(132-LF-E657-KR)
EVGA needs no introduction. They have now been around for more than a decade. Since the release of the X58 chipset, they have separated themselves as a top-notch enthusiast motherboard manufacturer. As of this writing, their X58 Classified holds two out of the top five world frequency records for i7’s in the CPU-z Hall of Fame. EVGA hopes to continue their dominance with their new lineup of boards based on the Intel P55 chipset.
Today, we are presented with the EVGA P55 FTW motherboard. This is one of quite a few motherboards EVGA has released for the P55 chipset, and is in the top three in terms of features.
| CPU Support | Intel Socket 1156 Processors |
| PCH | Intel P55 Express (PCH stands for Platform Controller Hub) |
| Memory | 4 x 240-pin DIMM sockets, dual channel, max 16GB DDR3-2600MHz+ (overclocked) |
| Graphics | Three PCIe graphics expansion slots. Supports a single 16x PCIe GPU + dedicated PhysX card operating at 4x -or- two nVidia cards in SLI at 8x/each card + dedicated PhysX card at 4x -or- two ATI cards in crossfire at 8x/each card. |
| Storage | 8 SATA II ports, 6 internal and 2 eSATA with support for RAID 0, RAID1, RAID 0+1, RAID5, JBOD |
| I/O | 1 P/S2 Keyboard; 13 USB2.0, 7 on the rear panel, 6 onboard; 2 IEE1394 Firewire, one rear panel, one onboard |
| Network | Two gigabit ethernet ports (10/100/1000) |
| Audio | 8-channel HD Audio with SPDIF & COAX. |
| Power | 12+2 Phase Analog CPU Power, 3 Phase Memory Power |
Now we’ll have a look at the EVGA-specific features, as stated by EVGA’s specs.
- 12+2 Phase PWM with switching frequency up to 1,189KHz – Superior power delivery. (This also goes along with their two 8-pin CPU power sockets.)
- EVGA E-LEET Tuning Utility – Enthusiast software for adjusting overclocking in O.S.
- Passive chipset heatsink – No fans, lower noise, longer lifespan.
- EVGA ECP V2 Support – Overclock and debug your PC on an external adapter!
- EVGA Double Play Heatsink – CPU mounting supports both socket 775 and socket 1156 heatsinks.
- EVGA EZ Voltage read points – Easily read voltage levels with dedicated read points.
- Triple BIOS Support – Always have a backup, compare BIOS versions or use 3 separate profiles!
- PCIe Disable Jumpers – Quickly and easily troubleshoot!
- EVGA Vdroop Control – Improves overclocking stability.
- Onboard Clear CMOS, Power and Reset Buttons – System essentials at the touch of a button.
- 100% Solid State Capacitors – Higher quality, longer lifespan.
- EVGA EVbot Support – Next generation handheld overclocking device from EVGA.
- EVGA dummy OC – Overclocking at the touch of a button.
- Onboard CPU Temperature Monitor – Monitor your CPU temps quickly and easily.
- Dual Clock Generators (CPU and PCIe) – Ensure best signals for CPU overclocking
Whew, that is quite a list! Based on all of that, this board already looks like a winner. We’ll see how it holds up in our testing, but they have definitely set up some high expectations.
Before diving into the motherboard however, we know this may be some of your first experience with the combination of Intel’s P55 chipset and i5/i7 LGA1156 processors. Let’s take a minute to explore what that means for you.
The release of Intel’s P55 chipset and socket LGA1156 processors are bringing them back into the mainstream. The combination of X58 and i7 is a total powerhouse, unparalleled in sheer processing power. Unfortunately, it is a bit pricey (to put it mildly). Enter P55. Ranging from the ‘much less expensive than X58′ end of the spectrum ($100 boards + $200 CPU) to the ‘as expensive as low-to-mid range X58′ ($250+ boards + $500 CPU) end, you can go as high or low as you want with P55. Built for the “average” user, this is what may just take LGA775 to its end-of-life point. There certainly won’t be any more LGA775 CPUs, that’s for sure!
