MSI’s K8N Diamond Plus motherboard

Manufacturer MSI
Model K8N Diamond Plus
Price (Street)
Availability Now

SEEMINGLY CONTENT TO MAKE DO with the original nForce4 SLI, motherboard manufacturers have been slow to adopt NVIDIA’s nForce4 SLI X16 chipset. For months, Asus’ A8N32-SLI was the only Socket 939 motherboard capable of feeding 16 lanes of PCI Express to each of a pair of graphics cards in SLI, giving enthusiasts seeking gobs of graphics bandwidth few options. Fortunately, that’s starting to change. Mobo makers appear to be warming to the X16 chipset, and MSI is the latest to take the plunge with the K8N Diamond Plus.

Like most enthusiast boards, the K8N Diamond Plus comes loaded with features, including dual Gigabit Ethernet options, two flavors of Serial ATA RAID, and a BIOS filled with tweaking, overclocking, and even fan speed controls. MSI has also fitted the board with a Sound Blaster Audigy chip that promises support for EAX Advanced HD and better audio quality than typical motherboard audio implementations.

Is the K8N Diamond Plus’ integrated Audigy audio really all that, and more importantly, can the board keep up with the latest nForce4 SLI, X16, and even CrossFire boards on the market? Let’s find out.

The specs

CPU support Socket 939-based Athlon 64 processors
North bridge NVIDIA C51D
South bridge NVIDIA nForce4 SLI
Interconnect HyperTransport (8GB/sec)
Expansion slots 2 PCI Express x16
1 PCI Express x4
2 PCI Express x1
2 32-bit/33MHz
Memory 4 184-pin DIMM sockets
Maximum of 4GB of DDR266/333/400 SDRAM
Storage I/O Floppy disk
2 channels ATA/133 with RAID 0, 1, 0+1 support
4 channels Serial ATA with RAID 0, 1, 0+1, 5 support
2 channels Serial ATA with RAID 0, 1 support via SiI 3112
Audio 8-channel HD audio via Creative Sound Blaster Audigy SE
Ports 1 PS/2 keyboard
1 PS/2 mouse
1 serial port
1 parallel port
USB 2.0 with headers for 6 more
1 Firewire via VIA VT630 with header for 2 more
1 RJ45 10/100/1000
1 RJ45 10/100/1000 via Marvell 88E8053

1 analog front out
1 analog bass/center out
1 analog rear out
1 analog surround out
1 analog line in
1 analog mic in
1 coaxial digital S/PDIF input
1 coaxial digital S/PDIF output

Bus speeds HT: 200-450MHz in 1MHz increments
HT multipliers: 1X-5X
PCI-E: 100-148MHz in 1MHz increments
Interconnect: 200, 400, 600, 800, 1000MHz
Bus dividers DRAM/HT: 1/2, 2/3, 5/6, 5/4, 1/1
Voltages CPU: auto, 0.8-1.55V in 0.025V increments
CPU overvolt: +0.05-0.75V in 0.05 increments
DDR: 2.6-4.1V in 0.05V increments
PCI-E: 1.5-1.85 in 0.05V increments
North bridge: 1.2-1.5 in 0.1V increments
Monitoring Voltage, fan status, and temperature monitoring
Fan speed control CPU, north bridge

The K8N Diamond Plus comes loaded with everything you’d expect from a high-end board, including a whopping 38 lanes of PCI Express. 18 of those lanes can be found on the chipset’s north bridge, which reserves 16 lanes for the motherboard’s primary PCI-E x16 slot. The remaining 20 lanes are on the south bridge, where 16 are consumed by the second x16 slot. That leaves a total of six PCI Express lanes up for grabs—two at the north bridge and four at the south bridge. The extra north bridge lanes are consumed by one of the PCI-E x1 slots and the Marvell 88E8053 Gigabit Ethernet controller, while the second x1 slot, Silicon Image 3132 SATA controller, and x4 slot are fed by the south bridge. However, the south bridge has just four lanes of PCI-E bandwidth to spare, so the x4 slot actually only gets two lanes of bandwidth.

Because its PCI Express lane distribution splits graphics cards in SLI between north and south bridge chipset components, some have charged that the nForce4 SLI X16’s chipset interconnect could bottleneck performance. The interconnect is a 16-bit/1GHz HyperTransport link that provides 8GB/s of bandwidth—just enough to fully saturate one PCI Express x16 slot. Once you throw in additional south bridge I/O traffic, things get a little more crowded. However, it’s important to note that the nForce4 SLI X16’s chipset interconnect provides just as much bandwidth between its north and south bridge components as the chipset has between itself and the Athlon 64 processor. Consolidating enough PCI Express lanes at the north bridge for two graphics cards is undoubtedly a more elegant approach, and one that ATI employs with its Radeon Xpress 3200, but that doesn’t necessarily mean that NVIDIA’s implementation is any slower in the real world.

Given the nForce SLI X16’s abundance of south bridge I/O components, it’s no surprise there are concerns about interconnect bandwidth. Of those components, the south bridge’s Gigabit Ethernet and ATA and RAID controllers are the most notable. And then there are the auxiliary GigE and Serial ATA RAID controllers. Both of those chips ride the PCI Express bus, so they won’t contend for PCI bandwidth.

