by Doors4ever
05/08/2008 | 09:05 PM
Mainboards on Intel X48 Express chipset spread all over the place slowly but surely. Slowly, because the only significant difference between the new chipset and its predecessor is high price, although even Intel X38 Express already cost quite a bit. So, the customers are not really rushing to stores to get their hands on the new mainboards and there is no buying frenzy of any kind. However, the mainboards and the buyers are there alright. The manufacturers, just like fortune tellers, realized a while ago that the magic of numbers has serious effect on people. From the customers’ standpoint, X48 core logic is indisputably better than X38, just because it is a higher number. The graphics card developers have taken advantage of this human perception peculiarity for a while: just change the working frequencies a little bit and give the card a new name and it will immediately become more valuable from the consumer’s prospective. Although you may even leave the frequencies untouched: just change the name and it will do the trick.
<%BANNER[article]%>Now it’s the mainboard manufacturers’ turn. Launching a new product has never been easier. Just replace X38 chip with X48 on the old mainboard and you will get a new product right there. So, we see a lot of new boards being dead ringers for the old ones around. And today we are going to take a closer look at DFI LANPARTY LT X48-T2R, which is practically identical to DFI LANPARTY LT X38-T2R, which we have already reviewed before. It is really bring to talk and read about exact same things all over again, so to spice things up I suggest playing a game. We will be looking for differences between the two mainboards. What if in the end they turn out pretty significant despite the seeming similarity?
Even if you can’t read you will immediately notice that the package design of the new DFI LANPARTY LT X48-T2R board is different. Although they used the same style and design principles, the front of the box only bears the series name, while the list of main mainboard features is on the back of the box.
I agree that it is a small difference and under other circumstances we may not have even mentioned it. However, the two mainboards have so few differences that we will be counting in even the smallest ones.
As for the accessories bundle, it remained exactly the same. All accessories are packed into a separate box. DFI LANPARTY LT X48-T2R still comes with FDD and IDE cables, four Serial ATA cables and SATA power adapters, I/O Shield for the case rear panel, user’s manual, CD disk with drivers and utilities, a floppy disk with RAID drivers, colorful installation guide for LGA 775 processors, four replacement jumpers with convenient long “tails”.
These are standard accessories that you may find bundled with any other mainboard. However, there is also something unique in the bundle. The chipset North Bridge cooler is too tall to be shipped preinstalled, so it comes separately with an additional installation manual, a syringe with the thermal compound and even a special card for even spreading of the compound on the chipset heat-spreader. According to DFI web-site, this cooler was designed y Thermalright that is why its superb efficiency pointed out in our previous articles comes as no surprise to us.
One of the comments we have about this cooling system is its huge size. Some large processor coolers may simply not fit next to it. However, this time we can name one more drawback. These cooling systems used to be fastened to DFI mainboards with secure steel brackets at four spots. For example, the cooler on DFI LANPARTY UT NF680i LT SLI-T2R.
Some time ago four secure steel hooks have been replaced with two weak plastic “spindles”:
No wonder that when we completed testing and took the system apart, one of these plastic clips broke.
The two mainboards we are comparing today look absolutely identical that is why everything we have already said about DFI LANPARTY LT X38-T2R may be repeated again about the new DFI LANPARTY LT X48-T2R.
As for minor drawbacks, we could mention not the best location for the FDD connector and the fact that the graphics card installed into the first PCI Express x16 slot will block the DIMM slot clips. We could also wish the board had a few external eSATA ports: all eight ports including those six that are provided by the ICH9R South Bridge and the other two from JMicron JMB363 chip are internal ones.
All other features of the DFI LANPARTY LT X48-T2R mainboard are indisputable advantages. It boasts eight-phase digital processor voltage regulator, very conveniently located 24-pin power connector and 8-pin ATX12V connector, solid-state capacitors, two second generation PCI Express x16 and an additional PCI-E x16 slot working at x4 speed, three PCI and one PCI-E x1 slots, six fan connectors (a four-pin one for the CPU fan), 12 USB 2.0 ports (six on the rear panel, two gigabit RJ45 connectors (Marvell 88E8052 and Marvell 88E8053), two-section POST indicator, Power On and Reset buttons on the PCB, color coded connectors and small additional heatsinks on all heating components.
