Gigabyte GA-P55A-UD7 LGA 1156 Motherboard Review

[lemonlime]

Included Software

Included Software

As expected, Gigabyte includes a very generous software bundle with the P55A-UD7.

The driver and utility CD is about what we’d expect from any major motherboard manufacturer. A slew of drivers and utilities can be downloaded, although we’d definitely recommend taking a look online in case newer revisions are available.

Easy Tune 6

Easy Tune 6 is an application that we’ve become familiar with over the past few Gigabyte board releases. This particular application is the sort of “technical centerpiece” of the software suite allowing for both viewing of system vitals as well as the application of both overclocking and other system tweaks.

Much like the very popular CPU-Z application, the CPU and Memory tabs of Easy Tune provide all of the essential system information.

The real fun happens at the “Tuner” tab. As you can see, there are several profiles to cater to both the novice and advanced overclocker. The “Quick Boost” feature is a handy one for those very uncomfortable messing around in the BIOS, or for those who probably shouldn’t be overclocking in the first place. Simply select your desired level of overclock, reboot and you are off to the races. It doesn’t get much easier than that.

Under the advanced tab, things get a lot more granular. Users can increase both frequencies as well as voltages from the operating system. Although this works great for the most part, we found ourselves encountering system instability when increasing the BCLK values from within Easy Tune. It wasn’t always consistent, but settings stable in the BIOS were not always stable when setting from Easy Tune.

Unfortunately, Easy Tune didn’t recognize our video drivers or hardware and video card frequency control was “grayed out”. Compatible drivers/hardware would allow control of core, memory and shader frequencies.

Smart fan control is another nifty feature found within Easy Tune. The fan speed can be adjusted based on selectable temperature thresholds.

Last but not least, we have the HWMonitor tab. As you probably have already guessed, voltages, fan speeds and temperatures can be monitored from here.


Dynamic Energy Saver 2

Next up is Gigabyte’s power saving application called “DES2” or “Dynamic Energy Saver 2”.

​

The application allows the control of numerous power saving features above and beyond what is provided by Microsoft and Intel. Of most notable interest is the ability to shut down unused power phases that are graphically depicted by the animated pistons turning away. Many other power saving features can be adjusted, such as fan speeds, sleep timers, base clock frequency reductions and more.

DES2 is disabled by default, and has to be enabled by clicking the large button at the top right hand corner of the board. Interestingly, the power phase LEDs on the board itself light up and begin to work as soon as this feature is activated. There is a definite hardware component to DES2, as we could see by the “Powered by Intersil” logo at the bottom of the application. More than one Intersil IC can be found around the power delivery area of the motherboard to make this functionality possible.

Given the interest in going “green” these days, we were very pleased to see Gigabyte’s innovative power saving features.

 

[lemonlime]

Included Software pg.2

Included Software pg.2

Face Wizard

Let’s face it. Most self-respecting system tweakers disable the “full screen logo” that appears during the boot-up process. But what if that logo could be changed into absolutely anything? That certainly changes things. Gigabyte’s “Face Wizard” application allows you customize this logo in any way you see fit. You can pull down the current image stored on the board, and upload a new one. It’s really as simple as that.


Auto Green

Moving on, we have a very unique application called “Auto Green”. Auto Green is a blue tooth enabled application that automatically puts the PC to sleep or wakes it up based on proximity to a blue tooth enabled cell phone. For example, simply coming home from work and walking near the PC causes it to spring to life. This is a really nifty feature in concept, but unfortunately, Gigabyte does not include a Bluetooth adapter to make this work. Given the relatively inexpensive cost of a Bluetooth adapter, and the high price tag associated with the P55A-UD7, it certainly would have been nice to see one bundled. We would have gladly traded a couple of SATA ports for one!


@BIOS

Given that the vast majority of PC users have never seen a PC BIOS, let alone flashed a new BIOS ROM, it was nice to see that Gigabyte created a software product for the masses. Flashing a new BIOS really couldn’t be easier. Simply click to download the latest BIOS, and then click to flash. The current BIOS can also be backed up to file for roll-back purposes.


Smart 6

Last but not least, we have Gigabyte’s “Smart 6” application.

As you’ve probably already guessed, “Smart 6” is comprised of six different applications bundled together. These applications do everything from improving boot time, to securing password lists to restricting system access hours. “Smart 6” is a useful software suite that we’d consider a nice bonus that comes bundled with numerous Gigabyte products.

