We will be putting each chip through its paces with a variety of computational- and gaming-centric tests. We couple each processor with a (roughly equivalent) 16GB dual-channel set of high-speed memory.
Pixel-pushing power comes from a reference GTX 980 Ti graphics card. While the ~£500 graphics card may not be the most appropriate partner for ~£150 CPUs, it does allow us to remove the GPU from the bottleneck equation thus shifting the frame rate onus further towards the CPU.
With that said, I like the idea of using a high-end graphics card of today, even with budget CPUs. In gaming machines, the processor tends to outlive the graphics card. In perhaps two years when the GPU is ready for an upgrade, GTX 980 Ti-level performance is likely to be available at a price point that falls more in line with the shopping radar of ~£150 CPU users.
Despite being capable of a maximum turbo speed of 4.3GHz, we never saw the FX 8370 boost higher than 4.1GHz (20.5x CPU ratio) using our Gigabyte 990FX-Gaming motherboard. This could be tied to the motherboard's BIOS profile, or it could be related to the CPU's ability and cooling solution (although the former is more likely). Our overclocking results bypass this issue by pinning AMD's chip at 4.62GHz.
Gigabyte's 990FX motherboard was unable to pick up our Corsair memory kit's 2400MHz XMP setting. Instead, the XMP configuration was selected (for voltage and timings), but the memory clock was set as 2133MHz.
We did not enable a multi-core turbo profile for the Skylake test system. This allowed the i5-6400 CPU to operate with its default Turbo Boost frequencies, rather than be pinned at the highest multiplier available.
We loaded the XMP configuration on our 3000MHz GeIL memory, but manually selected a 2800MHz speed at stock settings. This allows us to more closely match the memory speed achieved when overclocking the system, and 2800MHz kits are also in a price-vs-performance sweet-spot at this point in time.
Test Systems
CPU |
AMD FX 8370 ‘Vishera'
(Retail) 8 cores, 8 threads |
Intel Core i5-6400 ‘Skylake'
(Retail) 4 cores, 4 threads |
Base Speed |
4.0GHz
|
2.7GHz
|
Max Turbo Speed |
4.3GHz
|
3.3GHz |
Motherboard |
Gigabyte 990FX-Gaming
BIOS: F1 (AM3+, 990FX) |
MSI Z170A SLI Plus
BIOS (Non-K OC): E7998IMS.11T (LGA1151, Z170) |
Memory |
Corsair Vengeance Pro
16GB (2x8GB) @ 2133MHz 10-11-11 DDR3 @ 1.65V |
GeIL Dragon RAM
16GB (2x8GB) @ 2800MHz 14-16-16 DDR4 @ 1.35V |
Overclocked |
4.62GHz CPU @ 1.425V (Med LLC)
2.60GHz NB @ 1.30V 2133MHz CL10 DDR3 |
4.32GHz CPU @ 1.325V
4.32GHz Cache 2773MHz CL14 DDR4 |
Graphics Card |
Nvidia GTX 980 Ti 6GB
(GeForce 361.75 drivers) |
|
System Drive |
SanDisk Ultra Plus 256GB SATA SSD
|
Crucial BX100 250GB SATA SSD
|
Games Drive | Samsung 850 EVO 500GB SATA SSD | |
CPU Cooler |
AMD Wraith
|
Noctua NH-D14
|
Power Supply |
Seasonic SS-760XP 760W Platinum
|
|
Operating System |
Windows 7 Professional with SP1 (64-bit)
|
Comparison Test System:
- For CPU-heavy tests, we pulled a small amount of data from our Skylake launch review (where the changes would not affect the data's integrity). The test systems for those processors can be viewed HERE.
Tests:
- Cinebench R15 – All-core & single-core CPU benchmarks (CPU)
- HandBrake 0.10.2 – Convert 1.23GB 1080P game recording using the High Profile setting and MP4 container (CPU)
- 7-Zip 15.05 beta – Manual video archival (CPU)
- 3DMark 1.5.915 – Fire Strike Ultra (CPU)
- Battlefield 4 – 1920 x 1080, ultra preset (Gaming)
- Far Cry 4 – 1920 x 1080, ultra settings, SSBC, SSAA (Gaming)
- Grand Theft Auto V – 1920 x 1080, maximum quality, 0x AA, 16x AF (Gaming)
- Middle Earth: Shadow of Mordor – 1920 x 1080, ultra quality (Gaming)
- The Witcher 3 – 1920 x 1080, ultra settings, SSAO, DOF (Gaming)
- Tomb Raider – 1920 x 1080, maximum settings, TressFX disabled (Gaming)