Soltek QBIC EQ3801: nForce3 250Gb goes SFF
by Evan Lieb on July 4, 2004 10:35 AM EST- Posted in
- Systems
Soltek EQ3801: Overclocking and Memory Testing
Overclocking Testbed | |
Processor: | Athlon 64 3200+ 2.0GHz |
CPU Voltage: | 1.50V (default) |
Cooling: | AMD stock Athlon 64 Heatsink/Fan |
Power Supply: | Soltek 250SFD (250W) |
Maximum OC: (Default Ratio) |
222FSB x10 2222MHz (+11.1%) |
Maximum "FSB" (HT): (Lower Ratio) |
250FSB @ 8X multiplier |
During our routine overclocking testing, we were able to reach a core clock speed of 2.22GHz on our 3200+'s default 10X multiplier. Compared to Athlon XP and Pentium 4 processors, this is hardly an overclock to get excited about. However, as you may or may not be aware, Athlon 64 overclockability differs from CPU to CPU. Just look at AnandTech's very own nForce3 250Gb overclocking results from last month's Socket 754 roundup. As you can clearly see from that article, one of our other editors just happened to get an Athlon 64 processor (of the same 3200+ speed grade) that could overclock regularly to between 2442MHz and 2480MHz on stock cooling, default voltage, and default multiplier. I just happened to get an Athlon 64 3200+ that tops out around 2200MHz, give or take a few MHz.
Anyway, after seeing how low our Athlon 64's core clock speed ceiling was on default multiplier, we decided to see how well the Soltek EQ3801 could handle a high FSB at an artificially low multiplier (8X in this case). The EQ3801 was able to reach 250MHz FSB, which is the highest option available in the BIOS. This is unfortunate, as we are quite sure that we'd get past 250MHz FSB if the BIOS allowed us to. That being said, we had to lower our HT setting from 4X (800MB/s) to 3X (600MB/s) to reach 250MHz. Certain nForce3 desktop motherboards reached as high as 266MHz FSB with 4X HT enabled.The EQ3801's B9A-FGR motherboard is running off a relatively modest, yet capable, 250W PSU and simply cannot be as easily optimized as larger ATX desktop motherboards, so we can definitely cut it some slack.
Overall, our EQ3801's overclocking results are definitely impressive compared to other Athlon 64 SFFs, but pretty much not close at all compared to retail nForce3 150/250/250Gb desktop motherboards due to PSU (250W) and motherboard size limitations. If Soltek can somehow release a revved up B9A-FGR BIOS with better FSB adjustments and voltage settings soon, the Soltek EQ3801 will be able to compete comfortably with the best Socket 754 nForce3 150/250/250Gb desktop motherboards on the market. Realistically, the best that we can hope for from any current or future SFF competing with today's specialized overclocker/performance desktop systems is generally good performance. Not quite cream-of-the-crop performance, but good solid performance/overclocking.
Memory Stress Test Results:
This memory stress test is very basic, as it simply tests the ability of the Soltek EQ3801 to operate at its officially supported memory frequency (400MHz DDR), at the best performing memory timings that our 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 (2/2 DIMMs populated) |
|
Clock Speed: | 200MHz |
Timing Mode: | N/A |
CAS Latency: | 2.0 |
Bank Interleave: | N/A |
RAS to CAS Delay: | 3T |
RAS Precharge: | 6T (use 9T or 10T for best performance) |
Precharge Delay: | 2T |
Command Rate: | N/A |
Pretty much the only reason why you wouldn't be impressed with these timings running at 400MHz DDR is simply due to the fact that these same modules have run at CAS 2-2-2-6, at DDR400 on countless other motherboards, including nForce3 250Gb motherboards. Generally, these are perfectly fine timings, even if they commonplace nowadays, especially among single channel DDR motherboards.
22 Comments
View All Comments
Swaid - Sunday, July 4, 2004 - link
The graphs for most of your benchmark only display one motherboard and leave out the rest, so its pretty much impossible to distinguish what the numbers corrispond to.Z80 - Sunday, July 4, 2004 - link
Your encoding benchmarks are not displaying properly