We usually consider numerous jumpers, especially those that serve to change the FSB frequency, a drawback, but in this case they are necessary. Moreover, while many mainboard manufacturers usually forget about a jumper that allows disabling the integrated PC-speaker, DFI did implement it. Even such a minor thing as a Clear CMOS jumper, which is marked red adds another point to the company’s good name. By the way, you don’t have to use this jumper at all: all settings may be cleared by pressing and holding Power On and Reset buttons simultaneously.
Speaking of the accessories we forgot to mention one more important thing: Bernstein audio module:
It is based on an eight-channel Realtek ALC885 codec. And the peculiarity of this solution is that it is not installed into a special slot on the mainboard PCB, but is connected with a flexible cable, which means that you can move it around the back panel of your system case as far as the cable allows.
As for the differences from the previous model, the only one is the new chipset North Bridge: X48.
The most attention readers may have already noticed that “X48” is not written on the PCB, but on the sticker:
It may be hiding the old “X38” beneath it. Let’s try taking a peak under the sticker now:
Hm, nothing there. Ok… To prove that both mainboards have identical electrical CB layout we suggest taking a look at the components schematics, where the only difference is the chipset North Bridge chip:
Summing up our exterior examination of the DFI LANPARTY LT X48-T2R mainboard we would like to offer you a complete list of the mainboard specifications, where the only difference would be official support of 400 (1600) MHz FSB.
In reality it is not a factual difference, but a pure formality. BIOS already had dividers that allowed setting the memory at 800 or 1066MHz at 400MHz FSB, and they are still there. However, we are going to talk more about BIOS in the next part of our article.
DFI LANPARTY LT X48-T2R mainboard BIOS is based on AMI code and looks exactly as before. However, a few changes have been made to it and unfortunately, not all of them are for the best.
Let’s start with the good things. Now there is an option to disable the additional IEEE1394 controller if it is not used. For some strange reason it was absent before.
PC Health Status section remained the same. The mainboard’s distinguishing feature is Adjust CPU Temp option that allows adjusting the processor temperature. The option to adjust the rotation speed of three fans out of six that can be connected to the mainboard can be regarded as a drawback and as an advantage at the same time. However, the board’s inability to control the rotation speed of a three-pin processor fan is a definite low. Moreover, the mainboard allows controlling a wide range of voltages, three temperatures including the temperature of the processor voltage regulator area and rotation speeds of three fans.
Genie BIOS Setting section contains all overclocking related options including adjustment of frequencies, timings and voltages.
I was very inclined to list new CPU N/2 Ratio parameter as one of the advantages. It allows setting fractional clock frequency multipliers for the processor. However, you get access to this parameter only if there is a 45nm CPU from Penryn family supporting multipliers like that. If you use a 65nm Conroe, this parameter disappears from the BIOS Setup.
Genie BIOS Setting consists of three sub-sections. CPU Feature deals with the processor related technologies:
The list of options in the DRAM Timing sub-section is so long that it couldn’t fit into a single screen: the photo is put together from two screenshots:
But that is still not everything. Clock Setting Fine Delay and Read Delay Phase Adjust lead to separate pages with more additional settings, however, we have already seen it all. All the differences are in the Voltage Setting sub-section:
The section looks formally the same as before, the list of available parameters remained unchanged. The only thing that catches your eye is the CPU VID Special Add parameter highlighted red. It is the one, although red color doesn’t really matter. If you use a Conroe processor and not Penryn, it will not be red anymore, it will simply be inactive, i.e. its working principle remains different.
I would like to remind you that processor Vcore may be set using two parameters: CPU VID Control and CPU VID Special Add. They are not equal. CPU VID Control parameter sets the Vcore in an explicit form with a small increment of 0.00625V. In this case the mainboard will maintain the Vcore at the set rate even if the CPU is idling. CPU VID Special Add sets the voltage in percents from the nominal. It I not very convenient that you need to imagine what the resulting voltage is going to be, however, the current Vcore is displayed on the same page, so after a few attempts you will be able to set the desired voltage. In this case the CPU will drop the clock frequency multiplier to x6 and reduce the Vcore in idle mode and if Intel power-saving technologies are activated.