With auxiliary storage and networking controllers riding the PCI Express bus, the K8N Diamond Plus’ only PCI-based peripherals are its VIA VT6306 Firewire chip and a Sound Blaster Audigy SE. This is the first mobo we’ve encountered with an Audigy SE onboard, but unfortunately, the Special Edition appears to be of the short bus variety. Despite appearing as an Audigy in Windows’ Device Manager, the chip is actually a Creative CA0106-DAT. We’ve seen the same chip on other boards, but in those cases, it identified itself as a Sound Blaster Live! 24-bit. Adding to the intrigue, the K8N Diamond Plus ships with a Creative Live! 24-bit driver CD, which must be used to install the “Audigy” drivers for the board.

Regardless of its true origins, the Audigy SE claims to support high-definition audio and EAX Advanced HD. It doesn’t appear to accelerate positional 3D audio in hardware, though.


The board
The K8N Diamond Plus’ veritable cornucopia of peripheral chips, slots, and ports look great on the spec sheet, but squeezing them all onto a standard ATX form factor isn’t easy. Fortunately, MSI has done a pretty good job with the layout, even if it looks a little crowded from above.

We’re not thrilled with the position of some of the board’s power connectors, though. The 24-pin primary power connector is located along the right edge near the top of the board, just where we like it. However, the eight-pin auxiliary 12V connector is far enough down the left side of the board to create cable clutter around the CPU socket. Cable clutter can restrict air flow around the processor heat sink, potentially raising CPU temperatures in the process.

Despite its eight-pin power plug, the K8N Diamond Plus works just fine with older four-pin ATX 12V power connectors. The board also has a four-pin Molex connector above the top PCI Express x16 slot, but it’s only needed when running a pair of graphics cards in SLI.

As if cable clutter around the CPU weren’t enough, the K8N Diamond Plus’ north bridge cooler leaves little room between itself and the socket. The cooler is close enough to create clearance problems with wider processor heat sinks like Zalman’s CNPS7700, although there’s enough room for the slimmer CNPS9500. Given that MSI appears to have taken great care in ensuring that only low-profile capacitors surround the CPU socket, it’s particularly disappointing to see the north bridge cooler get in the way.

Were the cooler a fanless design, it might be easier to forgive. But it’s not. Despite a fancy heat pipe linking heat sinks on the north and south bridge chips, the cooler still relies on a tiny chipset fan to keep temperatures in check. We’d prefer passive chipset cooling, as smaller fans tend to develop increasingly annoying whines over time. The K8N Diamond Plus’ fan could be an exception, of course, but it’s rare to see generic chipset fan maintain reasonable noise levels with extended use.

The chipset cooler’s saving grace is its low-profile south bridge component, which leaves plenty of clearance for longer graphics cards. Even the heat pipe is carefully shaped to avoid conflict with PCI Express cards, and the K8N Diamond Plus can support a bunch of them. You see, in addition to a pair of standard x16 slots, the board’s yellow PCI-E x4 slot has also been notched to accept longer x16 graphics cards. Cards installed in the slot obviously won’t benefit from a full 16 lanes of bandwidth, and they won’t be able to accelerate 3D graphics as part of a three-card SLI configuration, but users should at least be able to squeeze a couple of extra monitor outputs from the slot.

Although the K8N Diamond Plus’ array of PCI Express slots is certainly generous, the board has only two standard PCI slots. With a double-wide SLI configuration, one of those PCI slots will be blocked, leaving only one slot available for audio cards, TV tuners, and the like. Double-wide SLI configurations will leave two PCI Express slots open, but given the dearth of PCI Express peripherals, we’d almost prefer that one of those slots be plain old PCI. We’d really rather have a wider selection of PCI Express x1 peripherals so we could ditch PCI completely, though.

Moving to the right, we see the K8N Diamond Plus’ Serial ATA ports all neatly arranged along the edge of the board. This is exactly where they should be, and although gargantuan secondary graphics cards like the GeForce 7900 GTX can crowd the ports a little, they should all still be usable.

Speaking of ports, the K8N Diamond Plus’ backplane is loaded. Here, you’ll find a handful of legacy ports in addition to plenty of USB, Ethernet, analog audio, and even Firewire. MSI has also equipped the board with coaxial and TOS-Link S/PDIF outputs for those with digital speakers or receivers, but there’s no provision for digital audio input.

If there aren’t enough ports on the backplane for your stack of Firewire and USB peripherals, the K8N Diamond Plus also comes bundled with PCI brackets for an additional two Firewire and USB ports. There are onboard headers for four more USB ports on top of that, too.


The BIOS and nTune support
While the K8N Diamond Plus is stacked with features and peripherals, the BIOS will have to be just as well equipped to cater to enthusiasts and overclockers.

Things start well enough, with the BIOS offering full control over a laundry list of Athlon 64 memory controller timings. There are just enough options to get you into trouble if you don’t know what you’re doing.

Speaking of trouble, the BIOS also has an extensive array of overclocking options. Users can crank the HyperTransport clock all the way up to 450MHz, but at those speeds, the fact that the BIOS’s CPU multiplier only goes down to 8x could be a problem. That’s a shame, as MSI hasn’t skimped in the voltage department. CPU voltages are available as high as 2.3V, and it’s possible to crank the memory up to an impressive 4.1V with an onboard jumper switch.

