Sign in with
Sign up | Sign in
Asus And DFI: Core i7 Micro-ATX Motherboards Compared
By ,
1. ATX Without The Waste

With support for full-sized components and up to four expansion cards, the Micro-ATX format has always been more than adequate for the majority of high-end builds. Yet while enthusiasts have typically cited inadequate quality or design as the primary reason for not considering this option, manufacturers have cited lack of demand as a reason for not putting their best efforts into a board this small. A few attempts by manufacturers to win loyalty among space-conscious enthusiasts have mostly been rejected by a market that maintained its traditional view of the former problems.

The popularity of Micro-ATX portable gaming enclosures is finally starting to break the cycle of negative assumptions as customers are forced to make a decision about what hardware to put inside. Current top products are undoubtedly as feature-laden as many of their full-sized counterparts, incorporating high-end devices and support for even the largest dual-slot graphics cards in CrossFire and SLI. Always a target of upper-range Micro-ATX motherboard sales, professional media and home theater enthusiasts may instead choose to load up to three media-centric devices in addition to a single-slot graphics card. With this much flexibility, confessing that they don’t actually need more expansion room could be the hardest problem for many builders.

Yet few of us will even consider smaller devices until we can see that they function as well in every respect as the larger parts they replace, so today we’ll compare these against the fastest of our full-ATX samples. Before we go into the details of that test, let’s take a closer look at the features we so eagerly endorsed.

Micro-ATX Core i7 Motherboard Features

 

Asus Rampage II Gene

DFI LANParty Jr
X58-T3H6

Northbridge

Intel X58 Express

Intel X58 Express

Southbridge

Intel ICH10R

Intel ICH10R

Voltage Regulator

Eight Phases

Six Phases

BIOS

0705 (04-09-2009)

217 (02-17-2009)

133.3 MHz Bclk

133.6 MHz (+0.20%)

133.0 (-0.25%)

Clock Generator

ICS 9LPRS918JKLF

ICS 9LPRS918JKLF

Internal Interfaces

PCIe 2.0 x16

2 (x16/x16)

2 (x16/x16)

PCIe x1/x4

0/1

0/1

Legacy PCI

1

1

USB 2.0

4 (8-ports)

3 (6-ports)

IEEE 1394

1

0

Serial Port

0

1

Parallel Port

0

0

Floppy

0

1

 Ultra ATA-133

1 (2-drives)

1 (2-drives)

SATA 3 Gb/s

7

6

4-Pin Fan

5

1

3-Pin Fan

0

5

FP-Audio

Yes

Yes

CD-Audio

Yes

Yes

S/PDIF I/O

Output Only

None

Power Button

Yes

Yes

Reset Button

Yes

Yes

CLR_CMOS Button

Jumper Only

Yes (by PWR+RST)

Diagnostics Panel

External Device Header

2-Character

I/O Panel Connectors

PS/2

1

2

USB 2.0

4

6

IEEE 1394

1

0

Network

1

1

eSATA

1

0

CLR_CMOS Button

Yes

By Jumper

Digital Audio Out

1

2

Digital Audio In

0

0

Analog Audio

6

6

Mass Storage Controllers

Chipset SATA

6x SATA 3.0 Gb/s

6x SATA 3.0 Gb/s

Chipset RAID Modes

0, 1, 5, 10

0, 1, 5, 10

Add-In SATA

JMB363 PCIe, 1x SATA 3.0 Gb/s, 1x eSATA 3.0 Gb/s

None

Add-In Ultra ATA

JMB363 PCIe

JMB368 PCIe

IEEE 1394

VT6315N PCIe, 2 x 400 Mb/s

None

Gigabit Ethernet

Primary LAN

RTL8111C PCIe

88E8053 PCIe

Secondary LAN

None

None

Audio

HD Audio Codec

AD2000B

ALC889

2. Asus Rampage II Gene

Features and Layout

Smaller, yet slightly updated compared to the Rampage II Extreme, the Micro-ATX Rampage II Gene lacks only a few slots, a few redundant circuits, and a secondary Gigabit network controller compared to its larger sibling.

Customers still get two full-bandwidth PCIe 2.0 x16 slots, plus full support for overclocking all available Core i7 processors. Anyone choosing additional displays rather than SLI/CrossFire modes will be pleased to see that the PCIe 2.0 x4 slot is open-ended to allow a third x16 graphics card to be installed.

