Soltek K8AN2E-GR: Overclocking and Stress Testing

FSB Overclocking Results


Front Side Bus Overclocking Testbed
Default Voltage
Processor: Athlon 64 3200+
2.0GHz
CPU Voltage: 1.5V (default)
Cooling: Thermaltake Silent Boost K8 Heatsink/Fan
Power Supply: OCZ Power Stream 520W
Memory: OCZ PC3200 EL Platinum Rev. 2
or Geil PC3200 Ultra X
Maximum OC:
(Standard Ratio)
246 x 10 (IDE), 230 x 10 (SATA)
2460MHz (+23%)
Maximum FSB:
(Lower Ratio)
250 x 9 (1:1 Memory, IDE, 2 DIMMs)
230 x 9 at (1:1 Memory, SATA, 2 DIMMs)

While the Soltek SL-K8AN2E-GR fared a bit better than the Asus with SATA drives, both sets of SATA connections still limited the maximum overclock. The board managed to reach a more respectable 250 CPU setting (the maximum available with the latest BIOS) with an IDE drive, but this is still far short of the 270 to 316 that we have seen with other nForce3-250 boards.

We suspect that the Soltek can do much better than 250 with lower ratios, but that is the current limit of the CPU range. If Soltek fixes this oversight with a BIOS update, you will likely be able to achieve higher performance with the K8AN2E-GR. The other controls that are needed for excellent overclocking with the Soltek are all here.

Memory Stress Test Results:

The memory stress test measures the ability of the Soltek K8AN2E-GR to operate at its officially supported memory frequency (400MHz DDR), at the best performing memory timings that our OCZ PC3200 Rev. 2 modules will support. Memory stress testing was conducted by running RAM at 400MHz with 2 DIMM slots filled. However, please keep in mind that Socket 754 is single-channel and will still operate at top speed with just one DIMM.

Stable DDR400 Timings - 2 DIMMs
(2/2 DIMMs populated)
Clock Speed: 200MHz
Timing Mode: N/A
CAS Latency: 2.0
RAS to CAS Delay: 2T
RAS Precharge: 10T*
Precharge Delay: 2T
Command Rate: 1T
*Several memory tests have shown that memory performs fastest on the nVidia nForce chipsets at a TRas (RAS Precharge) settings in the 9 to 13 range. We ran our own Memory Bandwidth tests with memtest86 with TRas settings from 5 to 15 at a wide range of different memory speeds. The best bandwidth was consistently at 9 to 11 at every speed, with TRas 10 always in the best range at every speed. The performance improvement at TRas 10 was only 2% to 4% over TRas 5 and 6 depending on the speed, but the performance advantage was consistent across all tests. All benchmarks were run at a TRas setting of 10.

The Soltek K8AN2E-GR was completely stable with 2 DIMMs at the best performing settings of 2-2-2-10, at default speed. The Soltek has just 2 DIMM slots and does not support 3 DIMMs.

Soltek K8AN2E-GR: Features and Layout DFI LANParty UT nF3 250Gb: Features and Layout
Comments Locked

39 Comments

View All Comments

  • thebluesgnr - Wednesday, September 15, 2004 - link

    #28,

    have you tested any SiS board from ASRock? They claim their K8S8X locks AGP/PCI. I've seen good OC results with that board - including this one:
    http://members.home.nl/ethanol/mem.JPG

    Also, OCWorkBench has a review of the ASRock K8-Upgrade-760GX. They overclocked the FSB to 252MHz on this mATX board, so I can only assume it locks the AGP/PCI buses.
  • Wesley Fink - Wednesday, September 15, 2004 - link

    #24 - justly
    There is actually another complaint about Sis. None of the Sis A64 cipsets I have tested, including the 939 Reference Board, have a working PCI/AGP lock. We sincerely hope this is fixed on the upcoming 939 chipset.

    We liked the Sis chipset very much, but major manufacturers just wouldn't support it. If you recall we awarded the Refernece Board our Editors Choice - as did other web sites - then we all waited for the boards that never came.

    I think Sis is an innovative chipset and we have reviewed all the Sis 754 boards we could find, including DFI and Foxconn. The people at Sis are also great to work with and we would personally love to see a significant win by Sis. Unfortunately, Sis is mainly seen in bargain boards. We agree it's a shame, but we also have to deal with reality in our testing.
  • Wesley Fink - Wednesday, September 15, 2004 - link

    #24 and #26 - Justly -
    Thanks for the benefit of the doubt here. I always compare new drivers to earlier scores to see if there are substantive differences. Frankly if there are I normally stick with the old drivers for consistency.

    That is the reason you have not seen us using Divx 5.2, for instance, in place of 5.1.1. When we tested 5.2 the performance differences from 5.1.1 were significant. Eventually we will replace 5.1.1 with the latest Divx when it fits the schedule to do retesting.

