Virtual WRSC and simulator

[Yu Cao 曹宇]

Hi sailing robotics fellows,

It's have been a while since our last meet at WRSC China 2019. Lots of things have changed and we are facing a challenge to this year's WRSC.

I have been talking with many teams and thinking about the option for this year's WRSC. I know many teams may not able to resume even the competition is postponed until late autumn.

Many colleagues are proposing a virtual challenge that hosts online to avoid international travel restrictions. 

Rhys and I are looking at the Gazebo based simulator for the virtual challenge and we would like to present some initial result from this project. 

This simulator is based on the popular robotics simulator Gazebo and containerised using docker. Physical simulation including the wind and wave are included in the model.

In addition, 

- Simulated sensor data such as anemometer, IMU, GPS are provided are ROS topic

- Sails and rudder are separately controllable with custom software

- Wind and wave parameters can be changed in the simulator

- Virtual camera plugin

- Near real-time simulation 

We would like to invite colleagues to try this simulator and give us feedback. If there are enough teams interested in a virtual WRSC we can develop challenges based on this simulator.

 

Link to the simulator:

https://github.com/srmainwaring/sail_sim_docker

Best regards,

Yu

[rhys.ma...@me.com]

Some extra detail as follow up to Yu's introduction:

The github repo https://github.com/srmainwaring/sail_sim_docker contains a dockerfile to build the image and a docker-compose file to run it. Currently it is not possible for you to build the image using the dockerfile because it depends on modifications to https://github.com/srmainwaring/asv_wave_sim and some additional repos that I have not pushed into the public domain.

If you'd like access to the docker image drop me a request with your docker ID and I'll add you as a collaborator. To pull the image use:

docker pull rhysmainwaring/sail-sim

To run the image:

docker-compose -f docker-compose-nvidia.yaml up (use docker-compose-vmware.yaml if you are running on VMware or machine with an AMD card) 

If you're interested to delve into the physics underpinning the simulation:

 

- A description and references to the approach used for the buoyancy calculations are in the wave sim repo. It is mostly based on Jacques Kerner's two part blog describing boat physics for games: Water interaction model for boats in video games and Water interaction model for boats in video games: Part 2.

- The waves in the docker image are generated using the approach outlined by Jerry Tessendorf's in this report: https://people.cs.clemson.edu/~jtessen/reports/papers_files/coursenotes2004.pdf. 

- The foils and sails use a variant of the lift-drag plugin described here: http://gazebosim.org/tutorials?tut=aerodynamics&cat=physics 

- Apart from that the boat has the usual rigid-body physics that an object in Gazebo would have, so it will bump off objects (other boats!), get stuck in shallow water if it runs aground, that type of thing.

If you have any questions or issues please post an issue in the sail_sim_docker repo.

Kind regards,

Rhys

[Colin Sauze]

That simulator looks very promising.

I would like to try and run some kind of virtual WRSC this year but i'm wondering when the best time to hold it would be. We usually have the WRSC in early September, but that's at least partially a weather related decision and to be at a time when there are fewer tourists but not much university work. Given that weather and competing with tourists aren't a problem online we're less constrained about timing and perhaps a bit later in the year would be better to give everyone some time to work on things.

Colin.

--
You received this message because you are subscribed to the Google Groups "World Robotic Sailing Championships Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrsc-discuss...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/e4e040a7-411c-410d-8a7c-9b48cede83a6o%40googlegroups.com.

-- 
Dr. Colin Sauze
Research Software Engineer
Supercomputing Wales Project
Room 2.16, Physical Sciences Building
Aberystwyth University,
Penglais, Aberystwyth
Ceredigion, UK, SY23 3DB

Tel: +44 (0)1970 622774
Webpage: http://users.aber.ac.uk/cos

----------------------------------------------------------------------------------------------------------------------

Pantycelyn yn agor mis Medi 2020. https://www.aber.ac.uk/cy/accommodation/accommodation-options/catered/pantycelyn/

Pantycelyn opening September 2020. https://www.aber.ac.uk/en/accommodation/accommodation-options/catered/pantycelyn/

----------------------------------------------------------------------------------------------------------------------

Prifysgol Aberystwyth www.aber.ac.uk/cy/

Prifysgol y Flwyddyn yng Nghymru - The Times & The Sunday Times 2020.

