AOpen AK89 Max: From Ugly Duckling to Swan
by Wesley Fink on April 21, 2004 4:16 PM EST- Posted in
- Motherboards
AOpen AK89 Max: Overclocking and Stress Testing
FSB Overclocking Results
BIOS options don't mean much unless they can translate into improved performance. We have complained for months that the major reason we could not reach higher overclocks was not the poor headroom of the Athlon 64, but rather, the lack of a fixed AGP/PCI lock on the VIA and nVidia chipset boards that we were testing. Now, we get a chance to put our money where our mouth is.| Front Side Bus Overclocking Testbed | |
| Default Voltage | |
| Processor: | Athlon64 2.0GHz |
| CPU Voltage: | 1.5V (default) |
| Cooling: | AMD Stock Athlon 64 Heatsink/Fan |
| Power Supply: | Antec TruePower 430W |
| Maximum OC: (Standard Ratio) |
233FSB x10 2333MHz (+16.5%) |
| Maximum FSB: (Lower Ratio) |
347FSB x 6 (328x7 for CPU limited 2296MHz) |
The above overclocking setup at default voltage allowed us to reach a stable FSB of 233. With such a limited top vCore, we had to consider the 1.55V as essentially the same as the default 1.50V. A voltage range to at least 1.7V to 1.85V would be preferred - especially since the AK89 Max turned out to be an incredible overclocker. Even without CPU voltage adjustment, we reached an astounding 347FSB using some common overclocking tricks. Take a look at the Extreme Overclocking section to see details on how we reached those levels. Screen captures are also included.
Front Side Bus Stress Test Results:
A full range of stress tests and benchmarks were run to ensure the AK89 Max was stable at each overclocked FSB speed. This included Prime95 torture tests, and the addition of 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 Veritest Winstone 2004 suite, Unreal Tournament 2003, SPECviewperf 7.1, and Aquamark 3. 230MHz was the highest overclock that we were able to achieve with the AOpen while running these tests at default voltage and default multiplier.Memory Stress Test Results:
The memory stress test is very basic, as it tests the ability of the AK89 Max to operate at its officially supported memory frequency (400MHz DDR) at the lowest supported memory timings that our Mushkin PC3500 Level 2 or OCZ PC3500 Platinum Ltd Modules will support. Memory stress testing was conducted by running RAM at 400MHz with 2 DIMM slots filled.| Stable DDR400 Timings - 2 DIMMs (2/3 DIMMs populated) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| 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 problem running 2 DS 512MB DIMMs of our standard Mushkin PC3500 Level 2 or OCZ PC3500 Platinum Ltd in the AOpen AK89 Max. Both of these memories were completely stable at the Jedec standard memory voltage of 2.6V at DDR400 with the fastest 2-2-2-5 timings that we could select on this board. We were able to achieve the same fast timings with either 1 or 2 DIMMs.
Filling all available memory banks is more strenuous on the memory subsystem than testing 2 DIMMs. AOpen states clearly that only 2 DIMMs or a maximum 2GB of memory are supported at DDR400 speed, but we decided to try 3 DIMMs at that speed. 3 DIMMs only worked at DDR333 speed. This was consistent with AOpen's specification of a higher memory capacity of 3GB at DDR333. For best performance on the AK89 Max, we would recommend using 1 or 2 DIMMs. If you want more memory, use 2 of the new 1GB memory modules from Corsair, Kingston, OCZ, and others to allow 2GB of total memory using 2 DIMMs.
| Stable DDR400 Timings - 3 DIMMs (3/3 DIMMs populated) |
|
| Clock Speed: | 166MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We tested the memory timings with 2 banks filled using several stress tests and general applications to guarantee stability. Prime95 torture tests were successfully run at the timings listed in the above charts. We also ran ScienceMark (memory tests only) and Super Pi. None of the three stress tests created any stability problems for the AK89 Max at these memory timings.
Facebook
Twitter
RSS
23 Comments
View All Comments
retrospooty - Friday, April 23, 2004 - link
Am I the only one that found this artcle hard to follow ? Its all over the place ... What the hell was being tested ?cowdog - Thursday, April 22, 2004 - link
Thanks for the information.BTW, I finally did get through to AOpen, and they have been very prompt and helpful. I must have fallen through the cracks or something earlier.
Wesley Fink - Thursday, April 22, 2004 - link
Cowdog -I used our standard Mushkin/OCZ 3500 at default speed. Above 233, I used OCZ 3500EB to DDR500, and OCZ 4200EL to DDR550. if you look closely at the screen shots the memory is identified on the overclocks. As you state, BH6 can't reach these levels.