So, what does P55 change? For starters, there is no more southbridge with a P55 motherboard. The P55 chipset is called a “Platform Controller Hub”, or PCH. It does everything a southbridge used to do and then some. Graphics PCI-e lanes and the memory controller have all moved to the i5/i7 LGA1156 CPUs. The rest is handled by the P55 chip. Here is Intel’s diagram for the combo (click to make it a little larger if you have trouble reading it).
So what are the main differences between X58 and P55?
- For starters, there are fewer PCIe lanes available for graphics cards. With the X58 chip controlling graphics, motherboards had 36 PCIe lanes available to use, enabling two full-bandwidth 16x PCIe slots (three or more if you add nVidia’s NF200 chip). That is no longer the case with P55. The on-chip PCIe controller has 16 PCIe lanes available. For Crossfire and if your board supports SLI (IMPORTANT – Not all P55 boards support SLI, check your specs carefully!), you will have two PCIe slots available to use, but they will be operating at 8x each. Adding the NF200 chip to this combination will enable additional PCIe slots, but each at 8x.
- The P55 chip itself has some PCIe lanes available on it, connected through the DMI (Direct Media Interface). Some manufacturers (EVGA in particular) are using those lanes creatively. EVGA has decided to use them to allow the use of dedicated PhysX GPU running at 4x.
- On the P55′s platform, the on-CPU memory controller operates in dual channel, as opposed to the X58 platform’s triple channel.
That’s all well and good. Yay for X58. However, while these changes look big on paper, how much do they really affect performance? Not as much as one might think.
Let’s start with the reduced number of PCIe lanes. There are several comparisons out there now comparing dual and triple GPU combinations clock-for-clock. One of my favorite happens to come from the eVGA forums. It’s a quick-hit Tri-SLI comparison between an X58/i7 920 and a P55/i7 860. Both of the CPUs were overclocked to 4GHz and both had triple GTX285′s. The tests show very minimal difference between the two setups. So, as far as graphics are concerned, unless you are a hardcore benchmarking enthusiast, you will probably never be able to tell the difference.
Moving on, let’s have a look at dual-channel vs. triple-channel. Drop over to InsideHW and have a look at this comparison if you have a moment. It compares dual and triple channel on the same X58 motherboard (they run both) and setup. To save yourself some time: the difference is negligible. You’ll never notice in every day use. Again, hardcore benchmarking may show a slight deviation, but even that is very small.
There are others, but those are the main differences between the platforms. Long story short – there is a performance hit, but every day users will likely never notice it and even hardcore benchers will still find it a thrilling platform.
Enough about the new technology though…let’s move on to the matter at hand!
The board comes very well packaged. All accessories except the I/O plate, manual & driver CD come in antistatic packaging. The board itself also comes shrouded in antistatic protection with padding underneath. Included are four SATA cables, two 4-pin MOLEX to SATA power adapters, an SLI bridge, the EVGA control panel and its connectors, a driver cd and the manual.
 |
 |
 |
 |
Let’s have a look around the board. The colors are a muted but elegant black and gray. The last photo is after the stickers were removed and the water-block retention bolts and CPU have been installed.
 |
 |
 |
 |
The power section of the board is very stout. Its 12+2 phase CPU power design is plenty to push your i5 or i7 CPU as far as it will go. The heatsinks are installed with screws instead of spring-loaded push pins, which is great. Even with the screws though, contact was not perfect. Compared to others we’ve seen, it was not bad. The smaller heatsink on the top was perfect. The longer heatsink on the side had better contact toward the edges and wasn’t quite as good in the middle. On the plus side, there was contact with every one of the MOSFETs. They chose to stick with the usual thermal pad solution like pretty much everyone. As usual, it was promptly replaced with quality thermal interface material; Arctic Cooling MX-2 in this case.