We won’t go any further into this application as it was already covered in detail during our X58A-UD7 review recently. Gigabyte also has a section of their website dedicated to “Smart 6” if you’d like to learn more.

 

[lemonlime]

Test Setup and Methodology

Test Setup and Methodology

Intel Testing Configuration:

• Intel Core i7 860 Processor (2.8GHz, 8MB L3)
• Thermalright TRUE Black Heatsink with 200CFM Delta 120x120x38mm fan.
• Gigabyte GA-P55A-UD7 Motherboard (P55 + NF200)
• 2x2048MB OCZ Platinum “Low Voltage” DDR3, 1600MHz @ 7-7-7 rated
• 2x1024MB Crucial D9JNM DDR3, 1333MHz @ 9-9-9 rated (For memory overclocking tests)
• Western Digital 320GB SATA2 Hard Drive (WD3200AAKS, 7200RPM)
• EVGA GeForce GTS 250 SC for general/OC testing (G92 based, 512MB GDDR3)
• 2x 8800GTs for SLI testing (G92 based, 512MB GDDR3)
• Corsair TX750 Power Supply
• Pioneer Optical Drive

OS/Driver Configuration:

For our testing today, we’ll be using Windows 7 Ultimate x64 Edition. Although Vista and Windows 7 bring some interesting performance enhancing features to the table – like SuperFetch for example – these features can cause inconsistencies in benchmark results. Some reviewers prefer to 'train' their systems to make the most of these features, but we have disabled them to ensure a higher degree of predictability. Other 'scheduled' or unnecessary background tasks have also been disabled (as listed below).

OS: Microsoft Windows 7 Ultimate (64-bit), With all updates.

Nvidia Drivers: Nvidia 196.21 WHQL Drivers

OS Disabled Features: “SuperFetch”, “Windows Search” service, Windows Defender, Windows Sidebar, Scheduled Defragmentation, System Restore Services and UAC.

OS Performance Features Enabled: “High Performance” power management profile is enabled and C1E processor state is disabled in the BIOS. This prevents the system from reducing CPU frequency at the operating system level and also prevents HDD spin-down and other events that may impact benchmark results.

Without further ado, let’s see how these boards overclock!

 

[lemonlime]

Overclocking Results

Overclocking Results

To test the overclocking ability of the P55A-UD7, we looked for the maximum base clock frequency, the maximum CPU frequency and the maximum memory frequency achievable. Although the CPU itself generally is the limiting factor on the socket 1156 platforms, the motherboard still plays a very important role.

For our testing, we limited our CPU voltage to about 1.4V, and VTT/QPI voltage to 1.34V. Memory voltage was also limited to 1.7V.


Highest Stable BCLK Overclock

​

The P55A-UD7 allowed us to take our i7 860 to it’s maximum base clock speed of 204MHz. To achieve this, we set a VTT voltage of about 1.34V. We could get pretty close to 210MHz, but for complete stability, 204MHz was the limit.

We definitely don’t believe that this 204MHz is due to any limitations of the P55A-UD7 and that has everything to do with the i7 860 sample we’re using today.


Highest Stable CPU Overclock

​

With respect to maximum CPU overclock, we had little difficulty breaking well beyond the 4GHz barrier and settled for a final frequency of 4180MHz at 1.4V. We did enable the highest load line calibration to ensure a steady Vcore, which is why CPU-Z is reporting a tad over 1.4V. At idle, the Vcore is actually a little lower than at load in this configuration.

Higher overclocks were definitely possible with additional Vcore, and the P55A-UD7 was more than happy to provide it. At this frequency and voltage, we were simply pushing the limits of our air cooling solution. We were very pleased with the result.


Highest Stable Memory Overclock

​

To put the P55A-UD7’s memory overclocking abilities to the test, we used a pair of single-sided Micron D9JNM modules that were known to clock beyond DDR2000 at 9-9-9 timings. As you can see, we had no difficulty pushing these particular DIMMs to their limits at our self-imposed 1.7V limit. Both the board and IMC still had some headroom left as additional voltage and looser timings allowed for greater overclocking headroom. Our VTT/QPI voltage was sitting at 1.34V for this particular test, although it didn’t need to be quite so high.

 

[lemonlime]

Memory Benchmarks

Memory Benchmarks

Since we were successful in breaking 2000MHz on this test platform, we’ll show you the impact of tightening timings versus increasing frequency from a bandwidth and latency perspective.


