The Impact of X299 Motherboards on i9 Thermals & Power

By Published September 29, 2017 at 2:00 pm
  •  

We’ve talked about this in the past, but it’s worth reviving: The reason or keeping motherboard consistency during CPU testing is the inherent variance, particularly when running auto settings. Auto voltage depends on a lookup table that’s built on a per-EFI basis for the motherboards, which means auto VIDs vary between not only motherboard vendors, but between EFI revisions. As voltage changes, power consumption changes – the two are directly related – and so too the wattage changes. As a function of volts and amps, watts consumed by the CPU will increase on motherboards that push more volts to the CPU, regardless of whether the CPU needs that voltage to be stable.

We previously found that Gigabyte’s Gaming 7 Z270 motherboard supplied way too much voltage to the 7700K when in auto settings, something that the company later resolved. The resolution was good enough that we now use the Gaming 7 Z270 for all of our GPU tests, following the fix of auto voltages that were too high.

Today, we’re looking at the impact of motherboards on Intel i9-7960X thermals primarily, though the 7980XE makes some appearances in our liquid metal testing. Unless otherwise noted, a Kraken X62 was used at max fan + pump RPMs.

Motherboards & EFI Used

We’re using three X299 motherboards for this content piece, with five different EFI revisions across them. The ASUS Rampage VI Extreme was our primary board for the review and liquid metal testing, using the latest 0802 BIOS revision for current and previous testing. This board was joined by the ASUS Prime X299-Deluxe motherboard, using both the pre-7980XE launch EFI and the latest 0802 revision. The Gigabyte Gaming 9 also joined these, using the latest EFI revision (F6) and the previous 7900X press launch EFI.

The first ASUS Prime and Gaming 9 tests were conducted with BIOS revisions that came out well before the new CPUs, which means that the profiles aren’t tuned for the new CPUs. Updating EFI changes results, as you’d expect, because the auto-voltage lookup table will behave differently.

We’re just using Blender today. Prime95 is a great test, but its power cycling makes it difficult to use head-to-head in some cases. We did run Prime in a lot of these tests anyway, but Blender gets the job done well enough to convey what’s happening.

We’re using the Intel i9-7960X for these tests, using our liquid metal modded version.

Auto Voltage ID for Blender Workload

x299 vid all

Let’s start with voltage ID behavior, since this directly influences everything else. Using all auto settings, the first Gaming 9 test, using pre-launch BIOS, pegged VID at roughly 1.01VID dead for the entire test, never really changing. The ASUS Rampage motherboard fluctuates and bounced around between 0.91VID and 1.0VID, generally sticking to 0.96-0.97VID for the entirety of the test. We only started using this board with the 7960X launch, and it had the latest 0802 BIOS on it. The board has a tune for the CPU, and seems to behave as you’d expect.

The ASUS Prime board with the original EFI also bounced around a lot – at least, a lot more than the version with EFI version 0802. We averaged 0.89 to 0.96VID with the pre-launch version, and 0.96 to 1.0 with the 0802 version.

The most noticeable jump is the Gigabyte Gaming 9 EFI update to version F6 from the 7900X reviewer EFI version. We’ve moved from 1.012VID to a range of 1.065 to 1.08VID, depending on which core is measured. This will have significant impact on both power and thermals.

Power Consumption vs. Motherboards

x299 pwr all

For power testing, we are measuring at the EPS12V rails, which means the bulk of our power consumption is the CPU. Some other board components will also pull power from here, but nothing close to the CPU’s consumption. That said, there is some error involved in changing motherboards for power tests, even when measuring at the EPS12V cables.

The Rampage VI Extreme on auto and 7960X are measuring about 217W at the EPS12V cables. Again, this is under auto conditions and with a Blender workload, so it’s not the most power we could pull – but also not the least.

The ASUS Prime with the first EFI revision, which ran overall lower voltages of 0.89 to 0.96VID versus the average 0.91 to 0.98 of the Rampage, managed to consume about 206W on average. Updating EFI to version 0802 moved our VID to 0.96 to 1.0, which consequently shifted power consumption up to 222W. That’s just from an EFI change.

The Gigabyte Gaming 9 with its original EFI had a higher sustained VID of 1.012 constantly, which produced a higher power consumption. The Gaming 9 with this EFI version measured at 230W power consumption, now about 30W more than the lowest board and about 15W higher than the previous 3-configuration average.

We ran two tests of the Gaming 9 with version F6, since the results were so much higher in power consumption than the others. The tests had some variance, but were both overall much higher in power consumption: we were between 255 and 268W, depending on which test. That’s now 50-60W higher than the lowest board on the stack, which shows just how much auto voltage tables can impact results.

Thermal Impact from Motherboards

x299 temp all

Thermally, voltage also impacts results. We have a tolerance of +/-3C for ambient and thermal application differences here, but can still see gaps in performance. The Rampage VI cooled the best – some of this is because of the spacing of the VRMs from the socket – and landed around 40-44C. The Gaming 9 with launch EFI plotted at 42-46C, with the ASUS Prime tests at around 46.5C for 0802. [cut orange line] The Gigabyte Gaming 9 with the F6 EFI plotted at around 50-51C, for a 10-degree gap versus the lowest temperatures on the plot.

Some Auto Numbers

blender auto chart

prime95 auto chart

This just shows the thermal performance under auto conditions with the ASUS motherboard. Note that TIM & LM cannot be compared head-to-head in these charts, because auto voltage means variable voltage – that means the tests are not identically executed on each configuration, as variance is introduced from the automatic voltage. That’s why we use fixed voltages and frequencies for TIM comparison tests.

Conclusion

The only real conclusion here is to be on the lookout for auto voltage supplied by a motherboard. If you have the time, we’d recommend manually tuning voltage down until the threshold of instability is found, then increment until the platform becomes stable. This will reduce power consumption and temperatures directly, while remaining just as functional as with the overvolted configuration of most stock boards.

Editorial: Steve Burke
Video: Andrew Coleman 

Last modified on September 29, 2017 at 2:00 pm
Steve Burke

Steve started GamersNexus back when it was just a cool name, and now it's grown into an expansive website with an overwhelming amount of features. He recalls his first difficult decision with GN's direction: "I didn't know whether or not I wanted 'Gamers' to have a possessive apostrophe -- I mean, grammatically it should, but I didn't like it in the name. It was ugly. I also had people who were typing apostrophes into the address bar - sigh. It made sense to just leave it as 'Gamers.'"

First world problems, Steve. First world problems.

We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.

Advertisement:

  VigLink badge