Aberystwyth University www.aber.ac.uk/en/

Welsh University of the Year - The Times & The Sunday Times 2020.

----------------------------------------------------------------------------------------------------------------------

[Yu Cao 曹宇]

Hosting the virtual challenge later this year seems a better option. This will allow the team to have some time to prepare for the new challenge. 

But if we decided to have a virtual challenge, it is better to announce it as early as possible. Most students won't have a block of time to work on this from next semester. 

We also need to form a technical group to decide the challenge and release instruction and documents. 

Best regards,

Yu

[rhys.ma...@me.com]

A quick update on the simulator:

I've released the software for the simulator under GPLv3 and updated the docker image to build from the public repos. Bryn Heveldt, the designer of the Racing Sparrow 750 on which the robot model is based, kindly gave his permission for me to publish the derivative work. As a result all the code, models etc. are available to review. Here are the links:

- https://github.com/srmainwaring/asv_wave_sim/tree/feature/fft_waves

- https://github.com/srmainwaring/asv_sim/tree/feature/wrsc-devel

- https://github.com/srmainwaring/rs750/tree/feature/wrsc-devel

- https://github.com/srmainwaring/sail_sim_docker

Rhys.

[Colin Sauze]

Any thoughts on when a good time to run it would be? I was thinking perhaps early November as this gives some time for student teams to work on things after they go back but doesn't go too late to start clashing with end of term deadlines and exams.

As for the challenges I think we could run most of the traditional events like a triangular course race, station keeping and area scan. But i'm not sure we'll see much difference between the different teams for these as what makes these challenging is more the hardware and boat design than it is the software. Although if different sailing conditions with variable winds, waves and tides can be simulated then these events might get more interesting.

Depending on how well the simulator works we could look at trying to do collision avoidance using a simulated camera and/or transponder. If the simulation can handle multiple boats then we could also look at doing a co-operative event where for example two boats try to maintain a formation or do the area scan task collaboratively.

Colin.

On 23/06/2020 19:54, Yu Cao 曹宇 wrote:

Hosting the virtual challenge later this year seems a better option. This will allow the team to have some time to prepare for the new challenge. 

But if we decided to have a virtual challenge, it is better to announce it as early as possible. Most students won't have a block of time to work on this from next semester. 

We also need to form a technical group to decide the challenge and release instruction and documents. 

Best regards,

Yu

On Tuesday, 23 June 2020 at 18:27:52 UTC+1 Colin Sauze wrote:

To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/fd3c1ce0-81b2-4110-a4e8-a5855c759156n%40googlegroups.com.

[Yu Cao 曹宇]

Early or mid-November sounds like good timing to me. For teams with existing software, it would be too hard for them to finish the classical challenges.

And for the new team, the virtual challenge allows them to get started with robotics sailing without purchasing hardware. 

What I am trying to do is to connect Southampton sailing robot code with this simulator. This could be a chance for us to evaluate the performance of the simulator and decide the format of challenge. The result of the work will be hosted at WRSC GitHub account and provide to teams as a demo. 

[rhys.ma...@me.com]

The proposed time sounds good to me. There is some work to do to get the simulation code production ready depending upon the features needed for the challenge. These include:

- parameters to control the wave environment

- reflections and refraction if visual sensing fidelity is important

- additional sensors

- controller for sail trim that sets travel limits rather than a fixed position  

In addition we'd need to set up a number of scenes / courses. I am hoping that the assets developed for the osrf/vrx challenge can be incorporated without too much modification. The best approach here would be to decide on a location and course and put together a prototype. If you could provide me we a venue and spec. for the course I can look into this. (e.g. location:= Aberystwyth harbour, course:=[list of buoy shapes, colours, GPS positions]). Alternatively I can just take the osrf/vrx sandisland course and adapt that if you're not particular about the venue.  

Yu, how are you getting on with adapting the Soton code to work with the simulator? Let me know if you'd like any help with that.