We’ve been around long enough to prefer overclocking the old fashioned way, but the K8N Diamond Plus also has a couple of automatic overclocking options for those seeking something a little simpler. Dynamic overclocking kicks in when the CPU is under load, and users can set it to increase speeds by between 1% and 15%. The BIOS also allows users to overclock select NVIDIA graphics cards by similar percentages. Most importantly, though, all automatic overclocking options are disabled by default.

Unfortunately, not all of the BIOS defaults are correct. The 1.00 revision of the BIOS incorrectly sets a slower HyperTransport link between the chipset’s north and south bridge components. It’s easy to bump this connection up to full speed, and MSI has corrected the problem in a beta BIOS it released to us for testing.

Moving along, the K8N Diamond Plus’ BIOS offers a pretty decent array of automatic fan speed control options. Users can set temperature targets for the CPU and north bridge fans, but sadly, there’s no fan speed control for the board’s system fan headers. The BIOS also lacks fan failure or temperature-based shutdown and alarm conditions.

Users can get their hands on fan speed and temperature-based alarm conditions by installing MSI’s CoreCenter Windows software, though.

Otherwise, CoreCenter is more of a glorified hardware monitoring app than a BIOS extension. It’s possible to adjust the level of dynamic overclocking, but serious overclockers will want to stick with the BIOS, especially since the board’s support for nTune is among the worst we’ve seen.

In addition to lacking support for nTune’s hardware monitoring capabilities, the K8N Diamond Plus also lacks the hooks necessary to support the app’s overclocking and memory timing controls. About the only thing you can do with nTune is adjust graphics card clock speeds, which really have nothing to do with the motherboard. I guess MSI would rather you use CoreCenter to monitor and tweak the board, but since CoreCenter doesn’t match many of nTune’s features and capabilities, the user ends up losing out.


Our testing methods
Today we’ll be comparing the K8N Diamond Plus’ performance with that of DFI’s LANParty UT RDX200 CD-DR and LANParty UT NF4 SLI-DR Expert, ECS’s KA1 MVP Extreme, and Asus’ A8N32-SLI and A8R-MVP.

All tests were run at least twice, and their results were averaged, using the following test systems.

Processor AMD Athlon 64 FX-53 2.4GHz
System bus HyperTransport 16-bit/1GHz
Motherboard Asus A8R-MVP Asus A8N32-SLI Deluxe DFI LANParty UT RDS200 CF-DR DFI LANParty UT NF4 SLI-DR Expert ECS KA1 MVP Extreme MSI K8N Diamond Plus
BIOS revision 0402 1009 RDXDC23 NF4EDC07 1.0e 1.0
North bridge ATI Radeon Xpress 200 CrossFire NVIDIA nForce4 SPP 100 ATI Radeon Xpress 200 CrossFire NVIDIA nForce4 SLI ATI Radeon Xpress 200 CrossFire NVIDIA nForce4 SPP 100
South bridge ULi M1575 NVIDIA nForce4 SLI ATI SB450 ATI SB450 NVIDIA nForce4 SLI
Chipset drivers ULi ForceWare 6.85 CATALYST 6.1 ForceWare 6.70 CATALYST 6.1 ForceWare 6.85
Memory size 2GB (2 DIMMs) 2GB (2 DIMMs) 2GB (2 DIMMs) 2GB (2 DIMMs) 2GB (2 DIMMs) 2GB (2 DIMMs)
Memory type  Corsair CMX1024-3500LLPRO DDR SDRAM at 400MHz
CAS latency (CL) 2 2 2 2 2 2
RAS to CAS delay (tRCD) 3 3 3 3 3 3
RAS precharge (tRP) 2 2 2 2 2 2
Cycle time (tRAS) 6 6 6 6 6 6
Command rate 1T 1T 1T 1T 1T 1T
Hard drives Western Digital Raptor WD360GD 37GB SATA
Audio M1575/AD1986A nForce4 SLI/ALC850 SB450/ALC882 nForce4 SLI/ALC850 SB450/ALC880 Audigy
Audio driver Realtek 3.82 Realtek HD 1.30 Realtek 3.82 Realtek HD 1.30 Creative
Graphics NVIDIA GeForce 7800 GTX with ForceWare 81.98 drivers
OS Microsoft Windows XP Professional
OS updates Service Pack 2, DirectX 9.0c

Thanks to Corsair for providing us with memory for our testing. 2GB of RAM seems to be the new standard for most folks, and Corsair hooked us up with some of its 1GB DIMMs for testing.

Our test systems were powered by OCZ PowerStream power supply units. The PowerStream was one of our Editor’s Choice winners in our latest PSU round-up.

We used the following versions of our test applications:

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests. Most of the 3D gaming tests used the Medium detail image quality settings, with the exception that the resolution was set to 640×480 in 32-bit color.

All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.


Memory performance

As usual, performance in our memory subsystem tests is pretty close. Any variation here is a result of how each board manufacturer has chosen to tune the Athlon 64’s on-die memory controller. Some are more aggressive than others, as evidenced by the scores from NF4 SLI-DR Expert, but that board’s tighter timings limit its compatibility with generic memory modules. The K8N Diamond Plus suffers no such compatibility problems, as its default memory timings are more conservative.



