Whitepapers

Ansys_Fluent_GPU

Issue link: https://resources.randsim.com/i/1508712

Contents of this Issue

Navigation

Page 4 of 6

E L I T E C H A N N E L PA R T N E R simulation@rand.com | RandSIM.com | 888.483.0674 Does Ansys Fluent with a Native GPU Solver Live Up to the Hype? Test 3 – Compressible Flow in Generic Combustor In this scenario, the flow reached Mach 1. It is important to note that currently, simulating a strong shock wave isn't possible, but that capability may be coming soon. This test is based on 24 million cells and was run on an RTX 8000 with 48 gigs of RAM. No GPU double precision was run due to memory shortage. The meshing topology used is poly-hex, which is becoming more and more prevalent nowadays. The GPU was equivalent to about 90 CPU cores in this test, the fastest speed achieved in this study. This test demonstrated that the GPU was the equivalent to a blistering 90 CPUs. We expect that the mesh topology along with the large number of cells allowed the thousands of cores in the GPU to unleash their power. These amazing results were based on a $5,000 graphics card installed on a CFD prep/post machine and used an Enterprise License, a $7,000 upgrade over a Premium license. On the HPC route, hardware plus software would cost closer to $100,000.

Articles in this issue

Links on this page

view archives of Whitepapers - Ansys_Fluent_GPU