 |
 |
 |
 |
The P55 heatsink is thin to allow video card clearance but large enough that it seems like it could dissipate a fair amount of heat. It too is mounted with screws, though these are spring-loaded to give the proper pressure on the P55 chip. Underneath, they did choose to go with thermal paste instead of a less effective thermal pad. Unfortunately, like most manufacturers that use it, they have globed a massive amount of TIM on there. From its shape, it is obvious that this is a manufacturing time saver. They appear to put a square “sticker”, for lack of a better word, of TIM and let the heatsink pressure push what isn’t needed out to the side. While understandable in light of the time saved, it would be nice to see a bit more care taken in TIM application on important chips; not just by EVGA, but by the vast majority of manufacturers. It too got a nice cleaning and some MX-2 applied in its place.
 |
 |
The rear I/O panel is loaded with a lot of connectivity. There are no less than six USB ports, a PS/2 keyboard port, one each of firewire, eSATA and USB+eSATA ports, dual LAN ports, two SPDIF connectors (one optical, one coax), and the standard HD audio 6-plug interface.
Moving on around the board, we see that it was designed with overclocking in mind. There are some nice features offered for a broad range of overclockers, from the extremely hardcore to the average joe. There are onboard power, reset & clear CMOS buttons. To their left, you will notice a BIOS selector switch. There are three BIOS chips on this board. Bad flash? Choose another BIOS, then hot-flash the bad one. Need more OC profiles? Use another BIOS. Care to test more than one BIOS to see which works best for you? Go for it! This is an excellent solution to aid people who play with their BIOS a lot and enjoy testing the latest to see if it clocks any better or has additions they need.
There is one bad point in this photograph we need to touch on. Toward the upper left above the bank of four capacitors, you will see the header to connect front-panel audio. The location of this makes it a wire management nightmare. A lot of front panel audio cables can barely reach the header they need to go to (which is often on the very bottom of a board) and remain hidden to begin with. With this placement, a bad situation just got worse. So one ding in the FTW’s armor; wire management freaks (this reviewer is among them), say goodbye to front panel audio.
On the lower right section of the board resides a POST display that shows the POST diagnostic codes should something go awry as well as the CPU temperature once the system is fully booted up. Below the POST display we have the headers for additional firewire and USB ports as well as power/reset/LED case connectors.
To the right of the POST display are the connectors for the EVGA Control Panel (ECP). This is a handy little gadget that serves basically as an extension cord for a POST display, another set of power/reset/clear CMOS buttons and PCIe disable jumpers.
The PCIe disabled jumpers are a nice touch as well. Should you have a GPU go bad, these allow you to disable one of the cards without the need to physically remove it; a definite plus for anyone water cooling their GPU(s)).
The ECP also has three buttons to increase voltages; two that increase Vcore 0.1V (when both are pressed, it increases Vcore by 0.2V) and one that increases VTT by 0.1v. On-the-fly hardware voltage increases…nice touch EVGA!
 |
 |
Last, but not least in our tour of the outer rim are the voltage read points. This is extremely handy for hardcore overclockers. Most OC’ers know that software has the potential of mis-reading voltages. Enter pre-installed, easy to use voltage read points. No more hunting down the tiny places on your board to read the actual voltages. More importantly, no more dead boards from touching your multimeter lead to the wrong item. You can quickly and easily read actual values for Vcore, Vdimm, PCH, CPU_PLL and VTT. These are located just above the DIMMs on the upper right hand of the board.
The BIOS of this board is well laid out and easy to understand.  That’s not to say there aren’t a lot of options to keep overclockers busy; just that it’s user friendly. Since most of the non-overclocking options (boot order, time settings, chipset features, etc.) are very similar on most boards, we’ll skip straight to the heart of the matter. On the right of the home screen you’ll find the “Frequency/Voltage Control” menu. It contains all overclocking options. It also has the RAM & CPU configuration options.
 |
 |
 |
 |
 |
 |
 |
 |
Remember that CPU power section we spoke about earlier? It allows you to take your CPU to an astounding 2.3V!
