Thank you it is very useful. Just something I did not understand is the comparison between the CPU only and CPU+GPU. Which one was more quicker? I felt that the speed of the convergence for both is the same, am I right?
Excellent video. Congratulations. In my case I have a workstation with an msi X99A TOMAHAWK motherboard with an Intel Xeon E5-2699 v3 processor, (and I currently use 3 monitors). Because of this I installed a GPU, AMD firepro w7100 which works very well for me in Solidworks. The RAM is Non-ECC 32 gigabytes. The problem is that I am learning to use ANSYS Fluent, and this software is married to Nvidea, and for GPU calculation acceleration, looking at the Ansys GPU compatibility lists, I see that the K80 is used, and taking into account Considering the second-hand price, I'm interested in purchasing one. How can I configure my system to install an Nvidea Tesla K80 as a GPU for GPU+CPU acceleration calculations, but the AMD GPU works as an image or video generator for my monitors as it currently does? Does the Nvidea K80 gpu have 24 GB of ram, can this be affected when using this gpu in conjunction with the AMD GPU that only has 8 GB of ram? Would the K80 be restricted to the RAM of the Firepro w7100? My PSU is 700 watts. Thank you.
Hi, How will this affect the model partition segmentation ? (more CPU = More partitions on solve area). Is it better to run with 1 gpu + 1 Processor OR 1 gpu + 8 processor ?
Thank you for the useful information. I just plan to buy a new computer, if I have 24 cores with nvidia tesla 80K(2GPU), so I should set the number of CPU to 12 and GPU to 2 ?.
many Thanks sir for your explanation however I want kindly to ask you I have core i7 and RTX 2060 what is the recommended configuration for me if you can help please
Hi, I have GPU the same as yours, NVIDIA Quadro RTX 400, when I type in TUI >parallel/gpgpu/show, it shows the number of GPUs 0, to make fluent detect my GPU? thanks.
Thank you for this video! I have an intel i7 8750H (6 core, 12 thread) processor, and an Nvidia Geforce GTX 1060 MaxQ. I read that the number of processes should be twice the number of cores on the cpu, so I'm using 12 processes and 1 gpgpu. Is this correct?
Hi friend, Ansys solver only uses the cores (it doesn't use the threads). For CFD simulations you must disable threads in BIOS or UEFI. According to the information from Ansys in your case you should use 3 cores + 1GPU, however I do not think it is very efficient because the GTX series is used for games. I recommend you do several tests with some simulations and see which configuration you get faster results. Bet regards
@@ansys_cfd Hi, thanks a lot for your video. As I understand from your comment, disabling hyperthread from BIOS and using 6 processors is faster than enabling hyperthread and using 11 processor in fluent. Is this right? But I did it and the results show that when I enabled the hyperthread and used more processors , the simulation became faster.
I am having Intel Xeon E3-1245 V3 processor with Nvidia Quadro K2200 Graphics and 32GB RAM. What configuration will be more efficient for me to run the simulations.
Hello Friend, great video! Sometime I Read that GPU use isn't compatible with VOF model. Do you know something about it? I have an Intel i7-10870H processor with 8 core and 16 thread , and an NVIDIA Geforce GTX 1650 Ti with 8 GB. Do you think I can use that GPU or is it definitely impossible because it is more suitable for Gaming? The Gamming GPU card don't are adequate for Fluent?
Hi friend, Ansys has tested only some models using GPUs, ideally each of us can do tests and obtain conclusions, because the times to obtain results vary depending not only on the solvers but also on our hardware. It is possible to use NVIDIA GTX but they are inefficient. I did a test with a GTX 2060 and the time increased while with an RTX 4000 the calculation time decreased. In this video the simulation is using VOF and in the end I came to the conclusion that using the GPU the time decreases, only for 100 time steps, if I had completed the whole simulation the time gain will be greater compared to only the use of the CPU. My advice is to do several tests and compare the results. Best regards
you have professional license right? in my case, with the student version, this configuration does not run. the message in the solution module is: The FLUENT application could not validate the connection.
Hello, how to use graphics card to do the computation any simulation on Solidworks (e.g. static study, topology optimization etc)? I have a low processor(intel core i5 13400f) with 16 GB RAM, and GPU RTX 3060. Thank You.
Hey Ninja, I have the same GPU as yours, with 18 cores. When I try to utilize my GPU, FLUENT shuts down during the initialization point. Any ideas why this might happen?
@@ansys_cfd Thanks for the reply! I have the Enterprise license, which I think is the commercial one? It allows me to run multiple processes at once. For reference, I can run a 15 core simulation, but the moment I try to use the GPU it shuts down with an error. Thanks again!
