DX11 vs DX12 Intel 4770K vs 5960X Framerate Scaling

This is our third installment of looking at the new DX12 API and how it works better with a game such as Ashes of the Singularity. We have looked at how DX12 is better at distributing workloads across multiple CPU cores than DX11 in AotS. This time we are compare Haswell and Haswell-E processors' AotS performance.


Test Setup and Reasoning

Our system configuration for this testing is below, and likely a bit different than you have seen in awhile.

Article Image

For our tests a couple weeks ago using our AMD CPU, we were focused on looking at results that were very much CPU bound, but we also wanted to have a lot of cores at our disposal. In our second article, we wanted to again look at CPU bound results with our Intel CPU.

For THIS ARTICLE we wanted to focus on scaling of cores, we could have easily gone in and turned cores on or off in software/BIOS, but I always feel a bit dirty about doing that, so I headed over to storage and picked up a Z97 Chipset motherboard and had a couple of 4770K processors on hand as we have been switching test machines over to Skylake processors.

This way we have matched clocks on an Intel Core i7-4770K Haswell processor and a Core i7-5960X Haswell-E processor. So 4C/8T vs 8C/16T respectively.

Drivers have changed on both the Red and Green side of things in the last week, but for the scope of our article in terms of DX11 vs DX12 I do not see that updating our driver is going to change any conclusions we have here after spot checking some of the numbers. We are using the AMD driver; Crimson Edition 16.4.1. We are using the latest NVIDIA WHQL driver v364.72.

I am using Intel XTU program to control CPU clocks.

So again, like we did last time, we are going to run through the Ashes of the Singularity benchmarks and graph all the data and see how it looks in both GPU bound and CPU bound scenarios.