As a matter of fact, there are broad voltage ranges for all of the settings. The granularity is excellent as well (especially for Vcore) so you can set the voltage you really need without overkill. Here are the voltages available.
| Vcore | 1.00000-2.30000v |
| Vdimm | 1.20v-2.60v |
| CPU VTT | 1.050v-2.000v |
| CPU PLL | 1.050-2.400v |
| PCH | 1.050-2.625 |
One setting stood out very strongly in testing: EVGA Vdroop Control. It’s not just without Vdroop, it’s the polar opposite of “with Vdroop”. When testing 4.0GHz and 1.32500Vcore set in BIOS, the CPU idled at 1.341v and loaded at 1.352v.  By contrast, at the same Vcore set in BIOS and “With Vdroop” enabled, the CPU idled at 1.304v. It was an experiment that resulted in being unable to see the loaded Vcore, as prime crashed the PC immediately, but you can clearly see a large difference between the two settings. From seeing several of these boards floating about on various forums, the Vdroop is negligible to begin with.
Speaking of BIOS, there was one very interesting bug discovered when trying the newer BIOS (A29). When running stability tests with Prime95, the blend test caused the hard disk to go berserk. Absolutely bonkers. It just worked and worked. We’ve no idea what they did, but Priming the CPU results in the HDD reading (we assume, and hope just reading) an incredible amount. That may be one to stay away from. There are other reports of different issues on EVGA’s forums as well. Let’s say it makes one glad there are three BIOSes to use.
Overclocking this board is relatively painless to a point. 3.6GHz was where we started; it was a very simple overclock and was perfectly stable through an hour of Prime95 blend test. An hour doesn’t deem an overclock 100% stable, but it’s a good starting point. These were the settings for 3.6GHz:
Frequency/Voltage Control
- CPU Multiplier Setting: 20
- CPU Frequency Setting: 180
- PCIe Frequency Setting: Auto
- QPI Frequency Setting: Auto
- MCH Strap: Auto
- Extreme Cooling: Auto
- EVGA Vdroop Control: Without Vdroop
- Bootup CPU Vcore: 1.25000v
- Eventual CPU Vcore: 1.25000v
- DIMM Voltage: 1.65v
- Bootup CPU VTT: Auto
- Eventual CPU VTT: Auto
- PCH: Auto
- CPU PLL: Auto
- DIMM 1/2 DQ Vref: +0mV
- DIMM 3/4 DQ Vref: +0mV
- DIMM 1/2 CA Vref: +0mV
- DIMM 3/4 CA Vref: +0mV
- CPU PWM Frequency: 477KHz
- VTT PWM Frequency: 240KHz
- DDR PWM Frequency: 240KHz
- CPU Signal 1: Auto
- CPU Signal 3: Auto
- CPU Signal 4: Auto
CPU Configuration
- Hardware Prefetcher: Enabled
- Adjacent Cache Line Prefetch: Enabled
- MPS and ACPI MADT ordering: Modern Ordering
- Max CPUID Value Limit: Disabled
- Intel (R) Virtualization Tech: Enabled
- Execute-Disable Bit Capability: Enabled
- Intel (R) HT Technology: Enabled
- Active Processor Cores: All
- A20M: Disabled
- C1E Support: Disabled
- Intel (R) SpeedStep(tm) Tech: Disabled
- Intel (R) TurboMode Tech: Disabled
- Intel (R) C-STATE Tech: Disabled
- C3 State: Disabled
- C6 State: Disabled
- C State package limit setting: Auto
- C1 Auto Demotion: Disabled
- C3 Auto Demotion: Disabled
Memory Configure
- DRAM SPD: Standard
- DRAM Frequency: Auto
- DRAM tCL: 9
- DRAM tRCD: 9
- DRAM tRP: 9
- DRAM tRAS: 24
- DRAM tRFC: Auto
- DRAM tRRD: Auto
- DRAM tFAW: Auto
- DRAM tRTP: Auto
- DRAM tWR: Auto
- DRAM tWTR: Auto
- Command Rate: Auto
- CHA Round Trip Latency: Auto
- CHB Round Trip Latency: Auto
- Memory Low Gap: Auto
- Memory Remap Feature: Enabled
- Memory Hole: Disabled
- DRAM Margin Ranks: Disabled
4.0GHz, however, was much, much more difficult. Six hours of attempting to stabilize a 4.0GHz overclock was fruitless. We can tell you that just throwing voltages at these chips does not work. We went as far as 1.40Vcore (loaded), 1.45v CPU VTT and 1.9v CPU PLL. It would simply not stabilize. For what it’s worth, we did attempt the default multi of 21 with a lower BCLK to no avail. Time being an issue, we have chosen to leave stability alone for now.