@@shitijvishwakarma1166 i think your gpu is running out of memory, to verify this create a transcript in fluent by going to file > write > start transcript. run your simulation again, if it closes, then go to where you stored your transcript file and open it using notepad, it should tell you the error. are you using coupled solver by any chance? if you face the same issue still, change your solver and try, it should help
Thank you for your useful video Sir I have some problem When i type this command "parallel/gpgpu/select. It says command invalid . How can i solve this issues? Thanks
I have RTX3060 laptop GPU + Ryzen 7 5800H.. When I select 4 cores CPU and 1 GPGPU per machine, CPU works fine in task manager but my CUDA still 0%.. Can you help me
Thank you!
You're welcome friend
Best regards
Thank you it is very useful. Just something I did not understand is the comparison between the CPU only and CPU+GPU. Which one was more quicker? I felt that the speed of the convergence for both is the same, am I right?
What license do you have?
does ansys 2022 work without graphic card, if not then which graphic card shall I install on my desktop
Excellent video. Congratulations.
In my case I have a workstation with an msi X99A TOMAHAWK motherboard with an Intel Xeon E5-2699 v3 processor, (and I currently use 3 monitors). Because of this I installed a GPU, AMD firepro w7100 which works very well for me in Solidworks.
The RAM is Non-ECC 32 gigabytes.
The problem is that I am learning to use ANSYS Fluent, and this software is married to Nvidea, and for GPU calculation acceleration, looking at the Ansys GPU compatibility lists, I see that the K80 is used, and taking into account Considering the second-hand price, I'm interested in purchasing one.
How can I configure my system to install an Nvidea Tesla K80 as a GPU for GPU+CPU acceleration calculations, but the AMD GPU works as an image or video generator for my monitors as it currently does? Does the Nvidea K80 gpu have 24 GB of ram, can this be affected when using this gpu in conjunction with the AMD GPU that only has 8 GB of ram? Would the K80 be restricted to the RAM of the Firepro w7100?
My PSU is 700 watts.
Thank you.
what happened if we don't divided the core numbers by 2? Thank you 😊
Hi,
How will this affect the model partition segmentation ? (more CPU = More partitions on solve area). Is it better to run with 1 gpu + 1 Processor OR 1 gpu + 8 processor ?
Thank you for the useful information. I just plan to buy a new computer, if I have 24 cores with nvidia tesla 80K(2GPU), so I should set the number of CPU to 12 and GPU to 2 ?.
im also planning on buying tesla k80 with 11900, so how is it going for you?
you can use all your cpu i.e. 24 cores with 2 GPUs, since 24 is divisible by 2 which is evenly distributed among the two GPUs
what do you think about the ansys discovery live it is faster how accurate its results ?
Hello, Please which type of AMD threadripper do you use for comparison?
3960X
If I recall correctly, to use GPU the scheming model has to be coupled right? You can't use any other schemes to use the gpu?
Can we use normal graphic card instead of Quadro?
hello do you know where we can find benchmark with rtx 4090 ?
many Thanks sir for your explanation however I want kindly to ask you I have core i7 and RTX 2060 what is the recommended configuration for me if you can help please
my cpu core 6 and gpu igame rtx 3050.....how to use gpu mood??
When i type this command "parallel/gpgpu/select. It says command invalid . How can i solve this issues? Thanks
If i have dual core with 14 core for each. So, to utilize all my resources i should put 28 core in processes tab is that right?
so it gets you 1:21 vs 1:18 :D 3.7% speed up?
Hey man, thanks for the info and tutorial! really helped
Hi, I have GPU the same as yours, NVIDIA Quadro RTX 400, when I type in TUI >parallel/gpgpu/show, it shows the number of GPUs 0, to make fluent detect my GPU?
thanks.
Thank you for this video! I have an intel i7 8750H (6 core, 12 thread) processor, and an Nvidia Geforce GTX 1060 MaxQ. I read that the number of processes should be twice the number of cores on the cpu, so I'm using 12 processes and 1 gpgpu. Is this correct?
Hi friend,
Ansys solver only uses the cores (it doesn't use the threads). For CFD simulations you must disable threads in BIOS or UEFI. According to the information from Ansys in your case you should use 3 cores + 1GPU, however I do not think it is very efficient because the GTX series is used for games. I recommend you do several tests with some simulations and see which configuration you get faster results.
Bet regards
@@ansys_cfd Okay, thanks!
@@ansys_cfd Hi, thanks a lot for your video. As I understand from your comment, disabling hyperthread from BIOS and using 6 processors is faster than enabling hyperthread and using 11 processor in fluent. Is this right? But I did it and the results show that when I enabled the hyperthread and used more processors , the simulation became faster.
@@powerpower7361 the best answer is trying and seeing.
