processors in our SOLIDWORKS systems soon, once their availability improves.

Introduction

AMD's Ryzen 5000 Series processors have shown stellar performance across a wide range of applications so far, and now we have the opportunity to test them head-to-head against Intel's Core series in SOLIDWORKS 2020. For a long time, Intel's lead in single-threaded performance has kept them at the forefront of general engineering work, while AMD has been chipping away at high-thread-count workloads like rendering and simulations. With their latest Ryzen models, however, AMD has taken the lead in instructions-per-clock and rivals Intel in terms of clock speed as well, so let's see how that affects real-world processing in this popular engineering software.

Our in-house SOLIDWORKS benchmark suite covers performance in modeling, rendering, and simulations. In the past we have found this application to vary greatly in how it uses the CPU, with some functions being single-threaded while others are able to use all the cores effectively - and, of course, a spread of behavior in-between those extremes.

Looking for a Workstation?

Puget Systems offers a range of powerful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!

Test Hardware

Here are the detailed specs of the test platforms we used:

Intel Core Test Platform
CPU Intel Core i9 10900K
Intel Core i7 10700K
CPU Cooler Noctua NH-U12S
Motherboard Gigabyte Z490 Vision D
Shared Hardware
RAM 4x DDR4-3200 16GB (64GB total)
Video Card NVIDIA Quadro RTX 6000 24GB
Hard Drive Samsung 960 Pro 1TB
Software Windows 10 Pro 64-bit
SOLIDWORKS 2020 SP5

Benchmark Details

Our SOLIDWORKS benchmark suite was originally developed by one of my colleagues 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, which I have updated and added to over the years. I have been aided in that process by the help of many readers who have suggested changes, provided additional files for testing, and more. I have done my best to cite their contributions in past articles as they were integrated into our benchmark. The results are broken up into individual graphs below, grouped by general usage, simulation, and rendering - and followed by our analysis.

Results & Analysis

Here are galleries of the results from our SOLIDWORKS testing. AMD Ryzen processors are shown in green with Intel's Core CPUs in blue:

SOLIDWORKS start up, file open / save, rebuild, and motion study performance

In this first set of data, we can see that all five of the tested processors are fairly close when it comes to starting up SOLIDWORKS itself as well as saving files; both show a spread of less than a second. Opening files is a little more stratified but without any clear winner. However, when it comes to file rebuilds and motion studies the new AMD Ryzen chips have a clear and substantial advantage! It is worth noting that the file we test rebuilding on is artificially complex, in order to make the process take long enough to see substantial differences between CPUs, so most SOLIDWORKS users would be looking at much shorter rebuild times in their day-to-day work.

In the past, we have sometimes looked at the CPU's impact on part and assembly manipulation, but that is primarily limited by the video card rather than the processor. For performance data on that, check out our SOLIDWORKS 2020 SP1 GPU performance article.

Various SOLIDWORKS simulation tests

Simulations in SOLIDWORKS come in many varieties, and in past tests with a wider range of processors have shown equally varied performance across different CPUs. In this comparison, however, AMD's new Ryzen 5000 series is a pretty clear winner. All three models we tested beat both of the Intel chips, except in the largest simulation where the 8-core Ryzen 7 5800X fell slightly behind the 10-core Core i9 10900K. That fits with past results as well, where we have seen higher core counts become more important as the complexity of simulations increased. Looking back at those previous articles, though, it looks like these new Ryzen processors should beat, or at least match, any other CPUs currently available when it comes to running simulations within SOLIDWORKS! The only limitation these models have in this regard is the limit of 128GB of memory - which is far more than any of our benchmarks require, but could potentially be a constraint with real-world workloads.

SOLIDWORKS PhotoView 360 rendering

CPU-based rendering tends to scale very well across multiple threads, so PhotoView 360 is a place where high core count processors really shine. Naturally, then, the 12- and 16-core Ryzen chips beat out Intel's Core models which don't support as many simultaneous threads. Interestingly, though, the 8-core Ryzen 7 5800X also beats out Intel's 10-core i9 10900K... showing that AMD's new CPUs are quite fast per-core to make up for that difference.

While AMD's Ryzen 5000 Series is very good here, the higher core count Threadripper models we've tested in the past are even better for dedicated rendering systems. If your primary pain point is waiting on PhotoView 360, or another CPU-based rendering engine, it might be good to look at some of our rendering workstations rather than our SOLIDWORKS-optimized systems.

Are AMD's Ryzen 5000 Series processors good for SOLIDWORKS?

Yes, AMD's Ryzen 5000 Series processors are fantastic in this application! These new chips continue to impress, and definitely take the performance lead in most aspects of SOLIDWORKS. Intel's 10th Gen Core models aren't too far behind in general usage, and for a lot of folks doing strictly part and assembly design, it would probably be hard to tell the difference. If you run a lot of simulations, though, or do a bit of rendering (but not enough to justify the cost of a Threadripper) then the new Ryzen chips are definitely the way to go.