However, we also never say die. Follow in the forums for updates on overclocking this board & CPU combo. Seeing various reports on 860′s, some are good clockers and some are absolute bears. It could be that we received one of the latter. Only time will tell and there isn’t enough time right now to warrant delaying the review any longer.
Thankfully, we were able to bench both 4.0GHz and 4.2GHz. The only exception was the Everest FPU Julia test at 4.2GHz, which was unable to complete due to (you guessed it) instability.
Bear in mind that there are a plethora of overclocking tweaks besides voltages on this board. If past experience has taught us anything, it’s that you can throw voltages at a chip all you want, but if something else is awry, nothing will work to increase your overclock. We’ve just scratched the surface on the overclocking options available. Again, stay tuned in the forums; hopefully we’ll all be pleasantly surprised!
So, let’s have a look at the numbers, shall we?
While this chip has a default multiplier of 21, we chose to run ours at 20 for the overclocked tests. There is a method to our madness. The i7 920 overclocked in previous reviews was clocked with a multiplier of 20. As such, not only can you see how this board and CPU perform, but you can directly compare the two platforms to see if X58 really is worth the increased premium to you!
We will also compare results of the two systems at stock settings. Bear in mind when seeing these that the i7 860 has a higher default clock speed (2.8GHz) and a stronger Turbo setting than the i7 920. So the results may be skewed in its favor. As more of a direct comparison, look at the overclocked benches. We’ll get to that later though. Before we do, let’s see how this combination performs by itself.
The system running these tests consists of the following:
- EVGA P55 FTW
- i7 860
- 2x2G G.Skill Trident (rated at DDR3-2000, 9-9-9-24)
- EVGA 8800GTX
- Corsair TX650
The cooling is a custom water loop, consisting of (and flowing in this order:
- Swiftech MCP355 Pump with an XSPC Reservoir Top
- Swiftech MCR-320 Radiator
- D-Tek Fuzion v.2 CPU Water Block (Many, many kudos to EVGA for being the only manufacturer to include socket 775 holes!)
- Swiftech MCW-60 GPU Water Block
- Back to the Reservoir
For your viewing pleasure, we took a couple photos of the test system.
 |
 |
First up, the overclocker’s favorite: SuperPi Mod 1.5 1M run. Sub-10s with only 4.2GHz; excellent!
Super Pi 1M
| Overclock | Score |
Screenshot |
| Stock | 14.265s |  |
| 3.6GHz | 11.641s |  |
| 4.0GHz | 10.484s |  |
| 4.2GHz | 9.969s |  |
Next up we have the battery of Everest tests. These bench your computer and then compare it to their database of results. It is important to note that (from appearances), the comparison results are all run at stock clocks.
The first four tests in the Everest suite are memory tests. These results were very strong for the “lesser” dual-channel controller. Starting in 2nd more often than not, every test ended up in first with only the lower overclock, remaining at the top for the remainder.