Rhys

[Rhys Mainwaring]

Hi everyone,

I'm not sure if there is still interest to run an event this year, but I've been working on improvements to the simulator in any case. I've put a proof of concept docker image out with ArduPilot integration - so you can run the simulation using SITL and MavProxy. The project is on a branch here: https://github.com/srmainwaring/sail_sim_docker/tree/feature/ardupilot_sailboat_poc

Kind regards,

Rhys

[David Meredith]

Hi Rhys,

Thanks for all your work on this! I'm planning to propose a student project for the coming semester where we'll have a group of masters students working with your simulation until Christmas. So there will be plenty of interest from here in Aalborg! I'm going to try to get to grips with the simulation today.

Best regards,

David Meredith

Aalborg University, Denmark

[Rhys Mainwaring]

Hi David,

That's exciting! I've some work to do on the FFT wave generation to move it from a prototype to a better state for general use which I'm working through at the moment:
1. provide an interface for setting parameters base on standard wave spectra

2. fix some issues with the scaling of waves with the FFT resolution

3. fix some issues with the sampling from ocean spectra (linked to 2)

If you have any issues or feature requests please add them to the github repo.

I've looked into running the simulation from Windows 10 with WSL for an Ardupilot developer but found that WSL does not support graphics hardware acceleration (WSL2 may do later this year). I've found it runs fine on macOS and an Ubuntu VM. Yu had it running on Linux natively with an NVIDA card. Running natively, from source, you can compile in the OpenCL acceleration for the FFT which also helps (depending on your machine).

Kind regards,

Rhys

[Colin Sauze]

Hi Rhys,

I'm still interested in trying to do something this year but haven't had much time recently as i've been trying to put a grant proposal together.

I've just run your SITL container and it looks pretty good. The only issue was that it only ran at about 0.3 times speed, but this was on a very slow virtual machine. I'll try again on a faster system and see if I get better performance.

Using ardupilot as a reference implementation seems like a good idea, I hadn't realised there was now a functional ardupilot for sailing boats available.

If there's already a good reference autopilot, this brings up the obvious question of what we are challenging the participants to produce for a virtual WRSC. The obvious things that I can think of is doing collaborative tasks between different boats or collision avoidance. This brings up the question, how difficult would it be to simultaneously simulate multiple boats? 

The simple way to do this could be to run a separate simulator instance for each participant and have an external API (possibly the same one we've used for tracking real boats at the WRSC since 2015) that covers detecting other boats. But it would be really nice if we could do it all in one simulator instance and then we could access things such as simulated RADAR/LIDAR or cameras.

I've started a document on the technical requirements and some general ideas for a simulated WRSC, please add any thoughts to:

https://github.com/WRSC/simulation-challenge/blob/master/requirements.md

Colin.

To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/8e323e5c-d55b-46ba-813a-a333e9659d41n%40googlegroups.com.

[Rhys Mainwaring]

Hi Colin,
 

I've posted some notes about the SITL performance on the Ardupilot sailboat thread https://discuss.ardupilot.org/t/sailboat-support/32060/626. Depending on your environment and VM, the CPU load can be large as it has to simulate the physics (twice for the waves because of Gazebo's client / server model) and translate the OpenGL calls if there is no native support for OpenGL 2 and greater on the VM. On a VM with 4 CPU, 8GB RAM and hardware acceleration it should run at 30 FPS with a RTF of about 0.9 - 0.95.

I think SITL can support multiple drones, but I've not experimented. The SITL docker image was my first foray into ArduPilot so I'm still familiarising myself with its capabilities.

Gazebo and ROS will support multiple robots in the same simulation, provided that all the namespace logic is set up correctly. I have not implemented robot namespace in the proof of concept, but have examples of how to do this. There will of course be some performance impact. There are ways of mitigating this by simplifying the mesh used for the buoyancy dynamics and further limiting which objects are buoyant (at the moment its everything - including spars etc.) 

 

If supporting multiple boats initially proves to be too computationally expensive, you could always simulate collision avoidance against static obstacles, or dynamic obstacles that are mocked up with no or very simplified buoyancy (e.g. to simulate a shipping lane or ferry crossing, or navigate an uncharted area). 