WorldBench is a wash, with five of six boards offering identical scores.


The K8N Diamond Plus bounces around in our low-resolution gaming tests. Overall, it’s about as fast as the rest of the pack.


SLI gaming performance
Our first round of gaming tests were conducted with low in-game detail levels and display resolutions, but we’ve cranked things up for a second round. These tests use high resolutions, high detail levels, and anisotropic filtering and antialiasing, so they should be fairly indicative of how gamers play in the real world. We’ve tested each board with a single GeForce 7800 GTX, and we’ve also benchmarked the Asus A8N32-SLI, DFI LANParty UT NF4 SLI-DR Expert, and MSI K8N Diamond Plus with a pair of 7800 GTXs running in SLI. We’ve also tested the KA1 MVP, RDX200, and A8R-MVP with single and dual Radeon X850 XT cards using ATI’s Catalyst 6.1 graphics drivers.

When looking at our SLI performance results, pay special attention to the jump in performance from single- to multi-card configurations. We’re not out to compare the Radeon X850 XT’s performance with that of the GeForce 7800 GTX; we just want to see how adding a second card can improve overall performance.

With high resolutions, detail levels, and antialiasing and aniso bottlenecking performance at the graphics card, the K8N Diamond Plus has no problem keeping up with other SLI implementations.


Cinebench rendering

Nothing to see here. Move along.

Sphinx speech recognition

Sphinx is particularly sensitive to memory bandwidth and latency, so it’s no surprise to see the aggressively-timed DFI board edge out the others. The K8N Diamond Plus does well to keep up with the rest of the field.


Audio performance

So much for the onboard Audigy. Although the K8N Diamond Plus’ audio chip supports more simultaneous DirectSound voices than the other onboard audio implementations, CPU utilization is nothing special. In fact, with EAX enabled, the K8N Diamond Plus’ Audigy has higher CPU utilization than boards using Realtek’s software audio drivers.

To be fair, we’ve encountered numerous positional audio problems with EAX and Realtek’s audio drivers. The Audigy’s higher CPU utilization here may simply be due to the fact that it’s doing EAX correctly. We’d prefer that the chip could do it correctly in hardware, though.

Audio quality
We used an M-Audio Revolution 7.1 card for recording in RightMark’s audio quality tests. Analog output ports were used on all systems. To keep things simple, I’ve translated RightMark’s word-based quality scale to numbers. Higher scores reflect better audio quality, and the scale tops out at 6, which corresponds to an “Excellent” rating in RightMark.

With the exception of the Total Harmonic Distortion test, the K8N Diamond Plus’ Audigy fares reasonably well in RightMark Audio Analyzer. To my ears, it sounds a little brighter than typical onboard audio implementations, with less of the muffled flatness that tends to plague Realtek audio.


ATA performance
ATA performance was tested with a Seagate Barracuda 7200.7 ATA/133 hard drive using HD Tach 3.01’s 8MB zone setting.

The K8N Diamond Plus’ ATA performance is about what we’d expect from the nForce4 SLI X16 chipset.


Serial ATA performance
Moving to Serial ATA, we tested performance with a Western Digital Raptor WD360GD SATA hard drive. Again, we used HD Tach 3.01’s 8MB zone test.

Everything’s pretty even in our Serial ATA performance tests, as both of the K8N Diamond Plus’ SATA controllers are competitive.


USB performance
Our USB transfer speed tests were conducted with a USB 2.0/Firewire external hard drive enclosure connected to a 7200RPM Seagate Barracuda 7200.7 hard drive. We tested with HD Tach 3.01’s 8MB zone setting.

MSI does well in our USB performance tests, especially when you consider the K8N Diamond Plus’ low CPU utilization.

Firewire performance
Our Firewire transfer speed tests were conducted with the same external enclosure and hard drive as our USB transfer speed tests.

Firewire performance is pretty even across the board, at least when you look at the top competitors, which include the K8N Diamond Plus.


Ethernet performance
We evaluated Ethernet performance using the NTttcp tool from Microsoft’s Windows DDK. The docs say this program “provides the customer with a multi-threaded, asynchronous performance benchmark for measuring achievable data transfer rate.”

We used the following command line options on the server machine:

ntttcps -m 4,0, -a

..and the same basic thing on each of our test systems acting as clients:

ntttcpr -m 4,0, -a

Our server was a Windows XP Pro system based on Asus’ P5WD2 Premium motherboard with a Pentium 4 3.4GHz Extreme Edition (800MHz front-side bus, Hyper-Threading enabled) and PCI Express-attached Gigabit Ethernet. A crossover CAT6 cable was used to connect the server to each system.

The nForce4 boards were tested with the NVIDIA Firewall and Jumbo Frames disabled.

NVIDIA recently confirmed that it had scaled back ActiveArmor Gigabit Ethernet acceleration in its latest nForce drivers, and we can see that in action here. The K8N Diamond Plus’s GigE CPU utilization isn’t as low as we’ve seen from the nForce4 before, but NVIDIA says that scaling back the offload engine was the only way to avoid data corruption issues that had plagued ActiveArmor in the past. NVIDIA’s so confident that it’s banished those data corruption issues for good that it’s actually enabling ActiveArmor by default in the latest drivers.