Memory Read
Results are in MB/second, higher is better.
| Overclock | Score | Screenshot |
| Stock | 12021 |  |
| 3.6GHz | 15680 |  |
| 4.0GHz | 17379 |  |
| 4.2GHz | 17617 |  |
Memory Write
Results are in MB/second, higher is better.
| Overclock | Score | Screenshot |
| Stock | 11663 | |
| 3.6GHz | 14213 |  |
| 4.0GHz | 15791 |  |
| 4.2GHz | 16576 |  |
Memory Copy
Results are in MB/second, higher is better.
| Overclock | Score | Screenshot |
| Stock | 13918 |  |
| 3.6GHz | 19720 |  |
| 4.0GHz | 20090 |  |
| 4.2GHz | 22048 |  |
Memory Latency
Results are in nanoseconds (ns), lower is better.
| Overclock | Score | Screenshot |
| Stock | 59.4 |  |
| 3.6GHz | 45.9 |  |
| 4.0GHz | 41.7 |  |
| 4.2GHz | 39.9 |  |
Now we’ll have a look at the Everest CPU tests. Much like the memory tests, this system performed at or near the top in all of the benches run. In Queen and Photoworxx, the system launched to the top (from starting very close) when overclocked. Zlib results started a little lower and finished in a healthy second place. AES was the exception. There were several systems that this CPU just couldn’t best, even at the 4.2GHz level.
CPU Queen
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 24359 |  |
| 3.6GHz | 29958 |  |
| 4.0GHz | 33279 |  |
| 4.2GHz | 34957 |  |
CPU Photoworxx
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 29028 |  |
| 3.6GHz | 36938 |  |
| 4.0GHz | 41172 |  |
| 4.2GHz | 43495 |  |
CPU Zlib
Results are in KB/s, higher is better.
| Overclock | Score | Screenshot |
| Stock | 92702 |  |
| 3.6GHz | 114051 |  |
| 4.0GHz | 126752 |  |
| 4.2GHz | 133665 |  |
CPU AES
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 23157 |  |
| 3.6GHz | 28518 |  |
| 4.0GHz | 31661 |  |
| 4.2GHz | 33271 |  |
Finally we have the Everest FPU tests. These are more stressful than the previous tests. As such, since we were having stability problems, one of the tests failed at 4.2GHz – the FPU Julia test. All others were able to run as normal. In these, the Julia and Mandel tests never quite got to the top of the heap, though a strong second is nothing to sneeze at. SinJulia was a surprise here. Starting in third place at stock, it catapulted to the top when overclocked and kept going from there.
FPU Julia
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 12185 |  |
| 3.6GHz | 14975 |  |
| 4.0GHz | 16652 |  |
| 4.2GHz | n/a | n/a |
FPU Mandel
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 6603 |  |
| 3.6GHz | 8119 |  |
| 4.0GHz | 9022 |  |
| 4.2GHz | 9474 |  |
FPU SinJulia
Results are scored, higher is better.
| Overclock | Score | Screenshot |
| Stock | 5454 |  |
| 3.6GHz | 6706 |  |
| 4.0GHz | 7450 |  |
| 4.2GHz | 7829 |  |
So, the pressing question now is: How do the two platforms compare to each other? From the looks of it, very admirably! The results in this review were compared against the results from Lvcoyote’s review of the EVGA X58 Classified. In that review, like this one, the tests were run with a multiplier of 20 and CPU: DRAM ratio of 2:8 with timings of 9-9-9-24. Everything was as equalized as possible to allow for a fair comparison.
When viewing these graphs, there are a couple things to keep in the back of your mind. First, as mentioned before, all stock clocks are not created equal here. The i7 860 starts with a higher stock clock (2.8GHz compared to 2.66GHz), and since it was true stock testing, all features like Turbo were turned on. The overclocked tests were run with HT on and nothing else, just like the i7 920 / X58 tests.
