Intel made quite a splash with its 8th generation Core desktop processor rollout. Bolstered by two more cores than its predecessor, the Core i7-8700K restored Intel's domination in both single-threaded and multi-threaded performance metrics. Among other things, Intel increased the standard DDR4 clock speed for these chips to DDR4-2666, up from DDR4-2400 in the previous generation. This presents us with an opportunity to try out the chip with various memory clock speeds and timings to test just how comfortable the memory controller is with higher-than-standard frequencies and tight memory timings.
Apart from the two extra cores, the new "Coffee Lake" silicon, on which these chips are based, shares quite a bit of platform-related machinery with its predecessor, the "Kaby Lake" silicon. The integrated graphics core is practically carried over at the silicon level, with minor performance increases; and so are other uncore components, such as the system agent. The system agent (integrated northbridge) consists of the dual-channel DDR4 integrated memory controller, display I/O, and PCI-Express root complex. The 8th generation Core "Coffee Lake" processors support Extreme Memory Profile (XMP) 2.0, which means high clock-speed memory kits designed for 6th and 7th generation Core processors should support one-click memory overclocking on even these new chips. You simply enable "XMP" over your UEFI setup program's home screen and all the advertised speeds and timings of your memory kit are enabled.
Unlike on the AMD Ryzen processor, where the memory clock is synced with the clock-speed of the InfinityFabric interconnect, translating into significant performance increases with memory overclocking, performance-increases on Intel chips aren't as pronounced. There should still be performance gains to be had when going up from the failsafe 2133 MHz, up to the reference 2666 MHz and beyond.
We paired our Core i7-8700K processor with G.SKILL Trident Z 3866 MHz memory on the Gigabyte AORUS Z370 Ultra Gaming and tried out a large number of memory settings, including DDR4-2133 (JEDEC failsafe), DDR4-2400 ("Kaby Lake" reference), DDR4-2666 ("Coffee Lake" reference); and overclocked speeds of DDR4-2800, DDR4-3000, DDR4-3200, DDR4-3600, DDR4-3866, and DDR4-4000. With some of these clocks, we tried two sets of timings and even tightened the command-rate to 1T on some settings to produce a total of fifteen combinations. These combinations are then put through the selection of tests we put processors through - only in this case, the processor is constant (i7-8700K) while the memory changes. We were careful to use the exact same settings in all tests otherwise, which included "Multi-Core optimization" off in all tests.
This review is particularly significant as DRAM shortages (*cough* price-fixing *cough*) have caused memory module prices to get out of hand. Some of the higher-clocked memory kits can be twice or even thrice as expensive as those running at reference or JEDEC clock speeds. We're about to find out at exactly which memory clock speed you should stop your quest for the "fastest" memory for your i7-8700K powered machine unless you're a professional overclocker and need the fastest memory money can buy for that microscopic edge on the leaderboards.
|Processor:||Intel Core i7-8700K|
|Motherboard:|| Gigabyte Z370 Aorus Ultra Gaming|
|Memory:|| 2x 8 GB G.SKILL Trident Z DDR4|
|Graphics:||NVIDIA GeForce GTX 1080|
|Storage:||Crucial MX300 750 GB|
|Power Supply:||Seasonic SS-860XP|
|Software:|| Windows 10 64-bit Creators Update|
All updates installed
|Drivers:||NVIDIA GeForce 382.05 WHQL|
- All tests were run on the platform listed above
- No BIOS settings other than the memory speeds were changed
- Multi-core optimizations, overclocking, and Turbo tweaks were disabled
Arithmetic CPU TestsIn the first round of tests, we put the processor through three popular synthetic benchmarks that are heavy on number crunching.
Our Patreon Silver Supporters can read articles in single-page format.