Asus PC-DL Deluxe: 875 with Dual Xeons
by Wesley Fink on September 6, 2003 12:06 AM EST- Posted in
- Motherboards
Asus PC-DL: Stress Testing
We performed stress tests on the PC-DL in these areas and configurations:1. Chipset and motherboard stress testing, which was conducted by running the FSB at 144 MHz with 2x512MB double-bank Mushkin PC3500 Level II at the fastest 2-2-2-5 timings.
2. Memory stress testing, which was conducted by running Corsair 3200LL RAM at 333MHz with all 4 DIMM slots filled. Two pairs of Corsair TwinX ver.1.2 was used for this test at the lowest memory timings (2-2-2-5) possible.
Front Side Bus Stress Test Results:
We ran a full range of stress tests and benchmarks to ensure the PC-DL was absolutely stable at each overclocked FSB speed. These stress tests included Prime95 torture tests, which were run in the background for a total of 24 hours. We also ran several other tasks — data compression, various DX8 and DX9 games, and apps like Word and Excel — while Prime95 was running in the background. Finally, we ran our benchmark suite, which includes ZD Winstone suite, Unreal Tournament 2003, SPECviewperf 7.0, and Gun Metal Benchmark 2.The Asus PC-DL was completely stable at the 144 or 3.3GHz setting.
Memory Stress Test Results:
The memory stress test is very simple, as it tests the ability of the Asus PC-DL to operate at its officially supported memory frequency (333MHz DDR) at the lowest supported memory timings our Corsair TwinX LL ver. 1.2 can achieve:| Stable DDR333 Timings (2 Dimms in 1 Dual-Channel Bank) |
|
| Clock Speed: | 333MHz |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We had no problems at all running 2 Corsair dimms at the fastest timings available at DDR333. Since the Canterwood chipset was designed for DDR400 operation, it is hardly a surprise that it breezes though DDR333 performance. As new options become available for Dual-Xeon boards it will be interesting to see if DDR400 will present any problem. If Intel intends the 875P to be targeted to any true server markets, there will also have to be the option for much more memory than 4 dimm slots. However, we suspect the 875 will be the workstation, gamer, performance enthusiast, and SOHO server solution.
Filling all 4 available dimm slots with 2 banks of dual-channel memory is more strenuous on the memory subsystem than testing 2 dimms. However, the Asus PC-DL had no trouble running 4 dimms at the fastest DDR333 timings available.
| Stable DDR333 Timings (4 Dimms in 2 Dual-Channel Banks) |
|
| Clock Speed: | 133MHz |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We really didn’t expect any problems running our DDR400 memory at the most aggressive DDR333 timings available. Most of the memory currently available in the marketplace will run at fast timings on this DDR333 motherboard.
Facebook
Twitter
RSS
29 Comments
View All Comments
Kiwi42084 - Tuesday, March 2, 2004 - link
This Benchmark is totally unfair!!!!!The PC-DL has the avalibility to produce 4 usable CPUs...2 Physical and 2 Logical....
Windows XP will only see 2 of these. The benchmarks are being done with Half of the POWER POTENTIAL.
vppaul - Wednesday, February 25, 2004 - link
I am having a problem with my PC-DL board. The System Management BIOS is reporting that there is 4096 MB of RAM, but Windows reports that 2048 MB is available. Anybody ideas? I am trying to get help from Asus tech support but any help would be appreciated. Running dual 3.06 with 4 x 1GB sticks.piperfect - Monday, February 23, 2004 - link
I totally agree with FutureShock999. Why not run several instances of the a divx encoder. For instance do a 2 part movie on an encoder that runs two threads per instance then run both of the parts at the same time in two instances of the program on the xeon and see who finishes first. What you guys are doing is like comparing an f-16 fighter jet to a f-15 and saying the f-15 can only use one of its engines.piperfect - Wednesday, February 11, 2004 - link
I got my pc-dl with 2.8GHz Xeons with raptor drives in RAID0 to 3361MHz 160FSB with a dram clock of 200MHz. 4:5 It ran Sandra burnin overnight. Maybe my board is a newer revision. I read this article after I bought the Raptor so I didn't try to overclock it until yeaterday but it did and it runs well!!!!!FutureShock999 - Monday, October 6, 2003 - link
Wesley,A nice review, but I believe the wrong benchmarks. No one should buy a dual-processor machine to execute a single-threaded application faster, especially when that single proc is slower than others on the market. Dual-proc boards are bought to do multi-threaded stuff, either runnig a single multi-threaded application, or running several different single-threaded applications. In no place did I see benchmarks that explicitly looked like that.
As such, your review was a good cautionary tale for people that didn't KNOW the above, and hopefully will stop some people from spending a lot of money on a this board hoping to have a great UT setup.
Now if you had shown what it was like to play a game, encode media, and download a few gigs of content SIMULTANEOUSLY, then we could really see how this board stacks up to the competition you evaluated it against. And I think it might have beaten them rather handily...
Wesley Fink - Tuesday, September 9, 2003 - link
#23 -We recently changed our standard video card from nVidia's Ti4600 to the ATI RADEON 9800 PRO. Theoretically this should have no impact on encoding scores, but we reran all benchmarks with our new standard hadware on a few of the highest performing boards. Evan ran about half the new encoding benchmarks on the west coast, and I ran the other half on the east coast. As you can see our new results compare very well to each other.
I have no other explanation, but perhaps Evan can shed some light on this. I have used the new ATI Radeon 9800 PRO from day 1 and my benchmarks have been cumulative over the last couple of months with no dramatic change that you point out.
Anonymous User - Tuesday, September 9, 2003 - link
The 865/875 P4 boards in this test all perform 50% faster in the media encoding benchmark in this review than they did in the round-up a couple weeks ago.Has anyone else noticed this dicrepancy? The Abit IS7-G has gone from 64.45 fps to 104 fps. Both times they used a P4 3.0Ghz, 800mhz FSB, with HT enabled. Nothing seems to have changed except the encoding speed. I wish I could do that to my rig ; )
A 50% increase in as linear and consistent a benchmark as DivX encoding is simply astounding.
I just wanted to point this out to everybody around here.
Thank you for your time.
Anonymous User - Tuesday, September 9, 2003 - link
It' will not overclock with the SATA drives on the intel controller. BUT it will if you run them on the onboard promise controller. I have 2 WD Raptor's running RAID0 and the 2800/533's running @3250 100% stable.Anonymous User - Monday, September 8, 2003 - link
Which brings up the question about whether or not HT was enabled on it. And the fact is they used a regular consumer card for video Most workstations would have a workstation class card in them such as a Quadro, FireGL, or even a 3DLabs card in them. It all depends on the applications that one uses. Most normal people wouldn't use a dual machine for gaming anyways. They'd use them for graphical processing or media encoding or file serving and such. Just talk to those guys over at 2CPU.com they know what I'm talking about. ;)Anonymous User - Monday, September 8, 2003 - link
From the looks of it, there would be little if any reason to spend gobs of extra money on a system that is beat by AMD in gaming, and by single P4 siblings in high-end workstation tasks.