Page 3 of 6 FirstFirst 123456 LastLast
Results 51 to 75 of 129

Thread: Highest DDR Sandra Benchmark

  1. #51
    Xtreme Member
    Join Date
    Sep 2002
    Location
    Melbourne, Australia
    Posts
    390
    Originally posted by Dirty_Punk
    ah... i'm not good with bios mods, i've only read in the site from pahntompunisher an improvement over original bios that Enhanced ram is always set to on... don't know if this is true or not...
    Actually I think it's always set to 16T now. Enhanced DRAM on Abit boards is DRAM Idle timer on other boards. Enhanced On = 0T, Enhanced Off = Infinity (shows as greyed out in CPU-Z).

    ...and yes, go go go Epox 4PEA

    F@H boxen 39,869 MHz - Standford stats - EOC stats - Folding 4000 Club
    Dugu 9 Tweaks: 3DMark 2001 guides || RAM timings || Recommended RAM settings for P4 DC DDR system || How-to: Overclocking Athlon64
    My PC: A64 3200+ :: Shuttle AN50R :: 2x256MB XMSX3500 @ 270Mhz 2.5-2-2-6 3.4V& 512MB XMS3200LL @ 245MHz 2-2-2-6 3.1V :: Excalibur HIS 9800Pro 128MB :: CMV CT-722D 17inLCD :: 3DMark 2001: 24515 || Watercooling coming "real soon now"...

  2. #52
    Xtreme X.I.P.
    Join Date
    Jun 2002
    Location
    Belgium
    Posts
    4,477
    Originally posted by Dirty_Punk

    PS: have you still got the modded bios with turbo options?? if you have it, can you send me?? e-mail is ghost_2512@yahoo.it
    tnx! [/B]
    For which version do you want the Turbo option enabled?? C7, CF or CV ?
    Rig #1
    Gigabyte P67A-UD4 trying to figure out this POS board
    2600k @ ?????
    2x2Gb GSkill RipJaws-X 1333 (7-7-7-21)
    ATI 5850
    Coba Nitrox 750W
    Watercooled with HK 3.0 CU, Watercool GPU-X³ 5870 Nickel, PA120.3, Laing Ultra with XSPC top


    Rig #2
    DFI UT P35-T2R (0317 bios)
    E8200 @ 4000 (1.216V) / 4100 (1.248V) / 4200 (1.296V) / 4300 (1.344V)
    2x2Gb Chaintech Apogee GT PC2-8500
    Powercolor 4870
    Corsair 520HX
    Watercooled with HK 3.0 CU, EK-FC4870, Feser tripple, Laing Ultra pump


    Quote Originally Posted by Movieman View Post
    I've got Supermicro boards that lasted longer than one of my marriages!

  3. #53
    Xtreme Member
    Join Date
    Aug 2002
    Posts
    361
    Originally posted by Jupiler
    For which version do you want the Turbo option enabled?? C7, CF or CV ?
    it's the same, there is no difference for my use
    tns again

  4. #54
    Xtreme X.I.P.
    Join Date
    Jun 2002
    Location
    Belgium
    Posts
    4,477
    Originally posted by Dirty_Punk
    it's the same, there is no difference for my use
    tns again
    All 3 modded bios's are underway. As usual, use at own risk.
    Rig #1
    Gigabyte P67A-UD4 trying to figure out this POS board
    2600k @ ?????
    2x2Gb GSkill RipJaws-X 1333 (7-7-7-21)
    ATI 5850
    Coba Nitrox 750W
    Watercooled with HK 3.0 CU, Watercool GPU-X³ 5870 Nickel, PA120.3, Laing Ultra with XSPC top


    Rig #2
    DFI UT P35-T2R (0317 bios)
    E8200 @ 4000 (1.216V) / 4100 (1.248V) / 4200 (1.296V) / 4300 (1.344V)
    2x2Gb Chaintech Apogee GT PC2-8500
    Powercolor 4870
    Corsair 520HX
    Watercooled with HK 3.0 CU, EK-FC4870, Feser tripple, Laing Ultra pump


    Quote Originally Posted by Movieman View Post
    I've got Supermicro boards that lasted longer than one of my marriages!