In ROS you have access to simulated cameras and rangefinders. A number of teams have published packages for various sensors so it's a matter of selecting which ones you want and adding them to the vessel . Cameras are quite expensive to simulate as each one has to render the scene, so we may need to think about how that works when running multiple boats (you'd only want to render the camera feed for the boat your controlling, but still see the other boats in the main scene).  

On the physics side there are a number of aspects which require tuning, and I have not spent a lot of time fitting the simulated vessel motion to experimental data (I'm in the process of putting a boat together at the moment which may eventually help fill that gap - but if you have real boat data or a VPP program that can generate it that would be very helpful). There will be limits to what can be achieved given the current approximations used.

The main variables for the hull are drag coefficients (effectively for form and viscous drag). We could tabulate expected constant velocities for a given hull form as a function of thrust force (or wind speed for a given sail plan), and try to fit the drag coefficients. I found that the hydrodynamics approximations did not give enough lift / drag for the fin and rudder, so there is an additional lift / drag plugin with its own coefficients. These can be adjusted to control side force, heel, leeway and turning radius. In the simulation you can see the rudder stall if turned too aggressively, and the fin stalls if the boat is pushed too close to the wind or at low velocities out of a tack.

I have some work in progress to improve the FFT waves. The current version has some scaling issues as the maximum wavelength and resolution are adjusted. There is also no spreading function to control wave direction. This may or may not be relevant for a first version. There is also work to support approx. reflection / refraction which can be relevant for camera work - again it's probably not a priority for an initial version which I'd recommend keeps to the basics but does them well. 

Rhys

[Yu Cao 曹宇]

Two Ardupilot based sailing robots (both from Huazhong University of Science and Technology) have participated in WRSC 2019 already : )  

On my computer (i9-9900K + 2080Ti), the simulation runs at 60 FPS with RTF about 0.97 ~ 0.98. I don't know if there are any losses due to the virtualization, but the hardware requirement for the virtual competition will at least similar to Rhys'. I have been contacting with some robotics research institutes to see if they are interested in cost-host this event and provide some free cloud service.

Wish we can find an hour or so for an online meeting to plan out the virtual simulation!

[Colin Sauze [cos]]

I've added a web based VNC server to the container (https://github.com/colinsauze/sail_sim_docker) so its easy to run it on any cloud system. I tried it on a few different configurations but had to allocate 32 CPU cores on AWS to get up to between 0.85 and 0.88 RTF. A 32 core cloud system costs about $1 per hour to run, so this wouldn't be too much to run a competition for a few days but might start to get expensive for teams to do development on.

I'm wondering if there's anything we can do to reduce demand on the CPU. I tried switching to wireframe view and making the gazebo window smaller but this didn't make much difference, perhaps 0.05 RTF at most. When I get a chance I'll see if I can work out which parts of the process are taking the most CPU time.

Colin.

________________________________________
From: wrsc-d...@googlegroups.com [wrsc-d...@googlegroups.com] on behalf of Yu Cao 曹宇 [tsa...@gmail.com]
Sent: 09 September 2020 11:52
To: World Robotic Sailing Championships Discussions
Subject: Re: [sailbot] Re: Virtual WRSC and simulator

[RHYBUDD! E-BOST ALLANOL / CAUTION! EXTERNAL E-MAIL]

Two Ardupilot based sailing robots (both from Huazhong University of Science and Technology) have participated in WRSC 2019 already : )

On my computer (i9-9900K + 2080Ti), the simulation runs at 60 FPS with RTF about 0.97 ~ 0.98. I don't know if there are any losses due to the virtualization, but the hardware requirement for the virtual competition will at least similar to Rhys'. I have been contacting with some robotics research institutes to see if they are interested in cost-host this event and provide some free cloud service.

Wish we can find an hour or so for an online meeting to plan out the virtual simulation!

On Thursday, 3 September 2020 at 00:32:06 UTC+1 Rhys Mainwaring wrote:

Hi Colin,

I've posted some notes about the SITL performance on the Ardupilot sailboat thread https://discuss.ardupilot.org/t/sailboat-support/32060/626<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdiscuss.ardupilot.org%2Ft%2Fsailboat-support%2F32060%2F626&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364380436&sdata=OrlosX7RsGR4iduhexZiAd59cfEzL1StB4wMqfyPVpc%3D&reserved=0>. Depending on your environment and VM, the CPU load can be large as it has to simulate the physics (twice for the waves because of Gazebo's client / server model) and translate the OpenGL calls if there is no native support for OpenGL 2 and greater on the VM. On a VM with 4 CPU, 8GB RAM and hardware acceleration it should run at 30 FPS with a RTF of about 0.9 - 0.95.