Moving the top x16 slot upward two positions compared to the full-ATX version required everything else to be placed slightly “northward.” Asus removed the lower DIMM latches to avoid interference with the closely-placed graphics card, but we didn’t have any issue with module retention. The CPU socket is likewise shifted toward the Rampage II Gene’s top edge, leaving no room for the additional voltage regulator components found on the top edges of its full-ATX predecessor.

Our overclocking test will determine the stability of the resulting eight-phase design.

The Rampage II Gene follows the Asus P6T and Foxconn BloodRage by providing mounting holes for both LGA-1366 and LGA-775 CPU coolers. This could be especially handy when choosing an air-cooling sink to fit a particular case, or when carrying forward a previous-generation liquid-cooling water block.

Asus finally gets rid of the floppy drive connector that nobody really wants, yet some Windows XP installations require (we get around this in the lab by keeping a USB disk drive handy). Conversely, an Ultra ATA header that even fewer users require has been added by way of JMicron’s JMB363 combination controller, a part that’s also responsible for a seventh internal and a single external SATA port.

Putting the front-panel connectors along the motherboards bottom edge hasn’t traditionally been challenging with Micro-ATX cases, but this is an SLI/CrossFire board. Stuffing cables under a card can be difficult, especially when round cables such as those used with most front-panel audio connectors must be smashed flat at the connector.

Power and reset buttons along the center of the Rampage II Gene’s bottom edge allow testing and configuration outside of any case. A third button labeled “MemOK” forces the motherboard to loosen memory timings significantly, to ease booting with problematic modules.

BIOS

A comparison of voltage and frequency settings can be found in the overclocking portion of this review.

The Asus Extreme Tweaker menu is detailed enough to require three pages of scrolling, yet it begins with a focus on Asus’ automatic overclocking techniques. Mixed in with those automatic overclocking controls, the Rampage II Gene properly detected the X.M.P. values for our memory at both 2,000 and 1,866 MHz.

DRAM timings are spread across nearly two pages, but Asus simplifies configuration by allowing some to be set manually and others automatically.

Asus O.C. Profile allows saving up to eight custom BIOS configurations, while the TweakIT Batch File provides simpler overclocking configuration and similar saving function to two files.

Accessories

New to the Rampage II Gene is an oversized case sticker. The LCD Poster provides useful information concerning system status, but connecting it to the board through a hole in the I/O plate leaves it dangling whenever the system is moved.

3. DFI LANParty Jr X58-T3H6

Features and Layout

The LANParty Jr X58-T3H6 looks almost identical to its full-sized predecessor with the exception of two missing slots, and we’re hoping to reach similar performance levels and overclocking capability. Yet DFI didn’t simply shorten the slot region of its previous design, as everything from the X58 northbridge to the bottom edge had to be moved upward to fill a previously unused slot position.

Also missing compared to the larger motherboard version are two SATA and two IEEE-1394 FireWire ports. The missing SATA ports can be blamed on the switch from the larger JMB363 combination controller to smaller JMB368. Complete obsolescence of Ultra ATA means that DFI could have just left the add-in controllers off entirely, and we’d certainly rather have a couple eSATA ports than an Ultra ATA header. But while FireWire is also obsolete, the fact that most high-end cases have front-panel FireWire presents a small problem for some builders.

DFI also kept the floppy header required by some Windows XP users to add RAID or AHCI drivers during installation. All current motherboards have a floppy controller built into the multi-I/O IC, but some manufacturers have omitted the cable connector in an effort to modernize the look of their product. AHCI mode is particularly beneficial for adding removable drives on the X58-T3H6’s hot-plug-capable BIOS.

The LANParty Jr’s raised slot positions cause long graphics cards to interfere with the lower DIMM latches, forcing users to remove the top graphics card prior to memory changes. But DFI apparently forgot to move its front-panel audio connector upward with the slots, placing it directly beneath any second graphics card so that any attached cable must be smashed flat in order to fit under a second graphics sink.

Like the Asus competitor, the LANParty Jr X58-T3H6 provides all sixteen PCIe 2.0 pathways to each x16 slot. But unlike Asus, DFI didn’t provide an open-ended connector on its x4 slot. Builders who wanted a small six-display “home-office workstation” for such tasks as investment analysis might be disappointed that they can’t cram a x16 card into DFI’s x4 slot.

