FIRST LOOK: Asus P5AD2-E - 1066FSB & DDR2-711
by Wesley Fink on November 29, 2004 12:01 AM EST- Posted in
- Motherboards
Asus P5AD2 Premium: Overclocking and Stress Testing
FSB Overclocking Results
| Front Side Bus Overclocking Testbed | |
| Processor: | Pentium 4EE LGA 775 (Gallatin Core) 3.46/3.2GHz |
| CPU Voltage: | 1.55V (default) |
| Cooling: | Thermaltake Jungle 502 |
| Power Supply: | OCZ PowerStream 520W |
| Maximum OC at Stock Multiplier: | 300x13 - 3900MHz (+12.7%) |
| Maximum FSB OC: | 324FSBx12 (+21.8% Bus OC) |
The 3.46EE CPU Intel supplied for testing was unlocked, but the only available multipliers were a stock 13X and 12X. Since the 3.46EE is the only current 1066FSB CPU, this means that options for bus overclocking are limited. However, having said that, we were still able to reach 324FSB at 12X for a DDR2 1:1 memory overclock of 648. Asus also provides an additional memory multiplier of 711 on the P5AD2-E, so we tried out the higher memory frequency with some of the latest OCZ 1GB DDR2 DIMMs rated at 3-2-2-8. This memory ran great in the P5AD2-E and also allowed us to reach DDR2-750 with a modest overclock at the DDR711 setting.
The P5AD2-E provides the best overclocking that we've seen on an Intel 925XE/925X board, although the Abit Fatal1ty AA8XE, based on the same 925XE xhipset, reaches virtually the same levels in overclocking. The 711 multiplier for memory opens new options for memory overclocking that are unique to the P5AD2-E.
Memory Stress Test Results:
The memory stress test measures the ability of the Asus P5AD2-E 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 1:1 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: | 1:1 |
| CAS Latency: | 3.0 |
| Bank Interleave: | Auto |
| RAS to CAS Delay: | 3 |
| RAS Precharge: | 3 |
| Cycle Time (tRAS): | 10* |
| Command Rate: | N/A |
The Asus P5AD2-E was completely stable with 2 DDR2 modules in Dual-Channel at the settings of 3-3-3-10, at 1.8V default voltage. Intel has updated memory timings on the 925XE series boards and now specifies 3-3-3 as default timings compared to the 4-4-4 timings specified at the 925X/915 launch.
Filling all four available memory slots is more strenuous on the memory subsystem 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: | 3.0 |
| RAS Precharge: | 10T* |
| Precharge Delay: | 3.0 |
| Command Rate: | N/A |
As we first saw in the 925X roundup, four DDR2 DIMMs are not stable at CAS 3 as 2 DIMMs are. We required 4-3-3 timings when using 4 DDR2 DIMMs, though the 4-3-3 timings did work fine at default voltage.
Facebook
Twitter
RSS
34 Comments
View All Comments
dariush - Friday, February 4, 2005 - link
For Gaming, AMD is better but not all computer users are gamer, When Talking about about video encoding and stability (must important in critical cases) you see Intel says the last words.Wesley Fink - Thursday, December 2, 2004 - link
#32 - I would sincerely appreciate your reading the review before posting comments. Overclocking is covered on p.3 - which is actually titled "Overclocking and Stress Testing".Second, this is a First Look which is less detailed that a full review - the idea is to bring you more reviews quicker. We used to include a page explaining this, but I removed it this go-round because I assumed the idea was well-known by now. It appears I was should have kept the explanation page in the review.
Third, First Look uses PCMark2004 as a General Performance comparison. However, as I stated in Comment #18 the complete Winstones and media encoding do not provide additional information. However, they are already posted in Comment #18 and I'll repeat them:
"We did run a full suite of benchmarks for future comparisons, but nothing really changes.
925XE/3.46EE - 925X/3.6E - nF4/FX55 - Benchmark
34.1 - 34.4 - 39.3 - MM Content Creation 04
26.7 - 26.5 - 31.1 - Business Winstone 04
73.1 - 73.4 - 69.1 - AutoGK DivX 5.1.1".
Fourth, we use a standard setup to allow easier comparisons. Where standard 1024x768 benchmarks are CPU or GPU bound we do use AA and/or higher resolutions to allow valid comparions of motherboard performance. For example Comanche 4 is run at 1280x1024 with 4X AA because 1024x768 without AA reveals nothing about true performance. The "Score" bench for Aquamark 3 uses 4AA by default. We run the newest games at highest settings as detailed in the chart descriptions.
T8000 - Thursday, December 2, 2004 - link
If you talk about enthousiast options, overclocking is usually one of them. Still, no single overclocked CPU was included in the benchmarks.There was also no mention about the 925XE containing more options to keep PCI-E within spec during overclocking.
Besides, there where only gaming benchmarks with high end GPU's running at 1024x768 with no AA or AF, thus creating an unlikely scenario with performance numbers that do not reflect real world performance differences, even for those who just play games.
Who would even consider buying a $500 GF6800U to play only at settings that a $200 GF6600GT does with ease?
danidentity - Wednesday, December 1, 2004 - link
#28 - The Dolby Digital Live encoding on this board is functionally the same as SoundStorm on the NF2 boards. No PCI sound cards currently support Dolby Digital Live.cALIx - Tuesday, November 30, 2004 - link
Not to change the subject, but in this review, Doom 3 is listed as a dx9 game. Isnt D3 an opengl based game? If you open the console you can easily see all the gl_ extensions that are open. Also Hardocp always lists D3 as an opengl game when they benchmark with it. So I'm just wondering if someone could clear that up for me. I understand what they are individually, but i'm still confused as to how they work/dont work together...LoneWolf15 - Tuesday, November 30, 2004 - link
Ouch...my brand new $140 MSI K8N Neo 2 mainboard and I feel sorry for those who choose to shell $250-300 for this board. Not that it's a bad piece of hardware, not that Intel's high-end P4 chips are bad hardware either...but you can do better for far less money, so why would you pay this kind of cash? Especially when you could balance out your system with the money you've saved by upping your graphics card or opting for more RAM or a faster hard disk. Guess I'm just someone who doesn't see the point in buying a Jaguar when an Acura RSX-S would do what I want for less.tc2k04 - Tuesday, November 30, 2004 - link
I know i'm late, but,Did anybody else notice the talk about the audio solution, is this encoding comparable to the output of nforce2 boards? they say its a first for on the motherboard audio, can you buy seperate pci cards at reasonable prices which can do this?
i'm trying to get a sound solution that can output 2 channel audio in 5.1 through a digital out like my nforce2 onboard does. If this can do this, i might end up buying an intel.
Googer - Tuesday, November 30, 2004 - link
Intel's engineering DEPT takes orders From queer mac using Marketers.Googer - Tuesday, November 30, 2004 - link
#3 it works. I have done it. Super7 is like a Swiss Army Knife.TrogdorJW - Monday, November 29, 2004 - link
Hey Wes,Any chance of some OC tests and numbers from a standard Prescott core?
I'm wondering if a 2.8E has the potential to reach a 1066 FSB overclock.
(I'm also including hard returns due to the overly long link someone posted.)
As the 2MB cache alleviates some of the pressures of the FSB, a 1MB cache might benefit more.
Besides, if a 2.8E could actually OC well to 3.73 GHz and a 1066 FSB, that might be worthwhile!