I think SITL can support multiple drones, but I've not experimented. The SITL docker image was my first foray into ArduPilot so I'm still familiarising myself with its capabilities.

Gazebo and ROS will support multiple robots in the same simulation, provided that all the namespace logic is set up correctly. I have not implemented robot namespace in the proof of concept, but have examples of how to do this. There will of course be some performance impact. There are ways of mitigating this by simplifying the mesh used for the buoyancy dynamics and further limiting which objects are buoyant (at the moment its everything - including spars etc.)

If supporting multiple boats initially proves to be too computationally expensive, you could always simulate collision avoidance against static obstacles, or dynamic obstacles that are mocked up with no or very simplified buoyancy (e.g. to simulate a shipping lane or ferry crossing, or navigate an uncharted area).

In ROS you have access to simulated cameras and rangefinders. A number of teams have published packages for various sensors so it's a matter of selecting which ones you want and adding them to the vessel . Cameras are quite expensive to simulate as each one has to render the scene, so we may need to think about how that works when running multiple boats (you'd only want to render the camera feed for the boat your controlling, but still see the other boats in the main scene).

On the physics side there are a number of aspects which require tuning, and I have not spent a lot of time fitting the simulated vessel motion to experimental data (I'm in the process of putting a boat together at the moment which may eventually help fill that gap - but if you have real boat data or a VPP program that can generate it that would be very helpful). There will be limits to what can be achieved given the current approximations used.

The main variables for the hull are drag coefficients (effectively for form and viscous drag). We could tabulate expected constant velocities for a given hull form as a function of thrust force (or wind speed for a given sail plan), and try to fit the drag coefficients. I found that the hydrodynamics approximations did not give enough lift / drag for the fin and rudder, so there is an additional lift / drag plugin with its own coefficients. These can be adjusted to control side force, heel, leeway and turning radius. In the simulation you can see the rudder stall if turned too aggressively, and the fin stalls if the boat is pushed too close to the wind or at low velocities out of a tack.

I have some work in progress to improve the FFT waves. The current version has some scaling issues as the maximum wavelength and resolution are adjusted. There is also no spreading function to control wave direction. This may or may not be relevant for a first version. There is also work to support approx. reflection / refraction which can be relevant for camera work - again it's probably not a priority for an initial version which I'd recommend keeps to the basics but does them well.

Rhys
On Wednesday, September 2, 2020 at 4:39:17 PM UTC+1 Colin Sauze wrote:
Hi Rhys,

I'm still interested in trying to do something this year but haven't had much time recently as i've been trying to put a grant proposal together.

I've just run your SITL container and it looks pretty good. The only issue was that it only ran at about 0.3 times speed, but this was on a very slow virtual machine. I'll try again on a faster system and see if I get better performance.

Using ardupilot as a reference implementation seems like a good idea, I hadn't realised there was now a functional ardupilot for sailing boats available.

If there's already a good reference autopilot, this brings up the obvious question of what we are challenging the participants to produce for a virtual WRSC. The obvious things that I can think of is doing collaborative tasks between different boats or collision avoidance. This brings up the question, how difficult would it be to simultaneously simulate multiple boats?

The simple way to do this could be to run a separate simulator instance for each participant and have an external API (possibly the same one we've used for tracking real boats at the WRSC since 2015) that covers detecting other boats. But it would be really nice if we could do it all in one simulator instance and then we could access things such as simulated RADAR/LIDAR or cameras.

I've started a document on the technical requirements and some general ideas for a simulated WRSC, please add any thoughts to:

https://github.com/WRSC/simulation-challenge/blob/master/requirements.md<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FWRSC%2Fsimulation-challenge%2Fblob%2Fmaster%2Frequirements.md&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364390431&sdata=r45LsfKIy3B0Wm8y9pWydzurWuyC9dWu34ELM2EypdI%3D&reserved=0>

Colin.