Even with ActiveArmor’s reduced effectiveness, the K8N Diamond Plus’s nForce4 GigE controller still uses fewer CPU resources than the board’s Marvell Gigabit chip. Throughput is excellent for both chips, too.


For our overclocking tests, we swapped in a pair of OCZ Platinum Rev 2 DDR400 memory modules and set them to run at 2.5-4-4-8-1T timings. These DIMMs use Samsung TCCD memory chips and have proven to be great overclockers in the past. We also backed off on our CPU and HyperTransport processor link multipliers to remove them as potential bottlenecks, and nudged the memory voltage up to 2.8V.

With the memory divider set to 1:1, we were able to get the K8N Diamond Plus stable with a 280MHz HyperTransport clock.

280MHz isn’t spectacular, but it’s the highest overclock we’ve been able to squeeze from these DIMMs. Besides, we weren’t done yet. To remove the memory as a potential bottleneck, we backed off on the memory divider and cranked the HT clock a little higher.

20MHz later, we settled on a 300MHz HyperTransport clock. The board was perfectly stable at this speed, but we couldn’t get it to post with a 310MHz HT clock. Not even MSI’s latest beta BIOS could coax 310MHz from the board, and neither could an array of voltage tweaks. That might not be entirely the board’s fault, though. Because the BIOS’ lowest CPU multiplier is 8x, our Athlon 64 FX-53 processor was forced to run above stock speeds with an HT clock higher than 300MHz. This FX-53 is an early sample, and it’s never been particularly comfortable running faster than its default 2.4GHz clock speed.

If MSI wants the K8N Diamond Plus to appeal to more extreme overclockers, it really needs to offer lower CPU multipliers in the BIOS. The board could be capable of HyperTransport speeds well in excess of 300MHz, but without lower CPU multipliers, many users’ CPUs may have a hard time keeping up.


With a street price as low as $180, the K8N Diamond Plus sells for about as much as other nForce4 SLI X16-based motherboards on the market. Its features are pretty comparable, as well, with most manufacturers relying on the same auxiliary Silicon Image Serial ATA RAID and Marvell Gigabit Ethernet chips. The K8N Diamond Plus does have a few unique tricks up its sleeve, though. It’s the only Socket 939 board we’ve seen that supports up to three PCI Express x16 graphics cards, and even if the third card can’t contribute to SLI, it could be useful for those seeking more than four monitor outputs. The K8N is also the first board we’ve seen with a Creative Audigy SE onboard, although the chip’s lack of hardware acceleration for 3D audio limits its appeal.

Apart from the Audigy SE’s lack of hardware acceleration for 3D audio—an omission that’s hard to harp on given the complete lack of hardware-accelerated 3D audio on Athlon 64 motherboards—our only real gripe with the K8N Diamond Plus is the board’s active chipset cooler. The nForce4 chipset has a reputation for running a little hot, but other manufacturers have managed to reel in its thermals with passive cooling designs, and we wish MSI could do the same. At the very least, it would be nice if the cooler didn’t encroach on the area around the CPU socket.

Overall, MSI has built a solid nForce4 SLI X16 motherboard with the K8N Diamond Plus. The board has plenty of extra peripherals, a decent array of BIOS tweaking and overclocking options, and enough unique flair to set itself apart from other offerings on the market. It’s not perfect, mind you, but it’s little more than a nip and tuck away from greatness. 

Comments closed
    • Mick Leong
    • 15 years ago

    I am considering purchasing this mobo – but there is something ’bout the ram capacity that is not mentioned in the review. At MSIComputer’s web page, it states that there are 4 banks and you can put uptp 4Gb memory but system only recognizes 3+Gb memory quote

    “Due to the South Bridge resource deployment, the system density will only be detected up to 3+ GB (not full 4GB) when each DIMM is installed with an 1GB memory module”.

    Shall I interpret this as a hardware design error? And specifically how much is the “+”? 3.1? 3.2? 3.3?

    • fatpipes
    • 15 years ago

    I like the MSI K8N boards, I’ve had much less trouble with them than with the Asus A8N boards. I’m glad they finally adopted the wider PCIe configuration, dual-width cards were only millimeters apart before. This makes the inner heat sink fan virtually useless.

    People are giving this board way too much flak about the nForce cooling solution. It’s actually very nice.
    – The heat pipe is a mostly copper alloy rather than a mostly aluminum alloy (A8N). (good)
    – The heat pipe is shorter than the A8N. (very good) This helps a lot too if you’re laying the board flat in a rackmount chassis. While heat passes UP the pipe nicely, I imagine it probably travels ALONG the pipe exponentially slower as a function of length.
    – The cooling fan actively cools both the voltage regulators and the chipset. (very very good) while the old MSI K8N-SLI only actively cools the chipset, and the A8N shares the heat between the voltage regulators and the nForce chip with no active cooling at all (REAL BAD).
    – The old K8N chipset sink was very small and had a cut-out so you could fit PCI-X cards into the first PCI slot without hitting it. Still… I could probably have digested that sink. I’ve since then replaced it with a better heat sink and a long card would not fit in it. The heat pipe base on the new K8N removes this limitation.
    – The chipset/vreg cooling fan is larger than the original chipset fan (I’ll say at least twice as big just by looking at it) and the low-profile capacitors means they’re probably higher quality and they’re probably lowering the power envelope by doing this too. Those capacitors are at least half the size of the original K8N’s, which were disruptively beefy.