Power and reset buttons found on the motherboard’s bottom edge allow easy bench-top testing, and pushing these in tandem provides a CLR_CMOS function. The Port 80 diagnostics display is also handy for bench testing, though all of these features become inaccessible in a fully-configured SLI system. Though internal buttons are inaccessible in fully-configured systems, the CLR_CMOS jumper on the motherboard’s I/O panel is less prone to accidental-engagement than Asus’ rear-panel button.

BIOS

DFI’s Genie BIOS settings take up far less page space than those of its competitor, but still provides full manual control of the most significant clock speeds and ratios.

DRAM timings and system voltage levels are accessed through separate sub-menus. The LANParty Jr X58-T3H6 offers fewer timing but more voltage controls compared to its competitor, and most of the “missing” memory settings are rarely used even by the most experienced overclockers.

The LANParty Jr X58-T3H6 provides four BIOS registers to save custom configurations, but perhaps more important is its ability to restore the “last bootable setting” once BIOS has been cleared.

Accessories

The LANParty Jr X58-T3H6 includes round cables for both floppy and Ultra-ATA drives, but its SATA cable set is unfortunately reduced to two. DFI also includes bridges for both SLI and CrossFire configurations and four easy-grip replacement jumpers.

4. Test Settings

Though most of our current articles contain updates to both hardware and software, a previous test configuration was required to make today’s benchmark results consistent with those of our other X58 motherboard comparisons.

Test System Configuration
CPU

Intel Core i7 920 (2.66 GHz, 8.0 MB Cache)

CPU Cooler

Swiftech Apogee GTZ Liquid Cooling

RAM

Kingston KHX16000D3ULT1K3/6GX (6.0 GB)
DDR3-2000 at DDR3-1866 CAS 7-8-7-20

Graphics

XFX GeForce GTX 285 XXX Edition
670MHz GPU, GDDR3-2500

Hard Drive

Western Digital WD5000AAKS, 500 GB
7,200 RPM, SATA 3 Gb/s, 16 MB cache

Sound

Integrated HD Audio

Network

Integrated Gigabit Networking

Power

Coolermaster RS850-EMBA
850W, ATX12V v2.2, EPS12V

Software

OS

Microsoft Windows Vista Ultimate x64 SP1

Graphics

Nvidia GeForce 181.20 WHQL

Chipset

Intel INF 9.1.0.1007


While other reviewers have standardized their tests using DDR3-1600 CAS 8 memory, faster modules are required to assess the full overclocking capabilities of X58 motherboards. Kingston’s DDR3-2000 was chosen for its win in our 6 GB DDR3 overclocking shootout.

Zalman’s ZM-STF1 thermal grease was chosen for its quick set-in time, low thermal resistance, and mess-free application.

Excellent cooling is required to reach our Core i7 920’s overclocking limit. Swiftech’s Apogee GTZ moves heat quickly away from the CPU, via its MCP-655b high-volume pump and 3x120mm radiator.

Top benchmark performance in our previous X58 Motherboard Shootout has made the Asus P6T the reference platform for most of our tests. Today it represents the “Full ATX standard” by which we can judge the effectiveness of micro-ATX alternatives.

Benchmark Configuration

3D Games

Call of Duty: World at War

Patch 1.1, FRAPS/saved game
High Textures, No AA / No AF, vsync off
Ultra Textures, 4x AA / Max AF, vsync off

Crysis

Patch 1.2.1, DirectX 10, 64-bit executable, benchmark tool
Test Set 1: High Quality, No AA
Test Set 2: Very High Quality, 8x AA

Far Cry 2

DirectX 10, Steam Version, in-game benchmark
Test Set 1: High Quality, No AA
Test Set 2: Ultra High Quality, 8x AA

World in Conflict

Patch 1009, DirectX 10, timedemo
Test 1: High Details, No AA / No AF
Test 2: Very HighDetails 4x AA / 16x AF

Audio/Video Encoding

iTunes

Version: 7.7.0.43
Audio CD (Terminator II SE), 53 min
Default format AAC

Lame MP3

Version: 3.98 Beta 3 (05-22-2007)
Audio CD "Terminator II SE", 53 min
wave to MP3