On 17/08/2020 16:23, 'Rhys Mainwaring' via World Robotic Sailing Championships Discussions wrote:
Hi everyone,

I'm not sure if there is still interest to run an event this year, but I've been working on improvements to the simulator in any case. I've put a proof of concept docker image out with ArduPilot integration - so you can run the simulation using SITL and MavProxy. The project is on a branch here: https://github.com/srmainwaring/sail_sim_docker/tree/feature/ardupilot_sailboat_poc<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fsrmainwaring%2Fsail_sim_docker%2Ftree%2Ffeature%2Fardupilot_sailboat_poc&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364390431&sdata=dxyO8fUzRXBs01MV%2B0yRmeCbWZOzpyCfHFwsGfEOYFo%3D&reserved=0>

[thumbnail_Screenshot from 2020-06-21 19-59-27.png]In addition,

- Simulated sensor data such as anemometer, IMU, GPS are provided are ROS topic
- Sails and rudder are separately controllable with custom software
- Wind and wave parameters can be changed in the simulator
- Virtual camera plugin
- Near real-time simulation










We would like to invite colleagues to try this simulator and give us feedback. If there are enough teams interested in a virtual WRSC we can develop challenges based on this simulator.


Link to the simulator:

https://github.com/srmainwaring/sail_sim_docker<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fsrmainwaring%2Fsail_sim_docker&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364400424&sdata=2EQ%2B3mWvx27KHxV6jwS2r05tKNbo1XZQrhccHm%2FVOwQ%3D&reserved=0>

Best regards,
Yu




--
You received this message because you are subscribed to the Google Groups "World Robotic Sailing Championships Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrsc-discuss...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/e4e040a7-411c-410d-8a7c-9b48cede83a6o%40googlegroups.com<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fwrsc-discuss%2Fe4e040a7-411c-410d-8a7c-9b48cede83a6o%2540googlegroups.com%3Futm_medium%3Demail%26utm_source%3Dfooter&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364400424&sdata=MibnppzM56GGz719bPdMLyZmpomyJsFbuvNLzkg1Bv8%3D&reserved=0>.

--
Dr. Colin Sauze
Research Software Engineer
Supercomputing Wales Project
Room 2.16, Physical Sciences Building
Aberystwyth University,
Penglais, Aberystwyth
Ceredigion, UK, SY23 3DB

Tel: +44 (0)1970 622774
Webpage: http://users.aber.ac.uk/cos


----------------------------------------------------------------------------------------------------------------------

Pantycelyn yn agor mis Medi 2020. https://www.aber.ac.uk/cy/accommodation/accommodation-options/catered/pantycelyn/

Pantycelyn opening September 2020. https://www.aber.ac.uk/en/accommodation/accommodation-options/catered/pantycelyn/

----------------------------------------------------------------------------------------------------------------------

Prifysgol Aberystwyth www.aber.ac.uk/cy/<http://www.aber.ac.uk/cy/>

Prifysgol y Flwyddyn yng Nghymru - The Times & The Sunday Times 2020.

Aberystwyth University www.aber.ac.uk/en/<http://www.aber.ac.uk/en/>

Welsh University of the Year - The Times & The Sunday Times 2020.

----------------------------------------------------------------------------------------------------------------------

--
You received this message because you are subscribed to the Google Groups "World Robotic Sailing Championships Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrsc-discuss...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/fd3c1ce0-81b2-4110-a4e8-a5855c759156n%40googlegroups.com<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fwrsc-discuss%2Ffd3c1ce0-81b2-4110-a4e8-a5855c759156n%2540googlegroups.com%3Futm_medium%3Demail%26utm_source%3Dfooter&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C1%7C637352455364410422&sdata=loX2%2B4qN4fuR2lCDWM4jw8hLfnO%2B%2BAkwPOIAEf0L138%3D&reserved=0>.

