Ryzen 1950x is very slow in TYFLOW

[A.kavyani]

Hi,
first of all thank's for your great work.

I tried Tyflow on my laptop with 7700 HQ cpu and speed was great but when i tried same scene on my computer ryzen 1950x 32gb ram corsair 3600 mhz speed was 4time slower than my 7700hq.

I turned of auto multi thread and selected 16thread of my cpu
result was faster than when I simulate with all of cpu threads.

does anyone have the same problem?

[A.kavyani]

Video download link

[Heerockss]

Hi,
first of all thank's for your great work.

I tried Tyflow on my laptop with 7700 HQ cpu and speed was great but when i tried same scene on my computer ryzen 1950x 32gb ram corsair 3600 mhz speed was 4time slower than my 7700hq.

I turned of auto multi thread and selected 16thread of my cpu
result was faster than when I simulate with all of cpu threads.

does anyone have the same problem?

That sounds like a similar problem some people do have with PhoenixFD and Ryzen CPU`s. They have some problems with NUMA that slows down the calculation considerably. 

I run into a similar problem with my dual xeon workstation. When calculating lower resolution simulations, a single numa node (less threads) are faster than with both nodes. I think its because the communication between the two nodes takes longer than the actual calculation process.

[A.kavyani]

Hi,
first of all thank's for your great work.

I tried Tyflow on my laptop with 7700 HQ cpu and speed was great but when i tried same scene on my computer ryzen 1950x 32gb ram corsair 3600 mhz speed was 4time slower than my 7700hq.

I turned of auto multi thread and selected 16thread of my cpu
result was faster than when I simulate with all of cpu threads.

does anyone have the same problem?

That sounds like a similar problem some people do have with PhoenixFD and Ryzen CPU`s. They have some problems with NUMA that slows down the calculation considerably. 

I run into a similar problem with my dual xeon workstation. When calculating lower resolution simulations, a single numa node (less threads) are faster than with both nodes. I think its because the communication between the two nodes takes longer than the actual calculation process.

Thank's Heerockss,
I compared a simple sim in phoenixFD
7700 HQ : 9m 40 s
1950x : 4m 28s
in this comparision 1950 is 2x faster than 7700HQ.
but in tyflow the result is opposite 7700 is 2x faster than 1950.
hope Tyson gives us information about this.

[Heerockss]

Hi,
first of all thank's for your great work.

I tried Tyflow on my laptop with 7700 HQ cpu and speed was great but when i tried same scene on my computer ryzen 1950x 32gb ram corsair 3600 mhz speed was 4time slower than my 7700hq.

I turned of auto multi thread and selected 16thread of my cpu
result was faster than when I simulate with all of cpu threads.

does anyone have the same problem?

That sounds like a similar problem some people do have with PhoenixFD and Ryzen CPU`s. They have some problems with NUMA that slows down the calculation considerably. 

I run into a similar problem with my dual xeon workstation. When calculating lower resolution simulations, a single numa node (less threads) are faster than with both nodes. I think its because the communication between the two nodes takes longer than the actual calculation process.

Thank's Heerockss,
I compared a simple sim in phoenixFD
7700 HQ : 9m 40 s
1950x : 4m 28s
in this comparision 1950 is 2x faster than 7700HQ.
but in tyflow the result is opposite 7700 is 2x faster than 1950.
hope Tyson gives us information about this.

I did also some testing with TyFlow on one of my dual Xeon machines: CPU: 2xXeon e5-2670v3 (24C/48T) @2.7ghz  GPU: Maxwell TitanX

2x 2670v3 (opencl off) 02:09 Min  <--- Both Numa Nodes at 100% usage but really slow


1x 2670v3 (opencl off) 01:11 Min  <--- Only one Numa Node!

1x 2670v3 (opencl on) 00:38 Min  <--- A lot faster with OpenCL and one Numa Node

2x 2670v3 (opencl on) 1:12 Min  <--- And again lot slower with both Numa Nodes


I think there is indeed a slowdown for multiple numa nodes. Why your AMD Chip performs well in Phoenxi but bad in TyFlow i cant answer but dont forget that not all Ryzen or Threadripped behaved like this in Phoenix, there are some exceptions.

Yes our king and savior Tyson Ibele might enlighten us

[tyFlow]

Unfortunately I don't have a Ryzen to test so I can't give conclusive answers as to why you're experiencing this...could be a multitude of factors....sorry Hopefully I'll have an opportunity to figure this out in the future.

[A.kavyani]

Thank's Heerockss for tests result was intresting.
Hope other TR4 users help to tyson about this.

---

Unfortunately I don't have a Ryzen to test so I can't give conclusive answers as to why you're experiencing this...could be a multitude of factors....sorry Hopefully I'll have an opportunity to figure this out in the future.

Hi tyson,
I think that's because of Opencl is turned off for amd.

[MeHoo]

I doubt this is it, but for one of my machines, max actually wasn't loading using the dedicated GPU. It was using the integrated video by default. I had to force it through the Nvidia settings. Perhaps similar issue with your setup? I have no idea why that happened to me, but it did lol.

[tyFlow]

Oh...that would definitely explain the speed difference!

Could you open the debugging rollout of your tyFlow and click the OpenCL button? That will print out information about your devices in the MAXScript listener, and might help explain why tyFlow can't find a valid OpenCL device.

[A.kavyani]

Hi tyson,


System OpenCL (GPU) Info:

OpenCL.DLL found...

No OpenCL platforms found...


OpenCL acceleration not available!

[tyFlow]

Could you make sure you have the latest drivers for your GPU?

You may also have to download the latest OpenCL drivers from Intel.

[A.kavyani]

Could you make sure you have the latest drivers for your GPU?

You may also have to download the latest OpenCL drivers from Intel.

My nvidia driver is 419.67 for 2x1080
didn't success to find opencl for my cpu,can anyone help?

[MeHoo]

Double check in manage 3d settings in the Nvidia driver and manually create a profile for Max if there isn't already one. Also make sure SLI is off and your dedicated physx card is the second card. This happened to me once for both Lucid Physx and Photoscan.

Edit: oh, also if you are using any remote desktop applications for this, ensure they are compatible with open cl. Windows remote desktop kills open cl. Chrome remote desktop doesn't. Just a thought.

[A.kavyani]

Double check in manage 3d settings in the Nvidia driver and manually create a profile for Max if there isn't already one.  Also make sure SLI is off and your dedicated physx card is the second card.  This happened to me once for both Lucid Physx and Photoscan.

Edit: oh, also if you are using any remote desktop applications for this, ensure they are compatible with open cl.  Windows remote desktop kills open cl.  Chrome remote desktop doesn't.  Just a thought.

Thank's MeHoo,
tried but did not succeed activate opencl.

[A.kavyani]

my 1950x is 2xfaster than 7700hq without opencl hope anyone can help.
updated driver to the 425.31
Created a new profile for max.
tried both cards as physx and opencl.
but I did not succeed opencl is disabled.