Both parameters used to work independently before. You could set the Vcore using CPU VID Control or using CPU VID Special Add. Now CPU VID Special Add parameter depends on CPU VID Control for some reason. It doesn’t work until you set the Vcore above 1.2V for 45nm CPUs or above 1.3V for 65nm CPUs. In other words, the whole procedure looks as follows: at first we increase CPU VID Control to 1.20625V or 1.30625V (depending n the CPU we use). Then we can change the CPU VID Special Add and set the desired voltage in percents. After that we can return CPU VID Control parameter to Auto again. Reboot the system, check what resulting voltage you’ve got and repeat the whole thing again if necessary.
Very inconvenient, to say the least of it. I don’t know why they had to make things so complicated. And most importantly, it makes overclocking harder and doesn’t allow using power-saving technologies in most cases. We are going to discuss a few examples in the next part of our article, and here we would only like to state that we managed to find two differences in the new mainboard’s BIOS: one good (an option to disable FireWire) and one bad (increasing processor Vcore has become more complicated). Everything else remained the same.
We are going to check the overclocking potential of the new DFI LANPARTY LT X48-T2R mainboard in an open testbed configured as follows:
This is the exact same testbed as the one we used for our DFI LANPARTY LT X38-T2R review. However, that X38 based board was lucky, if we can put it this way: at that time we used Intel Core 2 Duo E6300 and Intel Core 2 Extreme QX6700 processors that are pretty easy to overclock for any more or less decent overclocker mainboard. DFI LANPARTY LT X48-T2R mainboard is facing a more serious task. At first it should pass the test with an Intel Core 2 Duo E8400 processor (3.0GHz, 333MHz FSB, 6MB, Wolfdale, rev. C0).
Actually, this processor is also not too difficult to overclock: you just need to increase the voltage, raise the FSB frequency to 450-455MHz and enjoy the result. Most mainboards can do it just fine. Overclocking by lowering the processor clock frequency multiplier will show the actual overclocking potential of the mainboard: it would be interesting to see how far up the FSB will go in this case. Contemporary mainboards can usually work at 520-530MHz FSB. So, we increase the Vcore of the Intel Core 2 Duo E8400 processor, CPU VTT Voltage, chipset North Bridge voltage, lower the clock frequency multiplier and try booting at 530MHz FSB. No luck. Then we try the 520MHz frequency… Then 510MHz… And even at 500MHz the board is not running really stably.
During our experiments we found out that DFI LANPARTY LT X48-T2R mainboard doesn’t detect over-overclocking that well. The board would frequently loop at the reboot attempt: the board would start, we see the first data on the screen with the CPU name and then the board would reboot on its own and keep doing this over and over. On those rare occasions when the system rebooted in safe mode and reported over-overclocking, you could still do nothing, because the keyboard didn’t work.
Actually, DFI LANPARTY LT X48-T2R mainboard is one of the “correct” contemporary mainboards with USB keyboard support enabled by default, but in this case something didn’t work right. That is why if you are experiencing any problems upon system restart, it is much simpler and faster to Clear CMOS than wait for the mainboard to respond to over-overclocking (which would often be the wrong response anyway). Luckily, brilliantly implemented CMOS Reloaded technology makes up a little bit for this unpleasant situation and the need to clear the settings. It allows saving all BIOS settings, provide the profile with a detailed description and load it quickly and easily when needed.
While we tried to achieve stability, our OS got corrupted. We increased the voltages, changed the FSB and memory frequencies, but the board wouldn’t boot Windows Vista at all, or would boot it but then immediately report inability to start some services, put on the blue screen of death or reboot on its own soon after Prime95 stability check started. That is why it took us a while to notice that the “Start” button didn’t work: couldn’t be clicked with a mouse pointer or with a keyboard key. So, maybe DFI LANPARTY LT X48-T2R mainboard an in fact work at 500MHz FSB, but it is the corrupt OS that causes all problems? However, the very first attempt to boot the freshly reinstalled Windows Vista at 510MHz FSB ended up with a familiar message that Windows Explorer was not responding.