    I don’t O/C (not my hardware) so a lot of the limitations you’re all talking about with regards to large heat sinks don’t really affect me or I’m guessing most other people too. I bought a 120MM thermaltake heat sink, and frankly, I’ve never considered installing it because I value the ability to move my hands around inside the case. If you’re going to OC, get real and get a water cooling setup. And too bad for all you HTPC guys looking to use the massive wide heat sinks so you don’t need a CPU cooling fan, you’re better off with the tall vertical sinks anyway.

    The reason there isn’t a hardware accelerated sound card on this board is that it’s $180 rather than $215. And to be honest, any self-respecting audophile (the only people who will care) is a lot more interested in a discrete sound card having better power stabilization and discrete electrical components dedicated to it.

      • fatpipes
      • 15 years ago

      Good insight

      Yeah I thought so too

      It’s too bad people would rather argue about legacy ports and dells

        • Vrock
        • 15 years ago

        Responding to your own post to compliment yourself is lame.

      • Dissonance
      • 15 years ago

      A few things…

      Although MSI’s literature likes to talk about how the chipset fan cools the VRMS, there’s little direct airflow with even a stock Athlon 64 heat sink installed. The processor heat sink gets in the way, and unless you think air’s going to make two perfect 90-degree corners, there isn’t going to be much left for the VRMs. If MSI was really serious about VRM cooling, you’d think they would have at least put a few passive heat sinks on the VRMs, at the very least.

      Implying that a motherboard with no active cooling is a bad thing discounts the fact that in a properly-cooled chassis, there’s little need for active chipset and/or VRM cooling.

      The chipset fan is no larger than the average chipset cooler. Sorry.

      And finally, why would any audiophine, self-respecting or otherwise, care about hardware acceleration for 3D audio? Hardware acceleration is important for gamers seeking higher in-game frame rates; it has nothing to do with audio playback quality. We mentioned the lack of hardware acceleration because most versions of the Audigy support hardware acceleration, but this “SE” does not, as it appears to be little more than a SoundBlaster Live! 24-bit with software support for a greater number of 3D voices.

    • Chrispy_
    • 15 years ago

    Ugh! – Chipset fan.
    Next board please.

      • Usacomp2k3
      • 15 years ago

      I wonder if it will run w/o the fan plugged in. If you have something like that 7700 that blows air over there anyhow, then that could be enough airflow.

        • Chrispy_
        • 15 years ago

        You can’t build a PC with a chipset fan.

        It will, WITHOUT A SHADOW OF A DOUBT come back in three months time with either instability problems or owner complaining about alarming buzzing noises.

        FFS. It’s beyond a joke now and it all started when people “accepted” that Abit used useless little fans on their KT133 rather than spending $0.50 more on a higer quality one.

          • Usacomp2k3
          • 15 years ago

          Umm. That may be your enthusiast opinion, but believe me it has been done.

            • Chrispy_
            • 15 years ago


            These are for normal people who use their computers once a day for email and web-browsing. How far from enthusiast can you get?


            • Usacomp2k3
            • 15 years ago

            If i read it correctly, you are saying that anyone who buys a motherboard that doesn’t have passive northbridge cooling will return it due to overheating or due to the fan getting too loud?

            Have you seen the DFI boards recently? They sell like hot-cakes, and have a mag-lev fan.

            • Chrispy_
            • 15 years ago

            Yes, you are right. DFI boards are most godly (and most expensive).

            90% of all motherboards come with rubbishy sleeve-bearing fans that last less than 3 months. Asus, Abit, MSI, Gigabyte, Leadtek, Asrock, Foxconn, BFG, ECS, Albatron = I’ve tried them all (and replaced all the NB’s with the Zalman passive chipset cooler.)

            I stick to Asus A8N’s at the moment with the awesome heatpipe. Every moving part is a liability and the faster it spins the faster it will wear out. If it’s a budget PC I go for the ULI or SIS boards with chipset pairs that generate less heat and often come without fans.

            • Usacomp2k3
            • 15 years ago

            So then my conclusion is correct that you are an enthusiast that can’t stand any noise that you don’t understand or are more sensitive to. I wasn’t reading it wrong.

            • Chrispy_
            • 15 years ago

            No. Non-entusiasts are used to Dells, which are silent and always comment on the noise of a cheap build using stock parts (NB fans, GPU fans, AMD’s reference 70mm cooler.

            For the love of all that is obvious, how can you call people who are upgrading from Dells “enthusiasts”? These people can barely understand the difference between a CD and a DVD.

          • Krogoth
          • 15 years ago


          Ooops….. I’m bad and retarded for not reading chipset *[

            • Chrispy_
            • 15 years ago


            At least it’s Friday now and there’s only 4 hours left of this working week ๐Ÿ™‚

            • crabjokeman
            • 15 years ago

            Unless it’s got heatpipes (which ultimately transfer the heat to an active cooling location and/or provide more surface area), cooling an NF4 passively is unreliable, especially while OC’ing or applying more voltage to the chipset. Just ask Chaintech.

