Why you can trust Tom's Hardware Our expert reviewers spend hours testing and comparing products and services so you can choose the best for you. Find out more about how we test.
Benchmarks & Conclusion
Synthetic Benchmarks
The X99P suffers a slight, but noticeable and repeatable deficit in the graphics portion of our synthetic tests. This could be worrisome, except the real-world gaming benchmarks are completely normal, suggesting this is only on the synthetic bench side.
None of our Sandra tests tax the GPU so we see consistent performance here. The CPU centered test results are even, but we do see an impressive lead in memory bandwidth, even at stock JEDEC timings (which also bumps the crypto performance slightly). This will affect other benchmarks down the road.
Once again we see the X99P slightly lagging in synthetic GPU tests in Cinebench and Compubench. Cinebench CPU tests show a slight lead due to the X99P's small BCLK advantage. The Gigabyte board's storage performance is very middle of the pack.
Gaming Benchmarks
The Carbon and X99P are close enough in RAM bandwidth that any performance difference in F1 2015 would be within margin of error. Everywhere else it's a dead heat.
Application & Productivity Benchmarks
A longer running benchmark like Handbrake is a perfect place for a small BCLK advantage to translate into a large lead.
Power & Temperature
Whereas the other boards were quite steady, the X99P's idle power fluctuates quite a bit, bouncing between 73W and 84W. It also has a high draw of 4W even when turned off (compared to .7W for the other two). Under load we see the generous auto voltage settings kick in, though it still sneaks in just under the Taichi.
The heat pipe between the two motherboard heatsinks makes itself known with a very low VRM temperature. However, CPU temperature is greatly affected by the extra voltage.
Overall Performance & Efficiency
Nothing to see here. Let's move along.
Again, not a very exciting graph for the Gigabyte board as it falls almost perfectly halfway between the other two in efficiency.
Overclocking Performance
While the Carbon technically beats the X99P in total CPU overclocking, it's effectively a dead tie. As we said earlier, the X99P's small edge in BCLK overclocking may not matter to generalists, but those who want to fine tune clocks are well served by the extra granularity.
Despite a lower maximum RAM frequency than the Carbon, the X99P gets more out of the RAM at slower speeds. The exciting part here is the impressive bandwidth at DDR4-2666 levels. Those who don't want to spend extra money on DDR4-3200 or DDR4-3400 modules can still expect great RAM performance using this board.
Performance Value
Being able to match the Carbon in just about every way for much less money is a notable feat. Thunderbolt and 4-way SLI support are great bonuses on top of that.
Final Word
Like MSI's Gaming Pro Carbon before it, the X99P-SLI is full of highs and lows. Adding Thunderbolt 3 to a less costly X99 board is sure to please some folks, especially with Gigabyte throwing in the patch cables for use with Thunderbolt displays. Proper 4-way graphics support that doesn't cripple other connectivity options always pleases us. Perhaps most impressive of all is the RAM performance, which provides better balance of stability and sheer bandwidth than boards that cost significantly more. If any of these facets are high on your priority list, the X99P's relatively low price makes it a very attractive option.
However it's lacking in some small but very noticeable areas. The board is specifically marketed to the multi-graphics crowd (especially with SLI in the product name), but you don't even get enough bridges in the box to utilize the feature. While I don't specifically need bench power buttons and an exterior clear CMOS button, an error debug display is practically required for anyone doing even moderate overclocking. Not only that, but Gigabyte goes through the expense and effort of putting two UEFI chips on the board and doesn't give the user access to both of them. And perhaps I'm a fan snob, but I refuse to purchase a motherboard for my personal systems if it doesn't have at least two PWM fan headers. I don't want to hear fans constantly spinning up and down because the motherboard effectively only runs them at 60% and 100% duty speed. Configuring the fan curves in the UEFI is also frustratingly bad.
In the end, the good things outweigh the bad. This is a solid motherboard. The complaints are not performance or reliability related, but creature comforts, and those don't apply to all consumers. It's not as well rounded or as competitively priced as the X99 Taichi, so it doesn't get our Recommended stamp. But this is still an Approved board.
MORE: Best Motherboards
MORE: How To Choose A Motherboard
MORE: All Motherboard Content
