[smashy1985]

Hi everyone,
recently we changed the workstation where we are using Star CCM+ and other softwares for 2D and 3D modeling in the office of naval architecture I am working for.

The previous workstation was configured as follows:

- CPU Intel i7-4770K (Quad Core – 3,5/3,9 GHz – Cache 8 MB – HyperThreading)
- 16 GB Ram Dual Channel DDR3 12800 – 1600 MHz (2 x 8 GB)
- Motherboard GIGABYTE – Chipset Intel Z87
- SSD 240 GB + Hard Disk 1 TB
- Graphic Card ASUS GTX 650 – 2 GB GDDR5

While the new one is configured as follows:

- CPU Intel i9-9940X (14 core – 3,3/4,4 GHz – Cache 19,25 MB – HyperThreading)
- 64 GB Ram Quad Channel DDR4 – 2666 MHz (4 x 16 GB)
- Motherboard ASUS WS X299 Pro – Chipset Intel X299
- SSD NVMe 250 GB + Hard Disk 2 TB
- Graphic Card ASUS GTX 1660 Super – 6 GB GDDR6

After instaling the software on the new worstation, using the same licence that allowed us to make parallel simulations, we noticed that the new P.C. is slower than the previous one.

Just two examples:
- #1 : Test simulation (2D External Flow around a Cylinder – 22.000 Cells) started with 6 parallel processes on the same host.
The previous PC processed about 6000 interations in 5 minutes, while the new PC process about 1200 iterations in the same period of time.
- #2 : Simulation “Marine Resistance Prediction: KCS Hull with a Rudder” (Star CCM+ Tutorial – 530.000 Cells) started on the new workstation in Serial mode and in parallel mode with 6 processes on the same host. On both modes, the new workstation calculates the same number of iterations in the same period of time.

According to these examples, I can see that there may be a problem linked to the simulation in parallel mode on the new PC.

Does anyone have an idea in order to solve this issue?

Many thanks for your help.