            A reminder – the new ATI RD580 dissipates 8W and is easily cooled with a passive chipset. VIA’s K8T900 looks promising too, if they ever get it to the market.

            No more NFurnaces for me.

            • Krogoth
            • 15 years ago

            To restate my previous comment again that Iater I corrected. (sigh)

            I had replaced the stock NB cooler on my EpoX 9NPA+Ultra with a Coolmaster NB heatsink. It has endured a 278Mhz HT bus speed for several months without a problem. The heatsink is warm to touch, and I still have enough clearance for my 7800GTX.

            The real problem was when you try to install a passive NB cooler on some NF4 SLI boards. It usually caused clearance problems for the first or second video card. You could of course use a Dremel to tirm down some of fins on the NB heatsink to make the fit.

    • Vrock
    • 15 years ago

    It seems parallel, DB-9 serial, and PS2 ports will never die.

    And though the “Audigy” on board this thing isn’t the Crab, I still have to wonder: why put integrated audio on an enthusiast board anyway? I’d bet 9 of 10 users are going to disable it and plug in another, more competent sound card anyway.

      • flip-mode
      • 15 years ago

      huh, maybe not on this piece with only *1* pci slot.

        • Vrock
        • 15 years ago

        I’m counting two: one orange, one white, and the spec sheet says there’s two also. But there coulda been three if they left that onboard audio off of there. ๐Ÿ˜‰

          • SpotTheCat
          • 15 years ago

          a lot of people still use integrated audio.

            • Vrock
            • 15 years ago

            I know. My point was that people who buy motherboards like this overwhelmingly /[

          • flip-mode
          • 15 years ago

          Right, but only one with SLI (was what I was thinking).

            • Vrock
            • 15 years ago

            Only with double wide cards though. A pair of 7900GTs would be perfect on this board. But I guess the type of person who buys this board is gonna want GTXs in there….

            • flip-mode
            • 15 years ago

            Good point! I forgot single slot card existed anymore.

      • Bauxite
      • 15 years ago

      PS2 ports are critical for anyone that gives a shit about kvms. Life without kvms is not worth living.

      Its also nice to be able to actually have a chance to kill a process when all the cpus are hammered, usb will stop responding.

      If you want to kill PS2 ports for good, make a new kb/mouse input spec because usb is NOT it.

      Serial/parallel are still needed by some important folks too but they end up having to buy multiport addin cards these days.

        • Krogoth
        • 15 years ago

        Parallel and Serial ports are still more reliable and stable then USB.

        I rather have printers use parallel then a USB. Printers are not exactly the most bandwidth demanding device and parallel interface provides sufficent bandwidth and is proven to be very reliable, unlike USB.

          • Shintai
          • 15 years ago

          More stable? Got proff?

            • Krogoth
            • 15 years ago

            My old man’s older HP 940c (USB/Parallel) gave nothing but connection problems within XP when it used its USB port, however when we used the parallel port the problems went away.

            The old man’s current Kinolta Monica Lasercolor 2400W (USB only) needs to be turn on before his system in order for XP to detect it. Otherwise, you were forced to uninstall the existing printer profile and then disconnect/reconnect USB cable for the printer in order for XP to detect it.

            • Vrock
            • 15 years ago

            Really? I’ve an ancient HP 812C sitting on my desk right now that’s connected via USB, and it’s never given me a problem. Before Win XP it would flake if you installed it on one USB port and then for some reason moved it to another, but other than that it’s been grand.

            Though laser printers are like $80 after rebate now, I still can’t bring myself to give up my old reliable 812C. I’m cheap, I know.

            • Krogoth
            • 15 years ago

            Are you using a parallel to USB convertor or something?

            • Vrock
            • 15 years ago

            Nope, the 812C has jacks for both parallel and USB.

            • Krogoth
            • 15 years ago

            (After googling it)

            That printer is old, but certainly not ancient by printer standards. The really old impact units are the first printers come across my mind to be regrad as ancient.

            • Shintai
            • 15 years ago

            So all the stability discussion is based upon an old printer that doesnt work for your father, and another printer that just wont get detected?

            I can see you clearly did some statistic homework..

            • Krogoth
            • 15 years ago

            I am just saying that USB isn’t the ultimate replacement for serial and parallel interfaces. It still has a couple of issues with its devices and controllers despite being almost 10 years old. I am very certain that my case isn’t unique and is more commonplace then you think.

            • Shintai
            • 15 years ago

            What issues? Proff?

            I guess you are one of the few never had any problems with serial and parallel.

      • UberGerbil
      • 15 years ago

      I’m ambivalent about killing PS/2 (I like USB in theory, but I’m a KVM user), and I have no problem with the (overdue) death of parallel ports, but old-style serial ports still have a lot uses mostly because they operate at such a simple and low level. We had a recent discussion about this
      ยง[<<]ยง I'd be willing to buy a mobo without a serial port but I'd have to mentally add $30 or so to the price to account for a decent serial add-in card. And there goes one of my PCI slots.

    • sbarash
    • 15 years ago

    So if this board’s lowest multiplier is 8x, then does that mean that Cool & Quiet won’t lower the multiplier either?

    Also, how many fan headers does the board have, and which ones are managed?

    How loud is the chipset fan? Does it throttle with load?

