Original Link: https://www.anandtech.com/show/2375

Intel X48 - Much Ado about Nothing....

by Gary Key on November 14, 2007 9:00 AM EST


The launch of the Intel X38 chipset occurred just a few weeks ago and even by the most optimistic viewpoint, it was nowhere as successful as the P35 rollout last spring. In fact, we can honestly say this chipset launch is just about the only blemish on Intel's almost flawless product release schedule this year. We had grown comfortable over the past year with Intel's continuing evolution and roll out products, each one offering something distinctive whether it was performance related or a cost reduction that reset the price to performance ratios in an already very competitive market. The hype surrounding the X38 for the past several months had labeled it as the pinnacle in performance chipsets. From all early indications, it was going to be just that and then boom; it seemed like a major disappointment upon release.

The performance of the X38 was clearly not a step above the P35 upon release. Test results from new features such as PCI Express 2.0 and true dual x16 PCI Express capable slots were not available. The lack of performance oriented GPUs based upon PCI Express 2.0 from AMD meant CrossFire with the HD 2900XTs were not an alternative to NVIDIA's SLI technology featuring the class leading 8800GTX. Even without those new features, the whole launch process seemed unorganized with only a couple of motherboards available at launch. Even several weeks after launch, we are just now seeing boards from manufacturers other than ASUS and Gigabyte hitting the channels. It was not until this last series of BIOS releases that we would even consider purchasing an X38 based board over a P35 equivalent. It appears from the number of boards in the channel that many of you agreed with us.

Another kink in the X38 launch process centered on the continuing rumors of the upcoming X48 chipset. Once again, the hype machine or maybe more like the rumor machine got into full gear with claims touting the X48's greatly improved performance. If some of the rumors were true, it would soon wear the performance crown. Of course, a lack of official information from Intel about the chipset only propagated the continuing rumors around the X48 and the apparent demise of the X38 in short order. To complicate the situation, the release schedules for the X48 have bounced around from November to February and even at this point, we are still unsure.


What we are sure of at this point is that we have a mess on our hands. However, after testing the X48 for the past few days several areas of concern with the whole X38/X48 situation are clear to us now. The X48 is an upgrade to the X38 and not a replacement according to the information we have currently. Even with this statement, the motherboard manufacturers are still unsure at this point, as to how to market the chipsets or if they will even coexist in certain lineups. We see no reason as to why the X48 cannot be marketed into the very top end with the X38 replacing the upper tier P35 boards in the $150~$250 market. The X38 in our opinion is maturing quickly and we will see it replace the P35 at the top end, especially for those wanting a CrossFire setup.

With that in mind, we are going to provide a very early look at the X48 chipset and a few benchmarks that show its potential before spending the next few days concentrating on wide range of products from AMD. We will look at the X48 and DDR2 performance along with several X38 motherboards afterwards. Let us look at the feature set of the ASUS P5E3 Premium and check out its preliminary performance.



ASUS P5E3 Premium: Features and BIOS Changes

ASUS P5E3 Premium
Market Segment High Performance Enthusiast - $299.99 (estimated)
CPU Interface Socket T (Socket 775)
CPU Support LGA775-based Core 2 Duo, Core 2 Extreme, or Core 2 Quad Recommended
Chipset Intel X48 (MCH) Northbridge and ICH9R Southbridge
FSB Speeds Auto, 200 ~ 800 in 1MHz increments
Memory Speed (DDR3) Auto, Eight Ratios dependent upon Strap Setting
FSB Strap Auto, 200, 266, 333, 400
PCIe Speeds Auto, 100MHz - 180MHz
PCI Speeds Locked at 33.33MHz
Core Voltage Auto, 1.10000 to 1.70000 in .00625 increments, plus settings to 2.10V with OC jumper enabled.
CPU Clock Multiplier 6x ~ 12x, downward adjustable for Core 2, upward to 31 for Extreme
DRAM Voltage (DDR3) Auto, 1.50V ~ 2.78V in .02V increments, 1.50V standard
DRAM Timing Control Auto, Manual - 23 DRAM Timing Options (tCL, tRCD, tRP, tRAS, tRFC + 18 sub-timings)
DRAM Command Rate Auto, 1N, 2N
NB Voltage Auto, 1.25V ~ 1.91V in .02V increments, plus settings to 2.21V with OC jumper enabled.
SB Voltage Auto, 1.05V, 1.20V, 1.05V standard
FSB Termination Voltage Auto, 1.20V to 1.50V in .02V increments, 1.20V standard
Clock Over Charging Auto, .70V to 1.00V in .10V increments, .80V standard
CPU Voltage Damper Auto, Enabled, Disabled
CPU Voltage Reference Auto, .67, .65, .63, .62
NB Voltage Reference Auto, .67, .61
AI Clock Twister Auto, Light, Medium, Strong
DRAM CLK Skew CA/CB Auto, Normal, 50ps to 350ps either advance or decrease
DRAM CMD Skew CA/CB Auto, Normal, 25ps to 175ps either advance or decrease
AI Transaction Booster (tRD) Auto, Manual, settings from 31 to 1
Memory Slots Four 240-pin DDR3 DIMM Slots
Dual-Channel Configuration
Regular Unbuffered DDR3 Memory to 8GB Total
Expansion Slots 3 - PCIe x16 (2 - x16, 1 - x4 electrical)
2 - PCIe x1
2 - PCI Slot 2.2
Onboard SATA/RAID 6 SATA 3Gbps Ports - ICH9R
(RAID 0,1, 10, 5)
2 eSATA 3Gbps Port - JMicron JMB363
Onboard IDE 1 ATA133/100/66 Port (2 drives) - JMicron JMB363
Onboard USB 2.0/IEEE-1394 10 USB 2.0 Ports - 6 I/O Panel - 4 via Headers
2 FireWire 400 Ports by Agere FW3227 - 1 I/O Panel, 1 via Header
Onboard LAN Realtek RTL8110SC - PCI Gigabit Ethernet controller
Marvell 88E8056 PCI Express Gigabit Ethernet controller
ASUS 802.11n Wireless Port
Onboard Audio ADI 1988B - 8-channel HD audio codec
Power Connectors ATX 24-pin, 8-pin ATX 12V
I/O Panel 1 x PS/2 Keyboard
2 x eSATA
2 x SPDIF - Optical Out, Coaxial Out
1 x IEEE 1394
1 x Audio Panel
2 x RJ45
6 x USB 2.0/1.1
Fan Headers 6 - CPU, (5) Chassis
Fan Control CPU and Chassis Fan Control via BIOS/AI Suite, PC Probe II monitoring
BIOS Revision v110
Board Revision v1.00