Lavalys Everest Ultimate v5.02

Everest Ultimate is the most useful tool for any and all benchmarkers or overclockers. With the ability to pick up most voltage, temperature, and fan sensors on almost every motherboard available, Everest provides the ability to customize the outputs in a number of forms on your desktop. In addition to this, the memory benchmarking utility provides a useful tool of measuring the changes to your memory sub-system.

As you can see, high frequency is still king when it comes to memory bandwidth – especially read bandwidth. Interestingly, write bandwidth does not seem to be bottlenecked by the memory at all. This would likely only increase with further IMC frequency increases.

When it comes to latency, timings are a significant factor. As you can see, 2000MHz at 9-9-9 yields almost the same overall latency score as 1670MHz at 7-7-7.

 

[lemonlime]

Synthetic Benchmarks

Synthetic Benchmarks

To see just what kind of impact those wonderful overclocks and performance tweaks made, we ran through a small suite of synthetic benchmarks.


SuperPi Mod v1.5

When running the SuperPI 1M benchmark, we are calculating Pi to one million digits and timing the process. Obviously more CPU power helps in this intense calculation, but the memory sub-system also plays an important role, as does the operating system. Rather than timing the 32M process, which is frequently used for testing stability, we’ll stick with the 1M test that is more familiar to benchmarkers.

As expected, this single-threaded benchmark greatly favored the overclocked scenario. Thanks to the i7’s Turbo frequencies, a significant improvement was realized. You’ll recall that the CPU multiplier increases to it’s highest level when only one CPU core is utilized by an application.

On the flip side, turning off HyperThreading did nothing for this benchmark as the logical cores were not used in the first place.


PCMark Vantage X64

The latest iteration of the popular system benchmark is PCMark Vantage from the Futuremark crew. The PCMark series has always been a great way to either test specific areas of a system or to get a general over view of how your system is performing. For our results, we stuck with the TV and Movies test suite, which is taxing on the CPU and RAM.

This multi-threaded test does favor HyperThreading, but not by a very large margin. As you can see in our “Max OC” result, the 4.18GHz clock frequency had the biggest impact on performance.


Cinebench R10 64-Bit

Developed by MAXON, creators of Cinema 4D, Cinebench 10 is designed using the popular Cinema software and created to compare system performance in 3D Animation and Photo applications. There are two parts to the test; the first stresses only the primary CPU or Core, the second, makes use of up to 16 CPUs/Cores. Both are done rendering a realistic photo while utilizing various CPU-intensive features such as reflection, ambient occlusion, area lights and procedural shaders. For our testing today, we’ll be using the multi-core rendering option only.

Cinebench benefits greatly from multi-threading, but we were surprised to see such a huge performance boost from the high CPU and memory frequency in the overclocked scenario.


Lavalys Everest Ultimate v5.02 – Zlib Compression

Everest Ultimate is the most useful tool for any and all benchmarkers or overclockers. With the ability to pick up most voltage, temperature, and fan sensors on almost every motherboard available, Everest provides the ability to customize the outputs in a number of forms on your desktop. In addition to this, the memory benchmarking utility provides a useful tool of measuring the changes to your memory sub-system. To test processing horsepower, we ran the Zlib compression test, which is hard on both the CPU and memory subsystems.

Not surprisingly, this multi-threaded test yielded similar results to Cinebench. Again, the overclocked configuration came out on top by a large margin. HyperThreading seemed to play a very large role in the results as well, boosting the score by around one third compared to the disabled scenario.

 

[lemonlime]

Gaming Benchmarks

Gaming Benchmarks

We unfortunately didn’t have a matching pair of higher-end Nvidia cards to give SLI a run for it’s money, so we made due with a trusty pair of 8800GTs. We ran each benchmark with and without SLI as well as with and without the maximum CPU and memory overclock that we were able to achieve.


Crysis 1.2

We benchmarked Crysis using the “Crysis Benchmark Tool 1.05” and chose the “benchmark_gpu” timedemo to ensure that there was a heavy GPU load. Tests were conducted with all IQ settings at “High” and a resolution of 1680x1050.

“But can it play Crysis?” With a pair of 8800GTs, you bet it can. It is not a surprise that Crysis is GPU limited with a single card, but with SLI enabled, the CPU and memory frequency increase actually gets put to good use and increases frame rates marginally.


