Jump to content


  • Content Count

  • Joined

  • Last visited

Everything posted by DemonEyez_NL

  1. That's the reason why I bought this board I still use an AGP card, an heavily overclocked ATi X800 Pro whick runs as an X800 XT PE @ 640/1200. Which still gives me a lot performance in games. I think at the end of this summer/autumn I'll do another upgrade and get me an affordable nVidia 7800GS agp (still has to be released but will be a very fast AGP based card) and get me an affordable X2 proc
  2. Dear 306EXTREM, For Gskill to work correct, EVEN at default speed and settings , you will have to manually set each setting in the bios. I had a likewise problem when i first got this board and used my Gskill sticks. I left everthing at default and my PC refused to boot! There's perhaps some incompatibility issue with this board and the SPD's (SPD is the chip where al the stock settings of a memory module are kept in EEPROM) found on Gskill TCCD sticks. If those Gskill wont't work in an nF3 Ultra-D the don't have to be broken. Test them in another PC/mobo
  3. @ Bigjazz I myself run Kingston HyperX modules , 2x512MB, but I have to use bank 2 en 4. Those are the yellow slots, if I put the Kingston in the orange ones, the PC will not post at all, or keeps hanging at the Detecting Raid Array screen. I agree with [email protected] and Branjo. Get a decent PSU. @ Branjo I have the OCZ Powerstream 520 and it's a mighty fine PSU Would recommend to everybody
  4. hehe, i'm just as amazed as you are Perhaps it could be my Gskill crapping out, i've been running with the same kit for over a year know. I was one of the first few guys who ordered it directly from Taiwan when Gskill just started producing those kits and there weren't any resellers yet. So my two sticks have the first Samsung chips from Gskill ,431 TCCD chips. Nowadays its 519 or 550 I think, so in the TCCD line there've been some revisions Anayway, i just ordered some OCZ PC4800 Platinum Elite Edition TCCD and we'll see how that does in this board. I have to say, those boards from DFI are a challenge every time. I took me for over 3 months to get my old LanParty UT nf3 250Gb running on 10x295 with 2x512 @ 1T (BTW, My Opteron 146 ran Prime Stable voor 8 hours last night, doing 10x300 (4:3) at 1.525v. So at I think it has a decent memory contoller, and my board has no problems running a hight HTT )
  5. Okay, thanks for the info mate Oh well, I have the option of ordering some new ram and test it on this board, and if it's a no-go, can return it and pick some other brand of ram that works I tested my Opteron yesterday on a friend's board and it did 10x295HTT (1:1) so that one works great at least
  6. I know, I know... but come on! the Kingston is working fine and with those Gskills I can't even do more than 215Mhz. I realy feel like i'm missing something here.
  7. Thanx for all your responses guys, but i'm aware of divider table. But that doesn't explain the fact that I can run my Kingston Hyper X with a HTT of 280 and an 6:5 divider (which results in 233Mhz for the ram) but My Gskills will refuse to work with those exact same settings and also crap up the divider table. (btw, I never usehalf multi's) I used the Kingston Hyper X as an example to explain that with other memory than my Gskills the systeem runs fine and uses the correct mem frequency and dividers. It's also no problem to run the Kingston HyperX at 10x235 (1:1) , but yet with my Gskills the system will again refuse to work at those exact same settings But when I test those two Gskill sticks in MemTest, I recieve no errors and can run them up to 280HTT (1:1) doing like 40 passes with Test 5 and Test 8 So in short: 2x512 Hyper X at 10x280HTT (6:5) 233MHz (2.4-4-4-8 @2T) runs fine 2x512 Gskill at 10x280HTT (6:5) 233MHz (2.4-4-4-8 @2T) no boot, BSOD 2x512 Hyper X at 10x235 HTT (1:1) 235MHz (2.4-4-4-8 @1T) runs fine 2x512 Gskill at 10x235 HTT (1:1) 235MHz (2.4-4-4-8 @1T) no boot, BSOD 2x512 Gskill at 10x280HTT (1:1) 280MHz (2.4-4-4-8 @2T) MemTest gives 40 passes in test 5 and 8, but Windows will refuse to boot and gives BSOD's during loading screen (btw, on my old Lanparty nF3 250Gb my two Gskill sticks ran fine at 10x290HTT (1:1) @ [email protected] I've send you a PM
  8. I'v tried running ratios, but when i use the 6:5 divider and I go into memtest to check the correct MHz for the ram it most of the time is on 200MHz when it should be 235MHz for instance or my system wil not even post. So there'e something wrong when running dividers. I think I should be able to do 230HTT running 1:1 at least. I've installed some Kingston Hyper X temporarely and i'm now running 2x512MB at 10x280HTT (6:5) 235MHz 2.5-4-4-8 @ 1T. So go figure? It seems to me there's something wrong with my two sticks of Gskill. But the strangest thing is when I test those two sticks in MemTest they give absolutely no errors and I can do 50 passes on Test 5 and Test 8, running 2x512MB at 10x280 (1:1) (2.5-4-3-7 @ 2T or running 1x512MB at 10x280 (1:1) (2.5-4-3-7 @ 1T