AOpen tells us the board would require a revision to allow an increase in vcore, and with the nF3-250 near, a new rev will not appear with the 150. They can't modify the vCore on this board with just a BIOS update. vdimm and higher FSB may be a possibility - we have requested those changes - but it is more likely these will come with the 250 version.
cowdog - Thursday, April 22, 2004 - link
FYI: This seemed as good a place to post this as anywhere else. Just came across a new bios, 1.07"For AMD recommended,memory clock will set to DDR333 when DIMM1 and DIMM2 plugged double side DDR400 memory."
Two things:
1. BIOS defaulted to 200x11 for me, which isn't a good thing for an A64 3200
2. I have no option for DDR400 in the bios with 2x246MB Mushkin Lv II PC3500.
3. When I tried to lower the multipler, the bios gave me an error at POST (safe mode or something like that). Then even if I selected bios defaults I had the same message. I reset cmos and it worked, although cpu frequency went back to the 200x11.
Might want to be leary of bios 1.07, although my system may simply be whacked.
cowdog - Thursday, April 22, 2004 - link
Wesley, I must admit that I haven't tried Dr. AOpen; I missed that. I did fill out their tech support form (with serial number, etc.) three times. This third time (last night) I did receive an automated reply for the first time. I kept screen shots of the previous forms, and I know my email was right. I'm not sure what happened. Anyway, I'll try Dr. AOpen as well, thanks.Couple of things to add:
1. As far back as the official 1.06 bios, I was able to change multiplers (only downward, of course) using Clockgen. I found that clockgen would work if I set the multiplier lower in the bios, even though that did not change the multiplier at post. 1.06 m bios now changes the multiplier at post, as you note. Funny thing too, after I updated to 1.06m, the default multipler was 4.
2. I don't understand why AOpen won't open up the cpu and dimm voltages as they did with the AK86-L. As you note, the AK89 Max is their current A64 flagship. Why add features via bios update to the AK86-L but not do similar for the AK89 Max? Unless, of course, those features can't be added with a bios update. Do you know?
3. When you overlclocked the memory 1:1 to 250Mhz, were you using the OCZ or Mushkin memory that you referenced earlier in the article? I am unable run some Mushkin BH-5 memory 1:1 at anything higher than 230Mhz (clockgen) and 210Mhz (bios) with the multiplier way down and LDT at 3 or 2.5 even though I can run the same memory at speeds higher than 250Mz on another board. Same thing with some Corsair LL v.1.1 memory. Just curious. Perhaps the cpu is the limiting factor; is there that much variability?
Anyway, thanks for a really nice and thorough review. I am trying very hard to like this board (and AOpen) and your review renewed my motivation.
Wesley Fink - Wednesday, April 21, 2004 - link
#7 - The Asus K8V is based on the VIA K8T800 chipset which has no capability at all for a PCI/AGP lock. We have found the performance of the VIA and nForce3-150 to be about the same at stock speed, but the VIA is simply not cpable of avchieving the overclocks we have reported in this review.The K8V performs about like the MSI K8V Neo which is included in our benchmarks.
Modal - Wednesday, April 21, 2004 - link
I wish you would have done some comparisons with the ASUS K8V Deluxe as well, as this seems to be quite a popular board for Athlon64 machines (plus, I'm thinking about using it for my next build...). Oh well...Wesley Fink - Wednesday, April 21, 2004 - link
#5 - Have you tried Dr. AOpen on line? If nothing has worked please email me your details and I will forward your email to contacts at AOpen.Both Evan Lieb and I have had very positive Support expereinces with AOpen. If others are getting different treatment it is important to know that and report it in our reviews.
gimpsoft - Wednesday, April 21, 2004 - link
is this a Real AGP /PCI lock ???i heard there way to mod the bios to lock but not the same as hardware Agp/pci locks
well something like a sofmod ill see if i can find the link
may be a this must buy board i got a via KT800
i can do 3000+ 10 X 250= 2500MHZ but my radeon 9800pro & audigy bitch they say to high =) PCI: 41 APG: 83
bigtoe33 - Wednesday, April 21, 2004 - link
2-I was quite impressed by te FSB's the AOpen was able to hit and did feel the review was there just to show what the board is capable of.I also have this board and i would say 1:1 will be pretty limited due to 2.8V vdimm max.
Also it was nice to see a comparison of 3dmark although i know this is not a bench Wes uses.
Overall i think the review showed the board has awesome potential and proved that boards can become awesome with a "tuned" bios ;-)