TMPGEnc 4.5

Version: 4.5.1.254
Import File: Terminator 2 SE DVD (5 Minutes)
Resolution: 720x576 (PAL) 16:9

DivX 6.8.3

Encoding mode: Insane Quality
Enhanced multithreading enabled using SSE4
Quarter-pixel search

Xvid 1.1.3

Display encoding status = off

MainConcept Reference 1.5.1

MPEG2 to MPEG2 (H.264), MainConcept H.264/AVC Codec, 28 sec HDTV 1920x1080 (MPEG2), Audio: MPEG2 (44.1 kHz, 2 Channel, 16-Bit, 224 kbp/s), Mode: PAL (25 FPS)

Productivity

Autodesk 3ds Max 9

Version: 9.0, Rendering Dragon Image at 1920x1080 (HDTV)

Grisoft AVG Anti-Virus 8

Version: 8.0.134, Virus base: 270.4.5/1533, Benchmark: Scan 334 MB Folder of ZIP/RAR compressed files

WinRAR 3.80

Version 3.70 BETA 8, WinZIP Commandline Version 2.3, Compression = Best, Dictionary = 4,096 KB, Benchmark: THG-Workload (334 MB)

WinZip 11

Version 11.2, Compression = Best, Benchmark: THG-Workload (139 MB)

Sythetic Benchmarks and Settings

3DMark Vantage

Version: 1.02, GPU and CPU scores

PCMark Vantage

Version: 1.00, System, Memory, Hard Disk Drive benchmarks, Windows Media Player 10.00.00.3646

SiSoftware Sandra XII SP2

Version 2008.5.14.24, CPU Test = CPU Arithmetic / MultiMedia, Memory Test = Bandwidth Benchmark

5. Benchmark Results: 3D Games

The Rampage II Gene bursts from the gate in CoD:WaW, with performance that tops even full-ATX solutions. DFI lags slightly behind the full-sized P6T, but this could be due to less-aggressive use of Intel Turbo mode.

The Rampage II Gene wins again at Crysis High-Details, but turning up the eye-candy even more puts the full-sized P6T on top.

The tiny Rampage II Gene tears through Far Cry 2.

World in Conflict favors the full-sized P6T, our reference platform for 2009.

6. Benchmark Results: A/V Encoding And Productivity

The Rampage II Gene takes a big win in Apple iTunes format conversion. This type of leadership is so unusual that we re-tested while keeping an eye on clock speed, yet never saw the motherboard exceed its default 0.20% overclock.

More typical of our expectations, Lame MP3 encoding shows identical performance between different motherboard models.

The Rampage II Gene leads TMPGEnc video conversion slightly, while the full-sized P6T takes the prize in Mainconcept.

A dead heat in 3ds Max is once again typical of our previous expectations.

Yet typical results are the exception for the Rampage II Gene, as it again edges out competitors in AVG virus scans.

Both Asus motherboards edge out DFI in our file compression suite.

7. Benchmark Results: Synthetics

The Rampage II Gene leads 3DMark, but its advantage here is tiny.

The most impressive lead for Asus’ Rampage II Gene might be in PCMark, but it’s also DFI’s biggest downfall.

Slight advantage in clock control are often revealed in Sandra’s CPU tests, but the difference seen above is marginal.

The full-sized ATX motherboard leads Sandra’s Memory Bandwidth benchmark, but again by a trivial amount.

8. Overclocking, Power And Heat

Though available Core i7 processors are typically limited to around 220 MHz BCLK, a few select samples can at least exceed that limit. Extreme overclockers, especially though with sub-ambient cooling, will also try higher voltage levels than most of us would consider safe. Here’s what the two Micro ATX Core i7 motherboards offer in BIOS.