@@ansys_cfdWho told u? I m also using my threads for parallel solution and my solution is faster.
Just amazing, thanks a lot
I got the quadrro rtx 4000 with ryzen 2970wx but my fluent shuts off as i try to run any simulations.
please help!
I am having Intel Xeon E3-1245 V3 processor with Nvidia Quadro K2200 Graphics and 32GB RAM. What configuration will be more efficient for me to run the simulations.
since you only have a single gpu, you can use all the cores/threads and one gpu when starting with a new simulation in fluent.
Can amd gpus also use for run a simulation for workbench?
not from what I have read, but I am a noob
Hello Friend, great video!
Sometime I Read that GPU use isn't compatible with VOF model. Do you know something about it?
I have an Intel i7-10870H processor with 8 core and 16 thread , and an NVIDIA Geforce GTX 1650 Ti with 8 GB. Do you think I can use that GPU or is it definitely impossible because it is more suitable for Gaming? The Gamming GPU card don't are adequate for Fluent?
Hi friend,
Ansys has tested only some models using GPUs, ideally each of us can do tests and obtain conclusions, because the times to obtain results vary depending not only on the solvers but also on our hardware. It is possible to use NVIDIA GTX but they are inefficient. I did a test with a GTX 2060 and the time increased while with an RTX 4000 the calculation time decreased. In this video the simulation is using VOF and in the end I came to the conclusion that using the GPU the time decreases, only for 100 time steps, if I had completed the whole simulation the time gain will be greater compared to only the use of the CPU. My advice is to do several tests and compare the results.
Best regards
@@ansys_cfd thank you!
you have professional license right?
in my case, with the student version, this configuration does not run. the message in the solution module is: The FLUENT application could not validate the connection.
Hello, how to use graphics card to do the computation any simulation on Solidworks (e.g. static study, topology optimization etc)? I have a low processor(intel core i5 13400f) with 16 GB RAM, and GPU RTX 3060. Thank You.
Hi. Did you find an answer? I have same question
There is a solution and it worked for me. I have an RTX 2060: ua-cam.com/video/AG0UdUO4sfA/v-deo.html
Do I need to have installed CUDA toolkit ?
Hi friend,
You need a Nvidia graphic card only.
Best regards
@@ansys_cfd I have the student version and CUDA doesnt work , maybe its due to that
Hey Ninja, I have the same GPU as yours, with 18 cores. When I try to utilize my GPU, FLUENT shuts down during the initialization point. Any ideas why this might happen?
Hi friend,
Are you using commercial or research license?
@@ansys_cfd Thanks for the reply! I have the Enterprise license, which I think is the commercial one? It allows me to run multiple processes at once. For reference, I can run a 15 core simulation, but the moment I try to use the GPU it shuts down with an error. Thanks again!
@@shitijvishwakarma1166 i think your gpu is running out of memory, to verify this create a transcript in fluent by going to file > write > start transcript. run your simulation again, if it closes, then go to where you stored your transcript file and open it using notepad, it should tell you the error. are you using coupled solver by any chance? if you face the same issue still, change your solver and try, it should help
What computer specification do you use when operating that GPU?
5:34, Which threadripper model? and how many cores used in it?
Ryzen threadripper 2950x, ir has 16 cores, I usted 8 only.
@@ansys_cfd The comparison is 8 cores + GPU and 8 cores?
or 8 cores + GPU and 16 cores?
8 cores vs 8 cores + gpu
@@ansys_cfd Thanks for the video, It's excellent but I have a question, why did you use only 8 cores without GPU?
Thank you for your useful video Sir
I have some problem When i type this command "parallel/gpgpu/select. It says command invalid . How can i solve this issues? Thanks
Bro did you able to solve the issue
@@kartikgoyal3963 yes
i wanted to select the gpu, however, it said invlaid command for the following : parallel/gpgpu/select .
it's the same with my computer
did you chosse the right amount of gpus, in the set up? in my case "select" is only invalid when I chosse more gpus than what I really have.
Hi, can we use gpu for CFX?
Hi friend,
GPU works un Ansys Fluent only.
Regards
I have RTX3060 laptop GPU + Ryzen 7 5800H.. When I select 4 cores CPU and 1 GPGPU per machine, CPU works fine in task manager but my CUDA still 0%.. Can you help me
My friend, I have a Rtx 3070 TI with the same problem, did You solve it?
@@ingjoselo Friend I have Rtx 3080 and have same problem
is working well nvidia rtx 3050ti in fluent is rtx system problem for Ansys ,should I change my desktop
+
++
This software looks like a toy :D
I can do higher resolved simulations on the same hardware in seconds instead of hours :D
Ninja i need a little guidance in a project can you please assist me via email.