--
Dr. Colin Sauze
Research Software Engineer
Supercomputing Wales Project
Room 2.16, Physical Sciences Building
Aberystwyth University,
Penglais, Aberystwyth
Ceredigion, UK, SY23 3DB

Tel: +44 (0)1970 622774
Webpage: http://users.aber.ac.uk/cos


----------------------------------------------------------------------------------------------------------------------

Pantycelyn yn agor mis Medi 2020. https://www.aber.ac.uk/cy/accommodation/accommodation-options/catered/pantycelyn/

Pantycelyn opening September 2020. https://www.aber.ac.uk/en/accommodation/accommodation-options/catered/pantycelyn/

----------------------------------------------------------------------------------------------------------------------

Prifysgol Aberystwyth www.aber.ac.uk/cy/<http://www.aber.ac.uk/cy/>

Prifysgol y Flwyddyn yng Nghymru - The Times & The Sunday Times 2020.

Aberystwyth University www.aber.ac.uk/en/<http://www.aber.ac.uk/en/>

Welsh University of the Year - The Times & The Sunday Times 2020.

----------------------------------------------------------------------------------------------------------------------

--
You received this message because you are subscribed to the Google Groups "World Robotic Sailing Championships Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrsc-discuss...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/8e323e5c-d55b-46ba-813a-a333e9659d41n%40googlegroups.com<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fwrsc-discuss%2F8e323e5c-d55b-46ba-813a-a333e9659d41n%2540googlegroups.com%3Futm_medium%3Demail%26utm_source%3Dfooter&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C0%7C637352455364420415&sdata=GLnvpNHuWgxGfabPsw9wEyFsy%2BSbJ%2BIJOmwwsNO%2FLOI%3D&reserved=0>.

--
Dr. Colin Sauze
Research Software Engineer
Supercomputing Wales Project
Room 2.16, Physical Sciences Building
Aberystwyth University,
Penglais, Aberystwyth
Ceredigion, UK, SY23 3DB

Tel: +44 (0)1970 622774<tel:+44%201970%20622774>

Webpage: http://users.aber.ac.uk/cos


----------------------------------------------------------------------------------------------------------------------

Pantycelyn yn agor mis Medi 2020. https://www.aber.ac.uk/cy/accommodation/accommodation-options/catered/pantycelyn/

Pantycelyn opening September 2020. https://www.aber.ac.uk/en/accommodation/accommodation-options/catered/pantycelyn/

----------------------------------------------------------------------------------------------------------------------

Prifysgol Aberystwyth www.aber.ac.uk/cy/<http://www.aber.ac.uk/cy/>

Prifysgol y Flwyddyn yng Nghymru - The Times & The Sunday Times 2020.

Aberystwyth University www.aber.ac.uk/en/<http://www.aber.ac.uk/en/>

Welsh University of the Year - The Times & The Sunday Times 2020.

----------------------------------------------------------------------------------------------------------------------


--
You received this message because you are subscribed to the Google Groups "World Robotic Sailing Championships Discussions" group.

To unsubscribe from this group and stop receiving emails from it, send an email to wrsc-discuss...@googlegroups.com<mailto:wrsc-discuss...@googlegroups.com>.
To view this discussion on the web, visit https://groups.google.com/d/msgid/wrsc-discuss/63509e4e-da4d-4692-995b-9d216ba0ab4cn%40googlegroups.com<https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fwrsc-discuss%2F63509e4e-da4d-4692-995b-9d216ba0ab4cn%2540googlegroups.com%3Futm_medium%3Demail%26utm_source%3Dfooter&data=02%7C01%7C%7C483e299005e941c8569208d854ae6a10%7Cd47b090e3f5a4ca084d09f89d269f175%7C0%7C0%7C637352455364420415&sdata=vAMePJvsuCiGwnuR%2F%2FBrtTeNJYS32tXv%2ByqA2VqKnfo%3D&reserved=0>.

[Rhys Mainwaring]

Hi Colin,

I suspect the issue may be similar to the WSL docker image, that without hardware acceleration for OpenGL the CPU load is prohibitively high. Setting the simulation to wireframe probably won't help that much as there are still a set of OpenGL calls made each frame update from the plugin when the mesh vertices are updated. There are some articles on hardware accelerated GUI apps on docker with VNC using VirtualGL such as this: HW accelerated GUI apps on Docker, but I have not tried implementing the approach. Happy to take a look but let me know if you are more familiar with this as I'm coming at the google cloud services from a standing start.

Rhys