  9. Could someone please help me with the following. I can't get the board to run Dual Channel mode with two sticks of Gskill TCCD and an HTT of 215Mhz or higher. It wil boot, but during the loading screen of Windows XP i'll get nice BSOD. Even with the most relaxed timmings (like 3-4-4-8 @ 2T) it's a no go. My 2x512MB TCCD will complete 40 runs of Test 5 and 8 of MemTest flawlessly with the following settings: 10x285HTT, vdimm 2.8 2.5-4-3-7 @ 2T in dual channel mode (yellow and orange slots both seem to work) So it seems to me those two sticks themself are stable. With one stick I can boot into Windows running on 10x285 2.5-4-3-7 @ 1T and everything works perfect. I've been swapping the sticks in the yellow and orange slots, upping vdimm etc. Using dividers will result in no post at all, or BSOD's again What the deal here? i'm really running out of options Btw, my Gskill ran perfect on my old nF3 UT 250GB (10x290 2.5-3-3-7 @ 1T) (offcourse LDT = 3x and 1.75v)
  10. I've had the same error last night. What worked for me, to get rid of this error ,was setting DDR Output Driving to Weak and setting DDR DQ Drive Strength to 50% reduction. I'm using TCCD Gskill and found some info in another topic, one user reported that when setting DDR DQ drive Strength to Auto, 15% or 30% his rigg would not complete a boot at all, or could not boot fully to windows or get his raid array to be detected. So I thought; maybe the 'invoke problem' could be related to DDR DQ Drive Strentgh. I went back into the bios, and noticed that the setting was on auto. (i just bought the bord, so some settings were still on default) After setting it to 50%, the error has not appeared again. So i'm guessing this could be a workaround. Good luck
  11. Try Priming with only one stick of 512MB RAM inserted. There are some known problems with running Prime on the A64 platform with 2x512MB. It worked for me, I can prime @ 2900MHz. When I use 2x512MB Prime will fail in less than two minutes running also @ 2900MHz. All of my games, 3D Marks en DVD re-encoding software are rock solid with 2x512 @ 2900MHz. Only Prime can't handel it , if I use 1 stick Prime is also rock solid for 6 hrs torture test.
  12. Lol! Those 3400+'s can go very fast indeed I'm doing 2910MHz PiFast Stable, but Prime will fail in half an hour. 2880MHz seems to be stable for 5 hours. That's with vcore 1.7v chipset 1.7, vdimm 2.8v I have to mention i'm using waqtercooling with a parallel radiator setup en two pumps in series, combined with a heatspreaderless 3400+ New Castle and a DTek White Water. Final result is Hexus PiFast 42.66sec Maximum PiFast stable OC is: 10x291 with 2x512 @ 2.5-3-3-7-1T [bimg]http://home.versatel.nl/demoneyez/2910MHz_4266.jpg[/bimg]
  13. The new bios enables me to to do 4-way interleaving with two sticks of 512 TCCD. It gives me 100MB/s more in Sandra. It also boosts PiFast with 1 whole second Final result is 42.66sec Maximum PiFast stable OC is: 10x291 with 2x512 @ 2.5-3-3-7-1T [bimg]http://home.versatel.nl/demoneyez/2910MHz_4266.jpg[/bimg]
  14. RGone Can people that are not a member of DFI-street also do the poll? I've posted a link to this poll on a Dutch forum where several users are also having the multi problem, but a bunch of them arent a member, perhaps that's also the reason you don't get many votes for the poll? (not only from those Dutch people but other users as well)
  15. Yes, it is from week 9 this, unfortunately for you this one hasn't got the new SOI proces.But..., even without the new SOI my last A64 3000+ did 275MHz stable, so there's still a good change that changing vdimm and DRVstength accordingly will get you on your way
  16. You could try to put the sticks in slot 1 and 3, instead of 1 and 2. If that doesn't work try swapping the sticks again, but now put the stick you just put in slot 1 in slot 3 and the stick that you put in slot 3, in slot 1. If it doesn't resolve your problem, put the sticks back in slot 1 and 2. Now go into the 'GENIE BIOS' and look for a setting called 'DRIVE STRENGTH'. It maybe set to AUTO. Set it to '2' or '1' I discovered that using a vdimm of 2.5v - 2.7v requires DRIVE STENGTH set to '1' For a vdimm of 2.8v - 3.1v it requires DRIVE STRENGTH set to '2' First I was doing 2x512MB (Gskill PC4400) at 10x280MHz at 1T. Memstest would give about a hundred errors, so it wasn't stable. This was with vdimm set to 2.7v and DRIVE STENGTH set to '1' After setting vdimm to 2.8v and DRIVE STENGTH set to '2' , I now am able to do 10x290 without any errors, and can do 25 passes with MemTest which is telling me it's stable One could say that upping the vdimm alone is not eneough, it seems to be a very delicate balance between vdimm and DRIVE STRENGTH. DRIVE STRENGTH is the setting that controls the actual signal strength in the memory-controller (which nowadays is intergrated in the CPU itself by AMD). So you'll also need a good memory controller which can do high mem speeds. Mostly processors after week 34 of this year (eg. 0434 XPMW , will have a better mem-controller. This is due to the fact that AMD refined their SOI ( Slicon On Insulator) proces a bit further.