  5. #55
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Dirty Punk - I thought your DDR510 bench was a stable OC. I can fully believe you can run 246 through 3dmark with that stick.

    The benches I posted above are my STABLE settings. Ive currently started doing DVD rips to burn in the memory (Thanks BigToe ) and have had no problem at all with stability so far. I backed off to DDR462 and lowered the mem voltage to 3v, and have so far improved my mem FSB to DDR474 using 3v still. Before I would need 3.1 to 3.2v for total stability at these speeds.

    482 would just about complete a loop of 3dmark without dropping to desktop, but only from a cold boot. Now I can loop between 8 and 10 passes of 3dmark before it drops to desktop . Im slowly getting there.

    If BigToes theory about burn-in merely reducing errors at any given FSB rather than increasing the FSB is correct then I should be able to stabilise it at DDR494 using maxed timings. Now I need to find a CPU that will do that sort of FSB with a decent overall clockspeed. The 1.6a is nice, and I kept it due to the high fsb tolerance, but in order to compete at 3dmark you really need at LEAST 3.5ghz. Memory bandwidth helps to an extent, but you really cant make up for over 500mhz cpu speed using memory bandwidth alone.

    PiLsY.

  6. #56
    Xtreme Member
    Join Date
    Aug 2002
    Posts
    361
    your memory is stable at 520mhz??
    mine at 510 not, but at 492 is enought stable to run 3dmark, don't know if have stability problem, but works fine with 3dmark.. i'm working for 496, hope that will work with ram heatsinked, always 2-5-2-2 at 2.96-7v

    my 2400B works fine with that clocks (185fsb), for more mhz i need better cooling, watercooling with 3° water isn't enought... need to make a phase-change like prommy homemade, but i'm a newbie in that filed, so need to practice a bit

  7. #57
    Xtreme Legend
    Join Date
    Jun 2002
    Location
    Helsinki, Finland
    Posts
    2,559
    Dirty punk, My ram and mobo are acting exactly like yours: can do 3dmark @185/247 2-5-2-2 and sandra @253mhz but that seems to be the limit no matter how slow timings I use.
    ps. can you send the pcr data?

  8. #58
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Ive still got the 187fsb wall here too. 186fsb 100% stable, 187 and it wont even post. Tried 1:1 memory and max vcore (up from 1.63v to 1.85v) and everything else I can think of. It just wont post.

    However, I dont think ill be seeing 185fsb anytime soon with a top end chip, so im happy with it.

    Its strange how I could boot all the way up to 200fsb at first though. Very strange. This only happened after I put the winbond memory in the board. It wouldnt post with the winbond in, and now it wont go over 187fsb.

    Yes DDR520 was stable, however, performance was poor compared to tight timings and lower mem FSB, so I went back to DDR482 for 3dmark again.

    PiLsY.

  9. #59
    Xtreme Member
    Join Date
    Aug 2002
    Posts
    361
    Originally posted by macci
    Dirty punk, My ram and mobo are acting exactly like yours: can do 3dmark @185/247 2-5-2-2 and sandra @253mhz but that seems to be the limit no matter how slow timings I use.
    ps. can you send the pcr data?
    sorry, tomorrow i send it to you. Today i couldn't because i'd tryed a KT4 Ultra with a 2400+, but get very poor overclock and lost a lot of time do try to optimize this mobo... i haven't got a good image from KT4, poor voltages (ok, i've voltmodded it, but stock haven't good voltages), bug in bios, unlock doesn't work well..... hope to try a nforce 2 mobo like abit or epox soon, with better cpu i hope

  10. #60
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Picking up a Leadtek Nforce2 board next week. Should provide some interesting results. Its VERY cheap for a "Pro" board (ie has full soundstorm) at only £83 compared to £100 for Epox and £134 for Asus. The main reason in my choosing it though is that it has jumpers for the vdimm, not bios adjustment. At stock it provides 2.5/2.6/2.7v (although there is blatently an undocumented combination of the 2 jumpers which should provide 2.8v). Historically jumpered vdimm motherboards have been very easy to vdimm mod simply by either wrapping jumpers together or by jumpering between a jumper pin and ground. I really cba with lifting pins on that voltage controller - its TINY.

    www.leadtek.co.uk has the info on it if anyone else fancies taking a look. The layout looks ok, plenty of room between dimms and AGP and a ton of room around the northbridge. Its also got the 4 holes around the cpu socket.

