Overclocking: Gigabyte K8NXP-SLI

Gigabyte developed quite a reputation a couple of years ago as a large motherboard maker who paid attention to what enthusiasts wanted. In the last year, however, we have not seen the kind of attention to detail that distinguished Gigabyte in the past. With the K8NXP-9 and now the K8NXP-SLI, we are pleased to see Gigabyte back with the kind of features and the range of adjustments enthusiasts are looking for. The options in BIOS had us anxious to see where the Gigabyte nForce4 SLI board could go.

Front Side Bus Overclocking Testbed
Default Voltage
Processor: Athlon 64 FX55
CPU Voltage: 1.55V (default 1.50V)
Cooling: Thermaltake Silent Boost K8 Heatsink/Fan
Power Supply: OCZ Power Stream 520W
Memory: OCZ PC3200 EL Platinum Rev. 2
(Samsung TCCD Memory Chips)
Hard Drive: Seagate 120MB PATA (IDE) 8MB Cache
Maximum OC:
(Standard Ratio)
209x14 (5x HT, 2-2-2-10)
225x13 (5x HT, 2-2-3-10)
2926MHz (+12.5%)
Maximum FSB:
(Lower Ratio)
284 x 9 or 10 (2840MHz) (3x HT)
(1:1 Memory, 2 DIMMs in DC mode)
(+42% Bus Overclock)

The Gigabyte turned in one of the highest overclocks that we have seen on a Socket 939 Dual-Channel board. It is likely that the memory was holding us back at a 284 Clock Frequency, since we could only select 2.8V as the maximum memory voltage. Had higher options been available for memory voltage, it is likely that the Gigabyte could have grabbed the overclock record for 939. This is also supported by the 2.926GHz maximum overclock achieved at the stock multiplier. This is the highest stable overclock that we have seen with this FX55 processor, so we would expect a higher memory voltage to allow an even higher CPU clock at lower multipliers.

To put this in perspective, this is a board whose reason for being is SLI video, but the excellent K8NXP-9 underneath is showing through. If SLI excites you, then consider the additional performance options that the K8NXP-SLI opens up with its outstanding overclocking capabilities.

Features: Gigabyte K8NXP-SLI Memory Stress Testing


View All Comments

  • Googer - Friday, November 26, 2004 - link

    Jim, Your Hard Drive was not designed from the start to take advantage of newer Technolgical Features.
    With all things being the same, no ncq or tcq added and all modes Identical except the added bandwith of sata 3 Gb/s port there would be NO hard drive improvement with your older drive over 1.5 Gb/s. Since the drive you are using was not specified or designed to take advantage of it.

    If you owned a Yugo that could only go 55mph max on the highway and the Speed limit was 55mph. Would changing the law to make the speed limit 110mph make your Yugo go any faster? NO. Only getting a new car would alow you to go any faster. The new speed limit does not mean that your new Escourt will go any faster than 85mph. :-)

    I hope my analogy made some sence and helped to create a clearer understanding of drive interface bandwith. AIM GoogerSmith for any questens.
  • Gnoad - Friday, November 26, 2004 - link

    Just commenting on this great motherboard review, and reminding people that this was not a SLI review, hence the exclusion of 1600x1200 so we can see performance differences between motherboards. I understand many people want to see 1600x1200 benches because I do too, but this was a mobo review. Lets just thank wesley for a great mobo review, and save our SLI discussions for a SLI review. Reply
  • ChineseDemocracyGNR - Thursday, November 25, 2004 - link

    1) The four ports provided by the nForce4 chipset support NCQ.

    2) No. You need a SATA-300 drive to take advantage of the extra bandwidth available. Right now, even a Raptor is not limited by the SATA-150 interface.
  • jcromano - Thursday, November 25, 2004 - link

    I have a question or two about those SATA ports. There are 4 ports capable of 3 Gb/s and 4 capable of 1.5 Gb/s, if I'm reading things correctly.

    1) Do any or all of the ports support NCQ?
    2) If I had a fairly generic 7200RPM 8MB cache (Seagate or Maxtor or something similar), would I notice any difference at all between the fast port and the slow port? Or would I need some kind of super fancy drive to take advantage of the 3 Gb/s port? (And if so, what kind of drive would it take?)

  • Filibuster - Thursday, November 25, 2004 - link

    >#18 - 2x6600GT it $400 total, 2x6800 is $600 total, 2x6800GT is $800, and finally 2x6800 Ultra is $1000

    Where can anyone even get one of the 6800 PCIe cards for under $600? Even the GT is impossible to find for less than $550 (actually in stock).
    Is there a reason they are so limited? Like a new core to make the 6800 series PCIe native and to 'update' the video processor?

    This appears to be a very common question in many forums but there hasn't been an answer from any knowledgable source.

    Any information would be especially useful.
    Thank you.
  • flexy - Thursday, November 25, 2004 - link

    I understand your resolution decision, plus you probably didn't have time to bench it all... but no 8xAA/16xAF. Especially when the 6800Ultra came to a crawl with 8xAA enabled.
    why use 8xAA, and why use 16xAF ?

    4xAA and 8xAF is the current "standard" and IMHO a good compromise "image quality -> performance"

    How many run 8xAA ? See....then there's no point in covering this.

  • cnq - Thursday, November 25, 2004 - link

    OK, I think down deep Wesley knows he screwed up and wrote a fairly useless review. Let's give him a chance to stop denying it and make it right.

    Who was the Anandtech reviewer a couple months back who redid a review -- during his vacation no less -- after reader feedback? I think it was one of those Opteron vs. Xeon or Itanium reviews, when the mistake was made of using, well, a 3500+ instead of a true Opteron. Darn, I don't remember who it was (was it Kristopher?) but everyone immensely respected him afterward for standing up, acknowledging mistakes, and taking the time to redo things. Wesley has the title of Senior Editor for the web's #1 computer review site, so I'm sure he'll live up to those standards. Everyone knows that SLI users will be running at 16x12 and higher 100% of the time; I think the convincing has been done. Now let's kick back and give Wesley some breathing room and a chance to make the necessary fixes to his review.
  • Alkali - Thursday, November 25, 2004 - link

    "Why is playing a game on a 19" @16x12 difficult at all? If you have a decent 22" Diamondtron try having someone set one input up for you running 1280x1024 4xAA and the other input running 2048x1536 without AA and see what you think is better. I can't even comment about how 2048x1536 with AA would look as the only people that can run that type of setup right now are those with SLI parts in their hands and unfortunately they haven't deemed us worthy of obtaining that type of knowledge."

    Absolutely agree.

    There is absolutely NO point in any SLI review if there is no testing at 1600x1200, 1840x1440, and 2048x1536.

    There are many, many, many people like me who own 22" monitors, have tried FarCry etc at 2048x1536 and can't yet run that resolution because 1 card simply isnt good enough.
  • flexy - Thursday, November 25, 2004 - link


    i am envious you having the privilege having all this nice hardware to test....you know how many people are eagerly waiting for the nforce 4 boards.

    The gigabyte looks absolutely great (except the limited vdimm v which is kinda weird, 2.8V max ???)....but lets not forget that there are more boards coming up on nforce 4...who knows what the others (MSI, ASUS etc. will bring)

    I still have a burning question because i *need* to know whether the ThermalRight XP-120 HSF would fit on the Gigabyte board.
    This is supposed to be one of the best (if not THE best air cooling solution)...
  • Googer - Thursday, November 25, 2004 - link

    Does anybody know if you can use that second x16 PCI-E slot for anything else other than Graphics? Reply

Log in

Don't have an account? Sign up now