The one downside with these CPUs is that they have been in very short supply during the months following their launch in November of 2020. We look forward to offering these AMD Ryzen processors in our SOLIDWORKS systems soon, once their availability improves.

Looking for a SOLIDWORKS Workstation?

Puget Systems offers a range of poweful and reliable systems that are tailor-made for your unique workflow.

Configure a System!

Labs Consultation Service

Our Labs team is available to provide in-depth hardware recommendations based on your workflow.

Find Out More!
Tags: Dassault, Systemes, CPU, Processor, Performance, Intel, Core, i7, i9, Solidworks, AMD, Ryzen, Rendering, Intel 10th Gen, AMD Ryzen 5000 Series, 5950X, 5900X, 5800X, i9 10900K, i7 10700K
Nuno

I guess 3Ds Max is coming next...? :-)

Posted on 2020-12-17 22:25:44

That's not one of the applications I test, but rather in Kelly's area of focus. I don't think any more major CPU testing articles are coming this year, though - between the holidays, planning ahead for next year, and all hands being on deck to try and push through all of the orders we're rather hard pressed for time right now.

Posted on 2020-12-17 23:41:03
Nuno

Sorry William, didn't realize that. I guess Kelly is on it. Thank you for the detailed article and Happy Holydays!

Posted on 2020-12-18 11:32:43

You are very welcome! Happy Holidays to you as well :)

Posted on 2020-12-18 17:41:38
dan297

You should have paired the Ryzen CPUs with a Gen4 SSD in order to make most out of them.
I build a new system with R9 5950X lately and program start and file opening/saving is up to 25% faster than it was on my old i9 9900KS platform.
Your test bed config does not adequately represent this Zen3 benefit...

Posted on 2020-12-19 11:51:05

Hmm, that is a good point. We try to keep everything else as similar as possible when doing head-to-head CPU comparisons, so I still think that to see the CPU's performance in isolation this was worthwhile - but as fast as PCIe 3 based M.2 drives are, there are going to be some things where the additional speed of PCIe 4 could make a difference. This may well be one of them :) Perhaps I will be able to look at that specifically in another comparison next year.

Posted on 2020-12-22 01:46:29
dan297

True, it surely makes things more complex once you compare CPUs that are not based on the same platform and you have more variables like different chipsets, different drivers, etc...
I was stunned especially on program start. It went from 30+ seconds to 6 seconds. My toolbox is the culprit, but I keep it on autostart, just out of habit.
On the AMD system with a Gigabyte Aorus it starts in less than a second, on two different Intel systems (i7 9700K and I9 9900KS) it takes 21 seconds, both with Samsung 970 EVO Plus. Although this difference is so significant - I am not sure if the Gen4 disk can take all the credit...
Anyway, looking forward to your test. Love the stuff you do here, it is always great input. Keep going!

Posted on 2020-12-22 07:12:02
Izzy

While I agree with the importance of CPU in isolation situations, it's also good to compare these with newer technologies.

AMD has support for higher RAM speeds for example, and this has been shown to make a significant difference in many applications. Ryzen's support for 3200 mhz could make a difference compared to Intel's support of 2966 mhz.

The same can be said with PCIe 4.0 and PCIe 3.0

In the plausible scenario that one company would be first to release a CPU with DDR5 RAM for instance, we would have to give the company credit for their implementation of this technology.

When it came to Premiere benchmarks for example, you factored Intel's QuickSync technology. AMD's technological leads should be factored in if it could lead to a better performing system.

Posted on 2021-01-04 10:56:03
Alex Taguchi

I think having both scenarios tested is beneficial. I prefer to keep all variables the same to highlight the true CPU performance differences, and then separate benchmark that highlights the platform/chipset benefits in tests where it matters like file open. etc.

The biggest surprise here is that the jump from Ryzen 3000 to 5000 seems significant since the 9th gen i9s still topped most charts last time around and now they lag in almost every test.

Second biggest surprise is the difference in rebuild times. Rebuild times have almost always been limited by clock speed and the AMD outperforms the i9 by 40-50% which doesn't make a ton of sense since their turbospeeds are lower than the 10900k. Did you measure temperature or average clock speed during those rebuild tests? My first reaction is that the i9s were not running at their maximum turbo speeds while the AMDs were.

Fantastic benchmarks as usual.

Posted on 2021-01-15 22:57:48

A large performance gap like that could also mean it has something to due with instruction set optimizations. We see something similar in Lightroom where tasks like exporting are way faster with Ryzen compared to Intel 10th Gen: https://www.pugetsystems.co... . Apparently, that has something to due with Lightroom using AVX-2 rather than the newer AVX-512 for those tasks, and apparently AMD CPUs are better for AVX-2. Or something like that - I can't recall the exact details anymore and that is beyond the scope of what I generally look at.

Posted on 2021-01-15 23:39:04
Alex Taguchi

Thanks for the reply. That crossed my mind as well, but there have been very little technical specifications regarding instruction sets implemented across release notes. The last time anything was mentioned like this was in the Athlon XP days with SSE2. Curious results.

