| BIOS Frequency and Voltage settings (for overclocking) | ||
|---|---|---|
| Asus Rampage IV Extreme | EVGA X79 FTW 151-SE-E777-KR | |
| Base Clock | 80-300 MHz (0.1 MHz) | 85-287.5 MHz (1 MHz) |
| CPU Multiplier | 12x to 57x (1x) | 0x to 60x (1x) |
| DRAM Data Rates | 800-2666 (266.6 MHz) | 1067-2400 (266.6 MHz) |
| CPU Vcore | 0.80-2.10 V (5 mV) | 0.90-1.99 V (1 mV) |
| CPU VCCSA | 0.80-1.70 V (5 mV) | 0.90-1.99 V (1 mV) |
| VTT Voltage | 1.05-1.70 V (6.25 mV) | 0.90-1.55 V (1 mV) |
| X79 PCH Voltage | 0.80-1.60 V (6.25 mV) | 0.71-2.59 V (40 mV) |
| DRAM Voltage | 1.20-2.1 V (5 mV) | 0.90-1.99 V (1 mV) |
| CAS Latency | 3-15 Cycles | 3-15 Cycles |
| tRCD | 4-15 Cycles | 3-15 Cycles |
| tRP | 4-15 Cycles | 3-15 Cycles |
| tRAS | 4-40 Cycles | 9-63 Cycles |
While we attempted to use identical base clocks for our overclocking benchmarks, those weren’t necessarily the absolute overclocking limits of this processor. We found that by adding just a few hundred kilohertz to the base clock, the Rampage IV Extreme was capable of pushing an extra 67 MHz behind its highest 100 MHz BCLK result.

EVGA gained nothing, tolerating not even the tiniest increase beyond the 4.67 GHz previously found in our 133.3 MHz base clock attempts. The same stability limit could be reached at 101.5 MHz base clock by using a 46x multiplier.

Our processor doesn’t work well at the chipset's 1.66x strap, with Asus’ 153 MHz max representing an underclock of some components. Anything less than 153 MHz was too much of an underclock for those components (likely the PCIe and/or SATA controllers), and the X79 FTW couldn’t successfully fit the processor within the narrow frequency range that allowed this strap.

Experienced hands are required to reach DDR3-2133 on the X79 FTW, since its default timings weren’t stable using XMP values. We couldn’t use its 21.33x multiplier anyway, and this overclock required both manual timing configuration and a 133 MHz base clock.
- X79's Last Hurrah Before Ivy Bridge
- Asus Rampage IV Extreme
- Rampage IV Extreme Firmware Overclocking
- Rampage IV Extreme Software Overclocking
- EVGA X79 FTW
- X79 FTW Firmware Overclocking
- X79 FTW Software Overclocking
- Overclocking Stability Compared
- Breaking Boundaries
- Test Setup And Benchmarks
- Power And Heat
- Benchmark Results: SiSoftware Sandra
- Benchmark Results: DiRT 3
- Benchmark Results: Metro 2033
- Benchmark Results: The Elder Scrolls V: Skyrim
- Benchmark Results: Audio And Video Encoding
- Benchmark Results: Productivity
- Asus Versus EVGA: Who Wins?
Where is the MSI Big Bang Xpower II? That's known to be a great board for overclocking as well.
Just knowing the ASUS and EVGA from past history (LGA 1366) which may or may not play any role here, ASUS tends to (lets call it adjust) the CPU vCore and VTT/VCCSA higher than advertised vs EVGA which probably explains* the problems with both the high frequency RAM and CPU OC's.
The disturbing thing to 'me' was the regulator voltages. I know the EVGA uses 12+2 vs ASUS's 8+3+(2+2) PWM and it's all digital controlled on ASUS, (*)but IDK if the EVGA is digitally controlled which might explain the inefficiency and OC.
The EVGA has always been a very 'manual' MOBO, so in that regard I'm not surprised you had to dive into the BIOS. I have no doubts if you raised the EVGA's voltages vs a cloned ASUS optimized OC set that you'd have no problems obtaining the SAME 4.8GHz OC. Both boards offer voltage check points and I'd be very interesting how they compared.
Just the other day I updated my ASUS BIOS and right-off I noticed an increased vCore increase by +0.01v~+0.015v and as part of the documented (improvements) was 'Improved Stability' ; yeah sure if you raise the vCore or VTT/VCCSA, phase, etc profiles... Now I have to redo my validations.
The most important testing here, to me, is the Baseline Comparison which tells me EVGA has some work to do ASAP, and hopefully a BIOS update can close the gaps. Further, personally I won't buy or recommend any X79 MOBO unless it offers an 8xDIMM option. In the forum it's been very clear which X79's I recommend since day one, and the ASUS R4E has always been on top on my list if you can afford it!
Thanks for the Article! Nice couple of boards indeed.
Cheers!
If I was into extreme system building I would not hesitate for a moment for a $50 price difference between a board that works as advertised vs a board you have to coax into working properly. In the grand scheme of things $50 is not that much money, especially considering the platform cost of the rest of the system. It's not like you build these rigs for OC'd gaming with less than $1000 in GPU horsepower alone, not to mention nice big SSDs and RAID arrays, $100 coolers, oversized cases etc. $50 just disappears at that point.
The ironic thing to me however is that this proves more than anything that OCing a system has little to no effect on gaming (at least with a single GPU), and yet gamers are the ones more likely to OC as most productivity people are artists or cube dwellers by nature and not hardware junkies who would OC their system to the moon. Yet there is a near 1/3rd of untapped performance potential in their computers that will never be touched.
Another thing of interest to me is that on the 1155 platform it is possible to get a higher efficiency by doing an OC (because you can get to 4.4GHz on some before changing any voltages which only gives a minor wattage increase). I wonder of the 2011 platform gets a little more efficient with a minor OC compared to these higher OCs.
Last thought: It is odd to me how OCing works. Something like 90% hit 4.2GHz with no problem, only some 50% of the CPUs can hit 4.5GHz, and then only ~5% can hit anything above 4.8GHz. On my own rig I can hit 4.2GHz with no problems at all, and that is a 2600 non-K (turbo OC as I am locked out of the base clock, but according to CPUz it hits a consistent 4.2GHz when under load, which is the only time I need anything above stock anyways). Back when I use to OC a lot (way back in the P3/P4 days) you would be lucky to hit a 500MHz boost before having heat problems (much less stability issues). It just amazes me that you can easily hit a near 1GHz OC these days without any voltage changes, but then going much higher than that it requires exponential amounts of power. And now the thing holding OCers back is more timing stability and power regulation instead of heat.
Perhaps Tom's could do a modern 'extreme OC' and see how far you can get. 6GHz maybe?
Why you asking? Did you send one in to be included in the test?
Because in the article they wrote "We contacted that manufacturer (along with one of its closest competitors) to see how two of today’s top-rated enthusiast-oriented boards would compare to each other in terms of overclocking ease, stability, and features.", meaning they never contacted MSI(I'm assuming). I'm pretty sure MSI would be eager to put their flagship board to the test as well.
Hope I clarify something here and I never meant to be rude in the first comment.
I'm jealous. I don't need a Rampage IV. I don't. What I have is fine. Yeahp. You know, A Rampage IV and a 3820 wouldn't be that expensive. I could sell my Asrock Z68 Pro and 2500K... I don't need a Rampage IV...won't make a difference. Its just for light gaming. I have another machine for serious stuff. I...I....
Buy your board based on:
1. Reputation
2. Reviews
3. Warranty
4. Price
5. Options
I also opt to use integrated chips for NIC and Audio made from anyone other than Realtek!