Lost Planet: Extreme Condition

Lost Planet is a graphically demanding game released by CAPCOM. For testing, we set resolution to 1600 x 1000 and IQ settings to “High” with 16X AF, 4X AA enabled. Both the “Snow” and “Cave” tests were looped three times, and the average framerate taken.

We see a very similar picture painted in Lost Planet, albeit with higher frame rates. With a single GPU, the card is not at all CPU bottlenecked, but with SLI enabled, the extra horsepower actually gets put to use.


World In Conflict

World in conflict is a strategy title from Sierra. For our tests, we set detail to “High” and resolution to 1680x1050. The internal benchmarking tool was utilized and done three times so that an average could be obtained.

WIC’s maximum frame rates benefit quite a bit from SLI as can be seen above, but average frame rates increased by only about 10FPS, while minimum frame rates didn’t increase at all. A very small increase in performance could be attributed to the overclocked CPU and memory in the SLI configuration.


Futuremark 3DMark06

The Futuremark 3DMark series has been a part of the backbone in computer and hardware reviews since its conception. The trend continues today as 3DMark06 provides consumers with a solid synthetic benchmark geared for performance and comparison in the 3D gaming realm. This remains one of the most sought after statistics, as well as an excellent tool for accurate CPU comparison, and it will undoubtedly be used for years to come.

Even in 3DMark06, we see the same similar trend of extra CPU horsepower being put to good use with SLI enabled. It is hard to believe that more than 5000 points can be attributed to CPU and memory overclocking. In the non-SLI scenario, that delta shrinks to only about 800 points.

 

[lemonlime]

Voltage and Temperature Testing

Voltage and Temperature Testing

Most people don’t realize that stable system power is the responsibility of more than just the PSU. A motherboard is essentially a complex power supply on its own, converting 12V and other supplied voltages to the required outputs for various system components. We’ll be validating the P55A-UD7’s “dialed in” voltages in the BIOS using a digital multimeter.

For this particular test, we left our CPU at 2.8GHz and memory at the default 1333MHz frequency so that they could be consistent across all voltages selected. We took measurements of both CPU Vcore and QPI/Vtt as well as DRAM voltage. For a grounding point, we used the fan header at the bottom-left hand side of the board. The highest CPU Vcore selection was tested with all three load line calibration options to see their effect.

All of the voltages appeared very stable and close to their BIOS set points. CPU Vcore consistently undervolted without any additional load line calibration applied. This isn’t really a problem, but it means that voltage has to be dialed higher than what is desired.

Enabling the first LLC setting improved Vdroop significantly over Intel’s specification, and still allowed for about 0.04V of droop that helps to ease the burden on VRM components. For those really interested in eliminating Vdroop to achieve very high overclocks, the second level of LLC actually increases Vcore slightly at load. With 24-phases for CPU power delivery, this won’t be a problem at all for the P55A-UD7.

Both QPI/Vtt and DRAM voltage were nice and accurate for the most part, with QPI/Vtt overvolting slightly once increased beyond 1.26V.

Overall, we were very pleased with the voltage accuracy and stability of the P55A-UD7.


Temperature Testing

Next up is the moment of truth for the Silent-Pipe module. To test it’s functionality, we ran our system at full load with and without it installed. We set out Vcore to 1.30V and QPI/Vtt to 1.20V for this test to introduce a little bit of extra heat into the cooling system and installed only one 8800GT in the third PCI-E slot.

Temperature readings were taken with a digital thermometer in three locations – the PCH heatsink, the NF200 mid-heatsink and the rear MOSFET heatsink. Prime95 Large FFTs across all eight threads was used to load the system and measurements were taken after 30 minutes so that temperatures had a chance to stabilize. Ambient temperature was 18°C.

Although the Silent-Pipe does provide some benefit to the cooling of the board, but it is unfortunately pretty insignificant. The largest delta we saw was only 1.8°C, and the smallest was only 0.5°C.

Any improvement in cooling is always welcome, but one must weigh the pros and cons of attaining this small improvement. For a 1.8°C decrease to the NF200 and PEX8608 heatsink, the massive Silent-Pipe module must be installed that not only blocks a PCI-E 1X slot, but poses a short-circuit risk to the first PCI-E 16X slot, and potential heatsink compatibility issues.