    Will of course report my findings, but gut instincts count for a lot in this game, and my guts telling me its going to be a good board. 211fsb limit though, but tbh thats not really a big deal to me at this point. Im sure either leadtek will come out with a bios releasing higher FSBs, or if im really that bothered ill learn how to make a TurboPLL.

    PiLsY.

  11. #61
    Xtreme Addict
    Join Date
    Jun 2002
    Location
    Melbourne, Australia
    Posts
    1,118
    PiLsY is that the 8rDa+ or something? If so then theres a bios that goes up to 250fsb now A friend just got the board and he is having Car Low problems and APIC is enabled so he doesnt know what it is....

    Also with you being able to boot at a certain fsb but now cant, I was able to boot into windows at 171 for a screenie but now cant even boot into windows at 168 anymore....wtf??!!!

    3.06 should arrive next week sometime so just going to be a gamer till then :p
    hehehehe

    Intel P4 3.2 EE || Asus P4C800-E || Prometia Mach II || 2 x 256Meg HyperX 3500 || Antec TrueControl|| Sapphire Radeon 9800XT || 2x Raptors || Samsung 172T

    OCAU 3DMark Team || 3DMark Compare: 26847 || World of Britney|| PCDB Entry

    "Sundance is weird. The movies are weird. You actually have to think about them when you watch them. " -Britney Spears

  12. #62
    Registered User
    Join Date
    Nov 2002
    Location
    Singapore
    Posts
    9
    PiLsY, the last I saw there was no multiplier option of any kind on the Leadtek nforce2 board. I'm not sure if they have it now with a new bios.

  13. #63
    Xtreme Enthusiast
    Join Date
    Jun 2002
    Posts
    668
    Originally posted by PiLsY
    Picking up a Leadtek Nforce2 board next week. Should provide some interesting results. Its VERY cheap for a "Pro" board (ie has full soundstorm) at only £83 compared to £100 for Epox and £134 for Asus. The main reason in my choosing it though is that it has jumpers for the vdimm, not bios adjustment. At stock it provides 2.5/2.6/2.7v (although there is blatently an undocumented combination of the 2 jumpers which should provide 2.8v). Historically jumpered vdimm motherboards have been very easy to vdimm mod simply by either wrapping jumpers together or by jumpering between a jumper pin and ground. I really cba with lifting pins on that voltage controller - its TINY.

    www.leadtek.co.uk has the info on it if anyone else fancies taking a look. The layout looks ok, plenty of room between dimms and AGP and a ton of room around the northbridge. Its also got the 4 holes around the cpu socket.

    Will of course report my findings, but gut instincts count for a lot in this game, and my guts telling me its going to be a good board. 211fsb limit though, but tbh thats not really a big deal to me at this point. Im sure either leadtek will come out with a bios releasing higher FSBs, or if im really that bothered ill learn how to make a TurboPLL.

    PiLsY.
    LardArse had one of these (think it was in the Vmod's forum he was looking for info on Vdimm/Vcore modification). I'd definitely recommend the Epox over the Leadtek nForce2.

  14. #64
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    Hmm...

    Well the no multiplier thing WOULD be a problem. Theres no way im hacking bridges.

    Will look into it - thx for the heads up! I owe you one m8 .

    PiLsY.

  15. #65
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    Can you explain me something?????...........I payed a lot of money buying Corsair XMS3200 C2 witch can do on this crap Mobo(IT7-Max2 V.1) 256MHz(512MHz DDR) and with ONLY 100 Euros I bought a SIMPLE Kingston 333DDR witch can do WITHOUT a lot of burn-in 262MHz(524MHz DDR)........It has Windbond chips on it 6ns like Corsair XMS3200 C2 one.......It drived me CRAZY....!!!!!!
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  16. #66
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    S-Sandra Memory Bandwidth....Buffered
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  17. #67
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    S-Sandra Memory Bandwidth .....Unbuffered.....
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  18. #68
    Xtreme Member
    Join Date
    Nov 2002
    Location
    Netherlands
    Posts
    259
    Hmz..