    The performance differences I found cannot be explained with 4.8 vs. 4.5 ATI drivers. I suspect BIOS tweaks DO have something to do with it however, which, as you are suggesting, probably means the earlier boards with later BIOS' are probably also faster.

    There is always the trade off between changing driver versions for testing and keeping drivers up-to-date. I can only say AnandTech is very cautious about driver versions - particularly in ongoing test/database areas like motherboards.

    You will soon be seeing a new General Performance Benchmark at AnandTech, since Veritest and PC Magazine have discontinued support for Winstones (and they don't work well with SP2). We will be talking more about this in an upcoming review.
  • justly - Tuesday, September 14, 2004 - link

    #25 – I understand the delay involved in retesting. I am not trying to say that retesting is required for comparison, but if the scores where just copied then it would be nice to see it pointed out on the test configuration or in the final words that more than one video driver was used.

    Actually a while back (around the time of “the real slim shady” article) I noticed identical systems being retested (using identical software/drivers) where the scores fluctuated more than the difference I suspect this article might have. So I am really giving Wesley a lot of credit (although it may not sound like it) for being able to set up systems months apart with such consistency.

    Really the only complaint I have (other than leaving out SiS) is that if he did what I suspect, then it should have been mentioned in the article. Then again if what I suspect is true it might have been better to leave out the whole first paragraph on the final word page.
  • TrogdorJW - Tuesday, September 14, 2004 - link

    #24 - I understand the desire to see comparable scores, but retesting on old hardware with new drivers would probably add a week or two of work. I certainly wouldn't want to do it! Then there are BIOS revisions that need to be updated as well. Yuck. :p

    I think it's relatively safe to say that performance with most of the other Nforce3-250 boards is going to be about the same as these, and the only remaining factor tends to be overclocking and features. I'm perfectly happy with my MSI K8N Neo Platinum. Were I buying a new S754 board today, it would still be a tough call between the DFI and the MSI board. I don't like the Asus look (or lack of certain features), and the same goes for the Soltek.

    Really, I think I would still stick with my MSI board. It would probably end up coming down to what else was included with the motherboard - nice rounded IDE cables would be great, as would a rounded floppy cable (because I still use a floppy drive on occasion). Just one man's opinion, of course.
  • justly - Tuesday, September 14, 2004 - link

    Wesley, first off I would like to say that I think your articles are some of the best on Anandtech. The thing is I still see a few things in your articles that “ever so slightly” annoy me.
    One is this quote “If overclocking is not particularly important to you, then one of the first generation boards based on the VIA chipset might also meet your needs at a lower price.” WELL WHAT ABOUT SiS. It seems that Anandtech has amnesia when it comes to SiS chipsets. Other than the very first SiS based socket 754 motherboard (the ECS 755-A) all Anandtech reviews seem to have only one main complaint about the SiS chipset, it overclocking abilities. So why don’t you mention it?

    The other thing is that by looking through the Generation 2 Socket 754 Roundup it appears that the Generation 2 motherboard results where copied not retested. I think it would be fine to do (for comparison sake) if everything was the same but in the Generation 2 Socket 754 Roundup it shows a different video driver than what is listed for the Socket 754 Roundup, Part 3. So I have to ask is there absolutely no performance difference between the cat 4.5 and cat 4.8 drivers, or could the reason that the 3rd generation boards seem slightly faster have something to do with the video driver being used?

    I realize that my concerns are very trivial and probably have no effect on the outcome of the article, but to be fair to the other chipset and motherboard manufactures I still think they are valid questions, trivial yes but still valid.
  • Wesley Fink - Tuesday, September 14, 2004 - link

    #21 and #22 - The full implementation of nF3-250Gb is 4 SATA ports that can be combined in any way in RAID with the IDE ports. Asus implemented 2 nVidia SATA ports plus 4 Silicon Image SATA ports.

    The problem is ports 1 and 2 on nVidia are coupled with the PHY Gigabit LAN and generally will not overclock very well. Ports 3 and 4 generally perform as well as regular IDE on the nF3-250Gb chipset.
  • jediknight - Tuesday, September 14, 2004 - link

    OK... I'm confused here. I thought Asus added an *extra* RAID controller in addition to the one provided in the stock 250gb implementation. Am I wrong here?
  • Zebo - Tuesday, September 14, 2004 - link

    "The decision by Asus to use Silicon Image SATA instead is really a drawback in overclocking."

    Instead of what?
  • Wesley Fink - Tuesday, September 14, 2004 - link

    #18 -
    Quite a few users are reporting success using the Mobile Athlon 64 chips with the DFI LANParty UT. Many are stating the DFI recognizes the mobiles just fine and sets the correct settings for the mobile chips. In fact you will see this combination as an alternate in an upcoming OC Guide.

    The biggest issue with the 754 mobiles on a desktop, once the board compatability is fine, is the HSF. Most won't make good contact with the mobile that does not use a heatshield. I am hearing decent things about the Thermalright XP-90 sink with mobiles on a K8.

Log in

Don't have an account? Sign up now