BIOS Frequency And Voltage Settings (For Overclocking)

 

Asus Rampage II
Gene

DFI LANParty Jr
X58-T3H6

CPU Base Clock

100-500 MHz (1 MHz)

133-250 MHz (1 MHz)

CPU Multiplier

Yes

Yes

DRAM Data Rates

BCLK x6 - x16 (x2)

BCLK x6 - x16 (x2)

PCIe Clock

100-200 MHz (1 MHz)

100-250 MHz (1 MHz)

CPU Vcore

0.85-2.50V (6.25mV)

1.00-2.00V (12.5mV)

Uncore Voltage

1.20-2.50V (6.25mV)

1.21-1.61V (10mV)

IOH Core

1.11-2.20V (13.25mV)

1.10-1.45V (50mV)

ICH Core

1.11-2.20V (13.25mV)

1.05-1.35V (100mV)

DRAM Voltage

1.51-2.50V (13.25mV)

1.455-2.40V (15mV)

CAS Latency

3-11 Cycles

1-11 Cycles

tRCD

3-10 Cycles

1-31 Cycles

tRP

3-10 Cycles

1-10 Cycles

tRAS

3-31 Cycles

1-31 Cycles


While the Rampage II Gene offers a broader range of Bclk and voltage settings, the LANParty Jr X58-T3H6’s BIOS limits still exceed the tolerable levels for available hardware.

The toughest test is to see how far each motherboard can stably push our Core i7 920 processor using the same voltage settings and timings. We used eight threads of Prime95 x64 to stress each configuration.

The Rampage II Gene beat the LANParty Jr X58-T3H6 in top clock speed at 1.45V core, but not by enough to consider a definitive victory. More significant is that both Micro-ATX motherboards proved more stable than the full-sized Asus P6T, invalidating myths about Micro-ATX incapacity.

A base clock of 220MHz could potentially take a Core i7 920 to 4.40 GHz, but only with a very good core and advanced cooling. A lower multiplier allowed us to bypass those special conditions to find a retail processor’s Bclk limit on each motherboard, just shy of the 220 MHz mark for both the LANParty Jr X58-T3H6 and Rampage II Gene. Both motherboards beat the full-sized P6T, again putting to rest notions about Micro ATX instability.

DFI wins the memory overclocking comparison. The second-place Rampage II Gene beats the full-sized P6T on average due to its much higher six-module results, but the P6T rises to the middle when only three modules are installed.

Asus alters the standard voltage level slightly in addition to its 0.20% default CPU overclock, but the increase in power is still somewhat surprising. Our most recent experience has shown that most motherboards typically fall between the LANParty Jr X58-T3H6 and Rampage II Gene in power consumption when using these components, so DFI deserves about half the credit for the significant disparity in power consumption.

In spite of its higher power consumption, the Rampage II Gene runs slightly cooler than LANParty Jr X58-T3H6.

9. Conclusion

Top-quality Micro-ATX motherboards offer a level of performance, stability, and capability believed by many to be possible only with full-sized parts. The results make sense because of the equally-sized upper portion of Micro-ATX and full ATX motherboards, but this is the first time we’ve seen manufacturers put serious effort into the overclocking capabilities of the four-slot form factor. For those of you who’ve jumped ahead, the DFI LANParty Jr X58-T3H6 and Asus Rampage II Gene actually exceeded the overclocking capability of the full-sized Asus P6T on this article’s previous page.

But some readers are only interested in benchmarks, so here’s a quick recap of how each Micro-ATX motherboard performed against the full ATX Asus P6T.

The LANParty Jr X58-T3H6 lagged slightly behind the full-sized-class-leading P6T, but a look back at our earlier comparison reveals that performance levels for DFI’s ATX and Micro-ATX motherboards are almost identical. The real star of today’s performance shootout was the Rampage II Gene, with prowess to outclass even the best board of that previous comparison.

We’re sure that several readers will point out that Asus has an advantage due to running its default base clock at 0.20% over reference speed, but its performance lead is much greater than 0.20%. We even monitored Bclk through several benchmarks to make sure Asus wasn’t ramping up base clock in some other way. Instead it appears that Asus has figured out a way to better-manage Intel’s Turbo mode, while also finessing voltage levels to keep everything stable. But Asus’ added performance came at a sizeable power-consumption cost. Dividing the relative performance of each motherboard (from the chart above) by the average of its full load and idle power consumption (previous page) shows a stark contrast in efficiency.

Though it did beat even our fastest full-ATX P6T motherboard in both performance and efficiency, the Rampage II Gene’s improved efficiency wasn't able to best that of DFI’s LANParty Jr X58-T3H6.

Great results from both Micro-ATX motherboards prove that the only significant reason to “go big” is to get a couple extra slots. But most builders never use more than four slots, even in an SLI gaming configuration. Anyone who truly desires a smaller-format high-end PC can no longer use the motherboard as their excuse for not chasing their dream.