It is also important to note that we didn’t install a video card right next to the Silent-Pipe so that we could position our thermometer correctly. We’d question whether or not this 1.8°C delta would exist once a toasty video card PCB radiates heat into the Silent-Pipe fin array. With that said, our advice is to simply leave the Silent-Pipe module in the box, but if you insist, be sure to install your video card in the second PCI-E 16X slot.

Silent-Pipe module aside, we are very pleased to report that all of the heatsinks on the P55A-UD7 are very effective and that the temperatures we observed were very good for a passively cooled solution. The heatpipe system is also functional, as the Silent-Pipe module proved for us - cooling one part of the cooling system lowers temperatures to all heatsinks.

 

[lemonlime]

Conclusion

Conclusion

Beyond any doubt, the Gigabyte GA-P55A-UD7 is the Cadillac of P55 boards. There are so many good things to say about this board that we’re honestly not sure where to begin.

Let’s begin by summarizing some of the P55A-UD7’s strongest points – namely it’s next-generation connectivity features and it’s innovative ways around platform limitations. Both USB 3.0 and SATA 6Gbps connectivity are defining “next-gen” features that really set the P55A-UD7 apart. Although next-gen peripherals and storage devices are scarce today, buyers will be happy to know that their investment today won’t leave them stranded in the future.

Gigabyte doesn’t just slap these next-gen controllers onto the P55 and call it a day though. They were very careful to ensure that these bandwidth hungry controllers were provided PCI-E 2.0 lanes via their implementation of the PEX8608 PCI-E bridge controller. Without this bridge, the Marvell SATA 6Gbps controller would be completely bandwidth starved and essentially useless. Not only is the P55A-UD7 outfitted with next-gen connectivity features, but it’s also equipped to handle them effectively.

Gigabyte didn’t just stop there though; they also provided official Tri-SLI and 3-Way Crossfire-X support via their implementation of Nvidia’s NF200 PCI-E controller. Rather than being limited by Lynnfield’s sixteen PCI-E lanes, the NF200 allows up to four cards in an 8X configuration to be used without having to traverse the P55 PCH at all. This equates to fast and functional multi-card setups that are not severely bottlenecked like some lower-end 3-slot P55 boards on the market today.

The P55A-UD7 has also proven to be a very potent overclocker thanks to Gigabyte’s proven 24-phase power delivery system and well-designed BIOS. We had absolutely no difficulty pushing our i7 860 to it’s limits and gave our D9JNM memory a good run for it’s money. In all of the overclocking tests we conducted, the board never held us back.

We also couldn’t leave off without giving mention of Gigabyte’s very comprehensive suite of software utilities that go above and beyond what one would expect bundled with a motherboard.

Our gripes with the P55A-UD7 are few and far between, but there are two points are worth mentioning. First on the list is the Silent-Pipe cooling module. Although it provides some cooling benefit to the board, it was simply not significant enough to warrant having such a large cooler on the board that blocks a PCI-E slot and poses other clearance issues. Gigabyte could have easily made this device much more effective by mounting it closer to the source of heat. Until this happens, we’d simply recommend leaving it in the box.

Our second gripe is the omission of a Bluetooth adapter. Although we’d never normally look for something like this, Gigabyte’s nifty “AutoGreen” feature can’t be used without one. Given the high price tag of the board, we would have expected this to be included.

Although the P55A-UD7 isn’t cheap at over $300, we can say with confidence that you get what you pay for. You get an incredibly capable and feature packed board for your hard earned money, and we would not hesitate to recommend it to anyone looking for a top-end LGA1156 board. As such, we’re pleased to award the Gigabyte GA-P55A-UD7 Hardware Canucks’ “DAM GOOD” award.

Pros:

- USB 3.0 and SATA 6Gbps support
- Tri-SLI and 3-Way Crossfire-X Support
- NF200 + PLX8608 reduce P55 PCI-E limitations
- Convenient board layout
- Twelve SATA 3Gbps ports included!
- Excellent 24-phase power delivery system
- Solid voltage regulation with two levels of LLC
- Very functional passive cooling solution
- Very generous software suite
- Excellent overclocking features
- Three Year Warranty
- Great BIOS layout and Q-Flash
- Redundant BIOS chips

Cons:

- Silent-Pipe cooler too large and not very effective due to heatsink design
- Bluetooth adapter not included for bundled software suite to work

Gigabyte GA-P55A-UD7 Review Comment Thread

A special thanks to Gigabyte for providing us with this sample.​