  17. You're right, I meant a spring clip (My English isn't that good, sorry)
  18. New TCCD can handle high voltage very wel, but only at CAS2 not at CAS 2.5 or 3 My friend Kakaroto can do 255 at CAS2 with 3.1v on his 512 Gskill TCCD. (There's a thread on Xtreme with screenshots and guide ) Mind you, that you have to make sure you got the Samsung TCCD 431 chips and not the TCCD 437. I myself got the TCCD 437 chips on my Gskill and they don't seem to like high voltage on any CAS Latency. However, I am able to run 290Mhz (that's with two sticks of 512MB and CPC ON) with 2.8v and DRVstr set to Level2 If I do 280MHz with 2.6v, 2.7v or 2.9v with DRVstr set to Level1 MemTest will give nnumerous errors. There seems to be correlation between the correct vdimm with the appropiate DRVstr. As for misalignment of the mounting holes, my own board seems to have two mounting holes out of alignment. The AGP slot and Chipset have been put far to dense on eachother. I'm using a Radeon X800XT PE with BGA heatsinksinks on the ram chips. I'm also using watercooling so there's a block on my Chipset. This combination makes it imposible to reach the little leaver of the agp slot once the X800 is installed. To little space and to many condesators near the slot, make it a hell of a work to get my card out of the agp slot. So I broke that leaver of.
  19. Hah! And the ones that keep asking and give their opinion will get their posts or topic trashed! :mad: (multi 9x lives? http://www.dfi-street.com/forum/showthread...15&pagenumber=2
  20. With Jesse's bios I could change everything (HTT and multi) in the bios and only the multi with Clock-Gen, but not the HTT. That gave me a freeze With Oskar's bios I can change everything (HTT and multi) in the bios and only the HTT with Clock-Gen and not the Multi, which freezes when applied. Mind you that I can now choose 9x at startup, Windows will boot properly. But when I run Memtest, test 5 will give me ten thousands of errors at 98%. That's running with a HTT of 250MHz. 8x250 or 10x250 will give me 5 passes. I think there are two 'problems' here. One is that running a high HTT is possible with two sticks of 512 with 1T. But you'll need a good proc with a good overclockable mem-controller. Two, there is somekind of multi problem- which looks to me - is bios related. That would explain why with one bios it works better than with the other one. Also reflashing more than one time and in what manner you flash seems te be important for certain cases.
  21. RGone I've flashed the bios, and you are correct. Clock-Gen does work concerning the multiplier, I can now again change them on-the-fly. But what I can't change anymore is the HTT bus. When I change it from 250 to 260 the system freezes. But then again, as you noted, I'm using a CG en Jon was using an CO. Anyways, I'm curently running 8x325HTT 5/6 and everything is stable in Windows (Prime, Sandra PiFast) Memtest will still give few errors in test 5, but hey, so be it. I'm currently on the lookout for a A64 3400+ batch week 35 or later. Seems that AMD is getting better yields, which results in better mem-controllers. By the way, the 3400+ from batch week 35 still has the CG designation, so AMD apparently has not revised the mem-controller. One thing that struck my surprise was, when I flash with AWD Flash ( parameters /py /sn /cd /cp /cc /cks /R) in a DOS enviroment this results in a faster bootime and higher in-BIOS response than when I flash the same BIOS in Windows with Winflash. I get a 4 second faster boottime when I flash with a bootfloppy with AWD Flash. I thought it was a coincidence but I tried both methodes and I kept getting the same results.
  22. Hmmm, very interesting. I'll have to try this bios. I'm currently using Oskar 10/15 bios. If I remember correctly, Oskar's bios uses a newer version of nVidia Raid, namely 4.66 sata bios. Would that pose a problem? I'm currently using an Raid0 Array on SATA 3 and 4 Or is there any way to keep my 4.66 SATA bios while flashing the Jesse 8/27 bios?
  23. Here's mine: 2x512MB Gskill TCCD in slots 1 & 3 Command per Clock enabled 1T vCore = 1.75v vDimm = 2.6v vAgp = 1.6v vChipset = 1.7 CPU idle 29 degrees celcius, PWMIC 35 degrees celcius CPU stressed 39 degrees celcius, PWMIC 52 degrees celcius HTT 265 MHz Multi 10x LDT 3x CPU Speed 2650MHz (A64 3000+ NC) Command per Clock enabled 1T C'nC = off Tcl = 2.5 Trcd = 03 Tras = 05 Trp = 03 Trc = 07 Trfc =12 Trrd = 00 Twr = 02 Twtr = 01 Trwt = 02 Tref = 4708 166MHz, 1.95┬Ás Twcl = 01 Skew value = 255 Drive strength = Level 2 Max async = 7ns Preamble = 5ns R/W bypass = 16x Bypass max = 7x 32B Granularity =Disabled (8 burst) Sandra Bandwith buffered = 4070MB/s Pifast = 47,11s 3Dmark2001 = 28120 (with Ati X800 XT PE @ 600/1200)
  • Create New...