Posted on 2021-01-26 00:14:09
Andres Eljadue Tarud

I would love to see the test for Ryzen 5 5000 series against Intel i5 10th gen, an option for a lower budget, this option would works for rebuild medium assemblies with external measures?

Posted on 2020-12-31 01:57:09
Andres Eljadue Tarud

Hi William, what to do think would be better for a low budget pc, an Intel Core i7-9700K or a Ryzen 5 5600?

Posted on 2020-12-31 02:25:19

I own a 9700K in my personal home system, and it is a good CPU... but as much as it pains me to say, I'm pretty sure the Ryzen 5 5600X would out-perform it in most areas. It is effectively 2 generations newer, has higher per-core performance, and official support for faster memory speeds. The only place it isn't superior is the number of cores (6 vs 8) but since it does support SMT and the 9700K does *not* have HT that probably ends up being a wash. So today, on a budget? I'd go for the 5600X.

Posted on 2020-12-31 02:41:08
Andres Eljadue Tarud

Thanks!, Ive been reading a lot of your Articles, Amazing job

Posted on 2020-12-31 14:13:12
Andres Eljadue Tarud

I´ve follow your call and I've been looking for the 5600X for two weeks, but it is too much harder than it used to be, looks like the market doesnt have stock, so, there is just a few selling the CPU at not affordable prices. we have to wait a little longer

Posted on 2021-02-02 20:02:46
Evan

Just a heads up, sometime in Solidworks 2021(it might be from service pack 0, or only in a later service pack). FEA simulations will be limited to 8 cores, so that will play a role in future performance.

Posted on 2021-01-12 09:53:04

Oh, interesting - thank you for the heads up! That would definitely change things a bit :/ Do you happen to have a link to info about that? I'd love to get an understanding of why they would make a change like that when the clear trend of CPUs over the last decade has been to add more and more cores.

Posted on 2021-01-12 19:35:35
Evan

Hi William, for some reason I thought I had replied to this. I was partially wrong with this, it seems to only be a limitation on Solidworks Simulation Standard and lower, probably to create a bit more differentiation between the upper and lower tiers(and provide incentive to buy the higher tiers for users that only do static simulations). Please see bottom of page here. A similar limitation has been added to the mesher aswell, see here for that.

Posted on 2021-01-15 07:11:06

Ah, gotcha - that makes sense, and is a practice I've seen in other simulation frameworks (Ansys, for example, has licensing tiers supporting different numbers of CPU threads). It looks like SW 2021 is supposed to have a decent speed-up in mesh generation and solver performance as well, which is always nice!

Posted on 2021-01-15 18:23:03
Mr. Trickster

Can you testing actual AMD's APU like 4700(4750)g/4600(4650)g without discrete graphics card in SolidWorks?
Is the integrated graphics enough for 3d modeling, large assemblies and design documentation with many views, detail elements, etc.

Posted on 2021-02-22 09:54:07

Unfortunately we don't have any of those APUs, and since we don't carry them for our workstations I don't think it is likely that we'd purchase any or that AMD would send us samples.

Posted on 2021-02-22 17:45:36
Alex Taguchi

I'm thinking of creating a series of videos about swx performance on a variety of systems. Are you guys OK with me referencing this article?

Posted on 2021-04-22 17:40:46

Yeah, absolutely! Please just mention us (during the video and/or in the description) and if you use any of the charts from our articles please keep the logo intact :)

Posted on 2021-04-22 17:54:55
Alex Taguchi

Absolutely, love your work.

Posted on 2021-04-22 17:59:55

Thank you! :)

Posted on 2021-04-22 18:47:53
Alex Taguchi

Would you or one of your team be willing to participate in a live webinar for our customers? We regularly host webinars about SOLIDWORKS and hardware/software optimizations and it would be awesome to have a guest presenter from your team since we reference your benchmark data.

Posted on 2021-07-22 18:59:39
D Ford

I'm interested in seeing a comparison regarding importing neutral file formats - I frequently need to import large STEP files provided by customers which can leave me sitting for long periods of time waiting for solidworks to generate a model. I'm not sure if the import process is single-thread limited or well multithreaded, and would probably be the deciding factor for me in a new workstaiton build between a mid-range CPU with a 5800x and a higher end with a 5950x which conveniently matches our existing MasterCAM workstations.

Is this something puget has tested, or would be willing to test in future SolidWorks CPU benchmarks?

Posted on 2021-10-12 14:17:32

That is an interesting idea, certainly - but at the moment we don't have anyone working specifically on SW here in our lab. I was doing that, but have moved to more of a product development role. If / when we have another engineering-focused researcher, I will try to remember to suggest this addition :)

Posted on 2021-10-13 15:35:03
Michael Grimm

I have appreciated your SW hardware articles. I hope you find someone qualified to continue the analyses.

Posted on 2021-10-19 15:51:59