    Honestly, I wish TR would start putting a little more emphasis on describing features that one might look for in a quiet low power or even HTPC box. Extreme overclocking isn’t everything…

      • Dissonance
      • 15 years ago

      Cool’n’Quiet controls are separate from the multiplier options presented in the BIOS.

      There are two additional fan headers on the board. And twice in the review (once on the spec sheet on page 1, and again in the BIOS section on page 3) we outline that only the CPU and north bridge fans are temperature controlled.

        • sbarash
        • 15 years ago

        From your tone, it sounds like I offended you. I apologise.

        However, if the lowest multiplier allowed is 8X, they must have done that for a reason. A limitation of their bios programming skills, or of their circuitry? Who knows? Since Cool n’ Quiet does modify the multiplier, and must be enabled by the bios, it stands to reason that Cool n’ Quiet would be affected by this limitation. Normally Cool n’ Quiet can lower the multiplier a low as 4-5x. Did you check this mobo with Cool n’ Quiet?

        I must have passed page 3. Thanks for that info, you guys actually went into far more detail about fan managment features than you have with past motherboard reviews. Its sad that it seems that the latest motherboards seem to be backing away from fan managment. My new A8N-SLI Platinum only has two active headers also. All other mobo’s that I’ve previously purchased have had at least 3.


          • Dissonance
          • 15 years ago

          I’m a fan speed control junkie, so I try to go into as much depth in that area as possible. Unfortunately, not all boards have robust fan speed control, so it’s not something we can often go into much detail regarding.

          As for Cool’n’Quiet, MSI appears to have biased the board’s selection of manual CPU multipliers towards the high end (up to 24x, IIRC), so that may have limited the number of options they were able to provide at the lower end of the scale. That won’t necessarily affect C’n’Q, though, just a user’s ability to manually dictate a given CPU multiplier. We’ve seen boards with proper C’n’Q support that haven’t provided any BIOS-level multiplier control at all.

    • Usacomp2k3
    • 15 years ago

    I haven’t heard from MSI in awhile. Props to them for doing better than a crab ship though, even if it is the ‘short-bus’ variety. ๐Ÿ˜€

    • grug
    • 15 years ago

    Were the jpeg artifacts on the cpu-z screenshots really necessary?

    The goggles do nothing ๐Ÿ™

    • totoro
    • 15 years ago


      • indeego
      • 15 years ago

      lol. more pleaseg{<.<}g

    • Forge
    • 15 years ago

    Always good to roll out a new mobo a month or two before the socket is replaced with a newer one; A nice long upgrade plan is ensured.

    WRT to the OCing/multiplier gripe: WTF is the point of >300MHz HT with a CPU below 2.4GHz?? Even the most macho of ram will be at or beyond specs at 300MHz HT, so why would you need lower multipliers? Sure it made your testing a little less fun, but it shouldn’t ever affect an end user.

      • Severus
      • 15 years ago

      Because memory dividers have basically no performance hit on A64 platforms, so RAM can be kept in check while you send the HTT rocketing skywards.

        • Lazier_Said
        • 15 years ago

        That’s not answering his question, what tangible benefit do you get from the HTT “rocketing skywards”?

          • Krogoth
          • 15 years ago

          A64s and 1xx Opteron systems need at least a HTT speed of 600 to function well. ~800MHz is good for optimal function. 1Ghz quite frankly is overkill and pure marketing nonsense at this point.

          Anyway, HTT speed determines how much bandwidth your perpherials have. Until the chipset guys and AMD decide to add in more PCIe lines there’s no point to having a HTT speed greater then 800Mhz.

            • Lazier_Said
            • 15 years ago

            All the benchmarks I have seen suggest that in the usual benchmarked applications there is no performance difference between a 400mhz and a 1200mhz HTT setting. And even running it at 200mhz the performance loss is too small to notice without a benchmark.

            • Krogoth
            • 15 years ago

            Because, HTT bus determines the amount of bandwidth that is avaiable to the perperhrials (PCI, PCIe buses and intergrated stuff). Game framerates are the last thing to be sigificantly impacted by it. You’ll notice the difference in the system’s overall responsiviness to heavy bandwidth loads.

            I had already tried a 200Mhz speed and noticed that it took 10-20 more seconds for XP64 to load-up and any sigificantly I/O load made the system very sluggish compaired to my current HT speed of 833Mhz.

      • flip-mode
      • 15 years ago

      I’m actually more disappointed over the fact that TR didn’t throw in a more overclockable CPU so they could really make a credible statement about the motherboard. I’m guessing it’s because there wasn’t another CPU available, but damn! I don’t know, that just kinda let me down ๐Ÿ™

      Also, Geoff, you mention that the chipset cooler is close enough to create problems with wider heatsinks like the 7700. Could you clarify that, as far as, did you actually place a 7700 on the mobo or are you eyeballing it?

        • Dissonance
        • 15 years ago

        I actually put one on the board. And a CNPS9500, as well.

          • flip-mode
          • 15 years ago

          I see, bummer. Thanks for the reply.

    • DrDillyBar
    • 15 years ago

    Liking that slotted x4 slot… thing.
    Edit: Praise be to non-AC97 audio!

      • kvndoom
      • 15 years ago

      Death to the blue crab!!!

Pin It on Pinterest

Share This