The P5E3 Premium board is designed for the general enthusiast with an emphasis on features designed for the home user, whether it is for gaming or home theater usage. ASUS provides two eSATA ports via the JMicron JMB363 chipset, IEEE 1394 support from Agere, excellent on-board audio support from the ADI 1988B, an 802.11n Wireless port, and dual Gigabit LAN capability.

The board offers a very good mix of expansion slots although utilizing a CrossFire setup will create the physical loss of a PCI Express x1 and PCI slot. The third PCI Express x16 slot operates at x4 electrically and can be utilized for various PCI Express cards including AGEIA PhysX, RAID controllers, or a third video card at this time.

We will go into the BIOS settings in additional detail once we receive the retail kit and official launch BIOS. However, the following screenshots represent features that we know will be in the retail BIOS.





The P5E3 Premium BIOS has undergone a makeover of sorts. The AI Tweaker section is still in place with a wide variety of options for tuning the board. Although not as extensive as the latest DFI enthusiast boards, the options available have been expanded. The notable addition being the revamped the Transaction Booster setting. The Transaction Booster option is actually another name for tRD (Performance Level) and can be adjusted from 31 down to 1. ASUS joins DFI now in offering this setting that greatly affects latencies and overall system performance. The DRAM timing control sections now offers 23 different memory settings.

Our screenshots show the maximum value for each option. ASUS has listened to the power users and now offers CPU voltages to 2.10V and MCH voltages to 2.21V provided the OC jumper on the motherboard is enabled. ASUS has also included the ability to save BIOS profiles, although it is currently limited to two profiles.



Test Setup

ASUS P5E3 Deluxe / P5E3 Premium Testbed
Processor Intel Core 2 Quad QX9650
Quad Core, 3.0GHz, 2x6MB Unified Cache, 9x Multiplier, 1333FSB
CPU Voltage 1.250V Stock
Cooling CoolIT Freezone
Power Supply OCZ 1000W
Memory Corsair CM3X1024-1800C7D
Memory Settings 5-5-5-15 (DDR3-1333)
Video Cards MSI HD X2900 XT 512MB
Video Drivers ATI Catalyst 7.10
Hard Drive Western Digital 7200RPM 750GB SATA 3/Gbps 16MB Buffer
Optical Drives Plextor PX-B900A, Toshiba SD-H802A
Case Cooler Master Stacker 830 Evo
BIOS Deluxe - 601 / Premium 110
Operating System Windows Vista Ultimate 64-bit
.

Test conditions were maintained the same, as much as possible, over the platforms tested. Our game test was run at a setting of 1280x1024 HQ to ensure our MSI HD 2900 XT is not a bottleneck.

We selected the Intel Core 2 Quad QX9650 as our processor of choice since it allows us several different choices with the front side bus settings and represents Intel's latest CPU offering, well for now. We are utilizing Microsoft Vista Ultimate 64-bit as our operating system along with a 2GB memory configuration for this preview. We will also provide Vista 64-bit results with 4GB of memory in the full review. Our test results were not adversely affected with 4GB of memory. We did notice stability problems when overclocking the memory past DDR3-1800 with 4GB installed. ASUS is currently working on this and we attribute these problems to a very early BIOS release.

We utilize new drive images on each board in order to minimize any potential driver conflicts. Our 3DMark results are generated utilizing the standard benchmark resolution for each program. We run each benchmark five times, throw out the two low and high scores, and report the remaining score. All results at stock speeds for this article are with memory timings at 5-5-5-15 (DDR3-1333) and at 6-6-6-15 (DDR3-1600) for our overclocking tests. Where possible, memory sub-timings were set exactly the same to ensure consistency between the boards.