Unfortunately, we couldn’t succeed in overclocking a dual-core CPU on DFI LANPARTY LT X48-T2R mainboard. But maybe the board would do well with a quad-core processor? We decided to check things out using a new weapon: the youngest quad-core 45nm Intel Core 2 Quad Q9300 (2.5GHz, 333MHz FSB, 6MB, Yorkfield, rev. M1).
This is a new processor, we haven’t yet studied its overclocking potential very carefully. So far we have only tested it on abit IP35 Pro mainboard. Yes, this is no mistake: on abit IP35 Pro. The thing is that abit has finally released a new BIOS version 1.6 for their IP35 Pro mainboard that allows using fractional clock frequency multipliers for 45nm processors. However, despite the long six months it took them to do this, the new BIOS version looks more like a beta than a complete release.
To check out the processor features without hitting against its overclocking potential, we reduced its clock frequency multiplier to the minimal x6. However, when we booted Windows, we often discovered that for some reason the multiplier was set at x7 or even at the maximum possible x7.5! It turned out that abit IP35 Pro mainboard may change the processor multiplier on its own at any time. It was pretty absurd, when we set all parameters including the x6 multiplier in the BIOS Setup, pressed F10 (save and quit) and when the system wanted us to confirm the operation returned to the previous page to see that the multiplier has already been changed to x7!
Of course, this unpredicted behavior of abit IP35 Pro mainboard made things more complicated. But it could really overclock processors, we can’ take that away from it. Despite all the problems, we managed to get Intel Core 2 Quad Q9300 to work at 475MHz FSB with the default multiplier of x7.5. This result doesn’t look final, I don’t feel like saying that our CPU overclocked to 3562.5MHz (475x7.5). 3.6GHz looks much more attractive and can be reached at 480MHz FSB. So we will go for it.
Unfortunately, DFI LANPARTY LT X48-T2R mainboard didn’t help us to overclock our Intel Core 2 Quad Q9300 to 3.6GHz. It didn’t have any problems with spontaneous multiplier increase, but at 475MHz FSB with the x6 multiplier it could only start, but failed to boot Windows. The same happened at 465MHz FSB and even 460MHz FSB. We managed to start Windows at 460MHz FSB when we enabled GTL REF Voltage Control in the mainboard BIOS, but the system still froze dead when we launched Prime95 stability check. Only when we dropped the bus frequency to 455MHz, we could work normally for some time, which is hardly a satisfactory result for an overclocker solution.
By the way, you may have noticed that our Intel Core 2 Quad Q9300 processor features impressively low default Vcore of only 1.17V or 1.175V. We couldn’t be happier that our quad-core processor worked at pretty high frequencies with such low nominal voltage setting, but only not with DFI LANPARTY LT X48-T2R mainboard. You remember that CPU VID Special Add parameter doesn’t work until you increase CPU VID Control above 1.2V with a 45nm CPU, don’t you? But our processor has even lower nominal voltage. We increase CPU VID Control, set CPU VID Special Add at the desired rate, lower CPU VID Control to the nominal, but since 1.17V is lower than the required 1.2V CPU VID Special Add will be reset to Auto.
As a result, it turns out that we cannot use CPU VID Special Add to overclock our processor as well as any other 45nm processor with nominal Vcore below 1.2V and a 65nm processor with nominal Vcore below 1.3V. We will have to use CPU VID Control to increase the processor core voltage, i.e. it will not drop in idle mode, Intel power-saving technologies will not work, the processor fan will have to rotate at higher speed all the time.
We won’t claim that DFI LANPARTY LT X38-T2R mainboard is better than the new DFI LANPARTY LT X48-T2R one. We did remind you that we tested it in less aggressive conditions. Maybe it would have revealed some issues as well if it ever had to deal with a 45nm Intel Core 2 Duo E8400 and Intel Core 2 Quad Q9300 processors. We are not going to guess here. Yes, DFI LANPARTY LT X38-T2R is not an ideal mainboard, it does have a few drawbacks of its own, but we have to admit that the new DFI LANPARTY LT X48-T2R made not the best impression we were hoping for. It seems to be the last, though very subjective difference between the two mainboards.
By the way, we do not claim to be the only universal truth tellers here. Maybe you will find more differences of your own. In this case the winner of our today’s game will be the one who finds 6+m differences. No, you won’t be able to do it, try to spot at least 6+n and win!