Asus P5AD2 Premium: Overclocking and Stress Testing

FSB Overclocking Results

Front Side Bus Overclocking Testbed
Processor: Pentium 4 Prescott LGA 775
2.8GHz
CPU Voltage: 1.3875V (default)
Cooling: Thermaltake Jungle 502
Power Supply: HiPro 470W
Maximum OC: 280FSB (+40%)

The LGA 775 CPUs are multiplier locked, so the only way to overclock the CPU is to increase the FSB. Fixing the PCIe frequency tends to limit the overclocking of the 925X, so Asus, like others in the roundup, allows the PCI Express frequency to float. Asus then manipulates the PCIe frequency (or ratios to be more precise) during boot to achieve higher overclocks. For more information on how Asus manages to reach such outstanding overclock with the P5AD2, please check Breaking Intel's Overclock Lock: The REAL Story.

The Asus was able to reach a FSB of 280, a 40% overclock, with an ATI X800 XT and SATA hard drive. This is by far the highest overclock that we achieved with this hardware in the roundup. The P5AD2 was more limited with an nVidia 6800 Ultra video card, reaching 258. The nVidia appears to be less tolerant of out-of-spec PCIe frequency than the X800 XT.

Memory Stress Test Results:

The memory stress test measures the ability of the Asus P5AD2 to operate at its officially supported memory frequency (533MHz DDR2), at the best performing memory timings that Crucial/Micron PC2-4300U will support. Memory stress testing was conducted by running DDR2 at 533MHz (stock 3:4 ratio) with 2 DIMM slots operating in Dual-Channel mode.

Stable DDR533 Timings - 2 DIMMs
(2/4 DIMMs - 1 Dual-Channel Bank)
Clock Speed: 266MHz
Timing Mode: 3:4 (200:266 - Default)
CAS Latency: 3.0
Bank Interleave: Auto
RAS to CAS Delay: 3
RAS Precharge: 3
Cycle Time (tRAS): 10*
Command Rate: N/A
*SPD (Auto) timings for DDR2 are normally 4-4-4-12 at DDR2-533. A tRAS setting of 12 is normal. We ran a series of tests to measure memory bandwidth, and found the tRAS setting made very little difference in the performance of DDR2. The most effective range of tRAS was 8 to 13 for DDR2 on the 925X chipset, so a tRAS of 10 was chosen for benchmarking.

The Asus P5AD2 Premium was completely stable with 2 DDR2 modules in Dual-Channel at the settings of 3-3-3-10 at 1.8V default voltage. This matches the best timings that all the motherboards in this roundup were able to achieve with DDR2 memory. 3-3-3 is also faster than the specified 4-4-4 timings specified for the Crucial/Micron DDR2 modules used for these benchmarks.

Filling all four available memory slots is more strenuous on the memory sub-system than testing 2 DDR2 modules on a motherboard.

Stable DDR533 Timings - 4 DIMMs
(4/4 DIMMs - 2 Dual-Channel Banks)
Clock Speed: 266MHz
Timing Mode: N/A
CAS Latency: 4.0
Bank Interleave: N/A
RAS to CAS Delay: 3
RAS Precharge: 10T*
Precharge Delay: 3.0
Command Rate: N/A
*SPD (Auto) timings for DDR2 are normally 4-4-4-12 at DDR2-533. A tRAS setting of 12 is normal. We ran a series of tests to measure memory bandwidth, and found the tRAS setting made very little difference in the performance of DDR2. The most effective range of tRAS was 8 to 13 for DDR2 on the 925X chipset, so a tRAS of 10 was chosen for benchmarking.

As we first saw on the Abit, 4 DDR2 DIMMs are not stable at CAS 3 as 2 DIMMs are. Once again, we required 4-3-3 timings when using 4 DDR2 DIMMs, though 4-3-3 worked fine at default voltage. Perhaps more voltage could improve stability at 3-3-3 with 4 DIMMs, but our tests up to 2.0V (from default 1.8V) did not produce enough stability to complete our Super Pi or Aquamark 3 at the 3-3-3 timings.

Asus P5AD2 Premium: Features and Layout DFI LANParty 925X-T2: Features and Layout
Comments Locked

30 Comments

View All Comments

  • jdoor0 - Tuesday, October 26, 2004 - link

    This review has been reviewed:
    http://www.theinquirer.net/?article=18896
  • Nige - Tuesday, October 12, 2004 - link

    Does the ASUS P5AD2 Deluxe motherboard have the same overclocking capability as the P5AD2 Premium?
  • skiboysteve - Friday, August 13, 2004 - link

    Wow nice catch. i guess my "(I know... toms sucks)" disclaimer came true.


    yaeh i understand.
  • Wesley Fink - Thursday, August 12, 2004 - link

    #26 -
    There is now an apology to Asus up at THG. They measured the voltage wrong. We had also measured the voltage and found 1.5 to 1.55 which is well within spec, not 2.1 as they reported. They now acknowledge the correct voltage measurement for the P5AD2 is 1.53V.

    High Northbridge voltage is not the reason the Asus, or any other 925X/915 board, overclocks well. There are far too many simple and wrong explanations for the complex overclocking issues of the 925X/915 chipsets.
  • skiboysteve - Thursday, August 12, 2004 - link

    Page 10
    "...Broadcom attached to the faster PCI Express bus..."

    there is no PCI Express bus, its a point to point protocal.

    Just nitpicking.

    Great review.



    Also, over at Toms (I know... toms sucks) they looked at 9x5 Boards over there and showed that the Asus P5AD2 was running at an astounding 2.1v on the northbridge (1.5v is the stock)

    Something might have to be mentioned about reliability of such out of spec behavior, and cooling concerns. You might want to conduct your own quick test on the voltage with a multimeter.
  • broberts - Thursday, August 12, 2004 - link

    One of the problems with these arguments is that the FX-53 is almost 20% more expensive.

    I've been thinking for a while now that benchmarks should show some form of pricing index so that one can better judge the advantage/disadvantage of the various choices. Just quoting prices isn't ideal, for a host of reasons. I'd suggest, instead, a relative measure. And not just the cost of the particular component being benchmarked. Calculate the cost of the each system used in the benchmarks. Pick one, perhaps the lowest or highest cost one and calculate the relative difference in price. I suggest using the entire system because quite often the choice of one component dictates the available choices for other components. Ideally a relative measure for both the components and entire system would be calculated and published.
  • manno - Thursday, August 12, 2004 - link

    Moo Moo MOO.
  • manno - Thursday, August 12, 2004 - link

    why no Doom3?
  • Wesley Fink - Thursday, August 12, 2004 - link

    #21 -
    We will definitely be including Doom 3 benches in future reviews. The only reason they are not included in this 925X roundup is because most of the testing was completed before we had a working copy of Doom 3. You can get a clear idea of how the 925X/Intel 560 performs in Doom 3 in Anand's Doom 3: CPU Battlegrounds review published August 4th at http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
  • kherman - Thursday, August 12, 2004 - link

    Umm, Doom 3 benches?

Log in

Don't have an account? Sign up now