Our choice of software applications to test is based on programs that enjoy widespread usage and produce repeatable and consistent results during testing. Microsoft Vista has thrown a monkey wrench into testing as the aggressive nature of the operating system to constantly optimize application loading and retrieval from memory or the storage system presents some interesting obstacles. This along with what we still see as a lack of driver maturity will continue to present problems in the near future with benchmark selections. Our normal process is to change our power settings to performance, delete the contents of the Prefetch folder, and then reboot after each benchmark run. This process results in consistency over the course of benchmark testing. All applications are run with administer privileges.



Memory Testing-


There are no real differences between the X38 equipped Deluxe board and the X48 based Premium in memory performance at like settings. Except for the ability to set tRD levels in the BIOS on the Premium board instead of using MemSet, we could fine-tune our memory performance to the same levels on either board. The additional five memory settings in the Premium BIOS allowed some additional tweaking for improved latencies. However, the differences in scores are within the standard variation percentages for benchmark testing.

Memory Performance Comparisons-

QX9650 9x333



Our stock memory results are preliminary considering the beta nature of the 110 BIOS on the Premium board. Once again, there are not any noticeable differences between the two chipsets at like settings.



Application Testing-


Our application results mimic the memory scores. At like settings, there are not any measurable differences between the chipsets. The slight latency advantage of the Premium board generates minor improvements in the benchmark results, something not noticed during actual system usage.

FSB Overclock Levels-

QX9650 11x410



QX9650 9x465


Our top screenshots show our initial results when pushing the CPU and memory to an 11x410 setting. The Premium board allowed us to set tRD at 5 and tRAS at 15 compared to 6 (using MemSet) and 18 on the Deluxe board. The Premium board defaulted to a tRD setting of 7 and the Deluxe to a default setting of 8 when utilizing the auto option. Other sub-timings were set slightly lower on the Premium board. The biggest difference between the two boards was the CPU and MCH voltage settings. The Premium board required 1.59375V for the CPU and 1.69V for the MCH while the Deluxe board required 1.64375V and 1.73V respectively. Whether these improvements are due to the BIOS or MCH improvements is up for debate currently. Based on preliminary results with other processors, we think it is a combination of both at this time.

Our bottom screenshot shows the maximum FSB rate we have reached with this particular CPU on the Premium board. Our results on the Deluxe board are almost identical down to the latency score. We are currently testing additional processors and will provide those results in the full review. It appears at this time that our selection of processors all reach the same FSB limits on either board. However, the voltages required and the memory settings allowed all show a slight advantage to the X48 equipped Premium board.



First Thoughts-

Our initial test results indicate that except for official 1600FSB support, there are not any real performance differences between the X38 and X48 chipsets at stock or mildly overclocked settings. This does not surprise us, as Intel's chipset guidance outline has not indicated anything differently. It was not until we started pushing both boards that we could tell a difference, even then it was minor for the most part. The X48 based Premium board did allow us to reduce CPU and MCH voltages when compared to the Deluxe board. This could be a combination of factors ranging from BIOS and board design to improvements in the X48 chipset. In early testing, it is proving difficult to isolate the specific differences. We believe it is a combination of BIOS design and the X48 chipset. However, it appears on the DDR3 boards that the user will not be able to take advantage of these minor differences unless they are really pushing the board. This will be important for some but for the majority of users it will not matter which chipset is used.


We cannot provide benchmark results with the upcoming 3.2GHz QX9770 quad-core processor that features a native 8x400FSB setting at this time. We can say in preliminary testing that both ASUS boards generated almost identical benchmark results at stock settings with the X38 Deluxe board having no problems with the native 400FSB setting. However, the X48 Premium board provided minor improvements in memory performance when overclocking past the 1800MHz level.

Overall, we can state that with the latest BIOS releases the X38 boards are maturing quickly and the transition to the X48 will be seamless for the manufacturers and potential users. We are being told that the X48 will command a price premium (average $30~$50) over the X38 boards. In our early testing, that premium only seems justified for those who extract 98%+ out of their boards or for those who want the security of knowing their 1600MHz FSB capable processors will operate properly without the supplier having to guarantee 1600MHz compatibility on the X38 product. The latter statement is one that we have not experienced yet with any X38 board. Our only open question at this time is if DDR2 performance has improved with the X48 as rumor says it will. We will answer that question shortly with the GA-X48-DQ6 board from Gigabyte.

In regards to the ASUS P5E3 Premium board, it appears to us that ASUS fixed our few qualms with the BIOS design and options on the Deluxe board. Both boards pack a tremendous number of features on-board and provide excellent, if not class leading performance, provided you make the jump to DDR3. Although they are expensive, we believe both boards set the standard in the DDR3 arena. We look forward to providing final test results shortly with both boards. For now, it appears the hype surrounding the X48 chipset was overblown. While it is an improvement in certain cases, we have not seen it clearly outperforming the X38 yet. Maybe some additional octane from the upcoming Penryn lineup will provide additional horsepower, but for now, we have two almost equal performing engines under the hood.

Log in

Don't have an account? Sign up now