Second, you should remember that these are just benchmarks. SuperPi can vary by a few hundredths of a second with every run. Everest can vary as much as a few hundred points/KB/MB with every run. Neither review ran them repeatedly to try and get the best score; they were run once and recorded. So take the individual scores with a small grain of salt.
However, there are eleven Everest tests all told, with four tests run in each. Considering both systems had the potential for score anomalies, with the number of tests run it is safe to assume that there are equal numbers of high and low anomalies. Therefore, as a whole they should be a good indication of how the two platforms compare.
One last thing before we get started; due to the time they were published, the Classified benches were run on Vista and these were run on Windows 7. That does have the potential to have affected these benches. We’d like to think it has a negligible effect, but it’s worthy of pointing out.
The benches will go in the same order as displayed above. Results from the system in this review are displayed in blue and from Lvcoyote’s review in red. To start, we have SuperPi1M. As you can see, the 860 benefited from higher stock clocks. After that, the 920 consistently beat its time in all three subsequent overclocks. Score one for the higher-end platform.
Everest memory benchmarks. These tests were far and away the biggest surprise. The 860/P55 platform simply trounced the 920/X58. The latter is supposed to be much stronger on memory because of the benefit from tri-channel RAM. This most certainly did not play out in our testing. In one out of the sixteen tests run, the 920 scores highest (memory copy, 4.0GHz) and one additional test was close (memory read, 4.2GHz). So in 15/16 tests, the 860/P55 combo beat out the 920, often by a substantial margin. Point: 860/P55
The CPU tests were more along the lines of expected performance. A high-end targeted platform should out-score a mainstream targeted platform. The 860 barely edged out the 920 in Queen; likewise the 920 barely edged out the 860 in AES. Photoworxx and Zlib were the sole property of the 920, even at stock clocks. Point: 920/X58
The FPU tests were really, really close. The 920 won out in Julia, but the 860 eked out a win with the Mandel and SinJulia tests. Point: 860/P55.
So, the question one might ask: Is the X58 platform worth the extra premium you’ll pay for it? The answer is an annoying “it depends on what you’re going for.” If you do not overclock and want to run a single GPU, that one is a simple, unequivocal no. If you overclock and run a single GPU, we’d still go with a no; not necessarily because the X58 platform doesn’t perform better, because in many cases it does. We just feel the benefit an average user will see is not worth the premium you have to pay.
The only place the X58 really, truly, in all cases stands out over and above the P55 platform is with multiple very high bandwidth graphics cards. If you want to run SLI/Crossfire (or tri-SLI / tri-fire for that matter), you’ll want to stick with X58. The P55 platform just can’t compete with two or three full x16 PCIe lanes.
So what about the board? There is no doubt about it, the EVGA P55 FTW is a superb motherboard. At $229.99 (£142.91), it is priced toward the higher-end of the P55 spectrum. As of this date, Newegg has three ASUS and a Gigabyte board with higher prices. This is to be expected. The features, accessories (EVGA Control Panel especially) and to a lesser degree the appearance of this board make it well worth the investment in our opinion.
Overclocking was a mixed bag with this board and CPU. Unfortunately, we have no other P55 boards handy to compare right now to see if the CPU is just not a good clocker or if the board is contributing to the inability for higher stable overclocks. We hope to revisit this in further reviews; if justified, the award may be adjusted. Instinct tells us it is the CPU and without proof otherwise, this will not detract from our otherwise very positive opinion of the board.
| Pros | Cons |
|
|
Here at Gilgamesh Reviews we have a three tier rating system. A Silver Award, an Editors’ Choice Award, and our top honor the Gold Award. To achieve the Gold Award a product must demonstrate a performance level above and beyond the normal. This exquisite offering from EVGA does just that. While it is priced at the higher end of the P55 spectrum, its featureset and specifications more than justify that high price, especially when compared to other manufacturers’ options at (and above) this price point. Therefore, we are very pleased to bestow another Gold Award upon EVGA for their efforts!
Quesitions, Comments, Concerns? Feel free to share in our forums!
Posted in 
