We tested the AMD Radeon 5700 XT in two systems, a 6-core 3.7 GHz Intel i7 8700 in an EVGA Z370 Classified K motherboard, and a 12-core 3.8 GHz AMD Ryzen 9 3900x in an NZXT H500i system with an x570 motherboard.
We ran six different tests, at two resolutions (TimeSpy Extreme, TimeSpy, Civilization VI 1440 & 4K High, Division 2 1440 & 4K, Shadow of Tomb Raider 1440 & 4K, Metro Exodus 1440 & 4K RTX) on four different graphics AIBs (RTX 2070, RTX 2080, RTX 2060 Super, RTX 2070 Super) in each system (a total of 262 tests).
In addition, we ran PC Mark and the CPU test of PassMark.
We averaged all those tests and got the following results:
|Graphics benchmark differences between AMD Ryzen and Intel core i7 with AMD AIB|
The 6-core Intel i7-8700 on average gave a 17% higher performance on graphics tests.
On the PassMark CPU test, the systems got the same score.
However, when Nvidia AIBs were used on PC Mark, there is a totally different result.
|PC Mark benchmark differences between AMD Ryzen and Intel core i7 with Nvidia AIB|
What do we think?
Additional CPU cores don’t seem to help game performance very much. It’s a bit ironic considering the game developers, the ISVs, used to develop for consoles first and then port to PC. When AMD got the game console market with their multi-core APUs you’d think things would change. But the game developers have always striven for the bottom designing for the lowest common denominator—people with dual-core CPUs. And the irony of that is the gamers who are in the high-end category are often 33-year-old men with a job and discretionary income. They can afford to buy a PC with a 12-core CPU in it. They already buy high-end AIBs. And yet the game ISVs cling to their 1999 concepts.
So in overall PC performance as PC Mark measure AMD shines, but in games, well, no so hot.