    how does this 'unbuffered' thing work? if i do sandra mem benchmark unbuffered or at least, if i did it right (i de-selected the use buffer thingie or so, in mem bench options). i only get some score under 1k... that aint good uh?
    XP1700+ Watercooled @ 1660 mhz max
    Abit KD7-Raid
    Kingston HyperX PC3200 (512mb)
    Leadtek GF3 Ti200 TDH @ 215/530

  19. #69
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    ......and with Cas 2-2-3-7
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  20. #70
    boughetto overclocker
    Join Date
    Jun 2002
    Location
    NORAD Command Center
    Posts
    2,012
    Sounds like you've got some good DDR man!
    Dinosaurs are chortling.
    Lian Li PC1200B, Custom Water Cooling, Lamptron FC5 Fan Controller w/ 4x 120mmx38 Deltas
    i7 930 @ 3.8ghz, Rampage III Extreme, 12GB G.Skill PC2000, BFG 260GTXOC, PNY Quadro 380FX
    (1) HP LP2475w 24", (2) Dell 2007WFP 20", Corsair AX1200 1200w, 3Ware 9650SE Raid
    (2) 500gb Hybrid Seagate Raid 1, (2) 2TB WD Red NAS 64mb Raid 1, (5) 1TB Hitachi 32mb Raid 5

  21. #71
    Xtreme Member
    Join Date
    Sep 2002
    Posts
    261
    Originally posted by bigtoe
    Guys please DITCH the buffered scores as they PROVE NOTHING!! they scale with fsb and as such cannot be trusted.Just bench unbuffered and show how fast the ram really is.Also show a sceeny of memtest86 at startup as this shows an unbuffered test too.
    My best Unbuffered Test here
    RDRAM @ 162 4x Turbo ON

    [size=0.5]| P4 NorthWood 3.0C @ 4350 Mhz (290 x 15) 1,872V Cooled by Ybris T-Rex @ 13,5V and Powered by TTGI 520W | Abit IC7 | 2*256 mb A-Data Winbond BH-5 Dual CH | Watercooled & Modded HIS 9700 "non Pro" @ 441/351 @ 1,95V/3,40V | Monitor 19" Tatung Vibrant | [size=0.5] | ICQ: [/size][size=0.5]174732553[/size][size=0.5] | [/size][size=0.5]**[/size][size=0.5] Diablo.II@email.it [/size][size=0.5]**[/size]
    3DMark Score:
    [/size][size=0.5]Class Ti4600 ( [/size][size=0.5]16606 pts. [/size][size=0.5])[/size][size=0.5] , [/size][size=0.5]Class Radeon 9700 ([/size][size=0.5]21505 pts. [/size][size=0.5])[/size]

  22. #72
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    Updated Info.....

    Memtest-86 with NO ERRORS at 254MHz(508MHz DDR)with Cas 2-2-3-7
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  23. #73
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    ......and after a few minutes more.....
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  24. #74
    k|ngp|n/Sham my brothers
    Join Date
    Aug 2002
    Location
    Athens---Hellas
    Posts
    5,693
    .....and Winbond Monitor(for accual Vmem....
    INTEL PWA FOR EVER

    Dr. Who my arss...

    .........



  25. #75
    Technician
    Join Date
    Jun 2002
    Location
    Merseyside, UK
    Posts
    2,661
    HiPro5

    I test for stability by looping memtest on test #5 only for between 2 and 6 hours.

    I start off at 3 passes of test 5 before moving up, when im close to the limit I test to 30 minutes, and when im at the limit 2 hours. After that I find 6 hours of memtest looping test 5 followed by 6 hours of seti and about 12 ripping a dvd will give me an extra mhz FSB.

    The full test of memtest really doesnt tax your memory much. Its only test 5 that really pushes it.

    Try looping test 5 for 30 minutes then touching the memory with your finger :p. Mine got to 62c after 15 minutes before I chickened out and heatsinked it. That was at only 2.8v and DDR466 btw...

    If you can show me a photo of memtest having run at that speed for 2 hours looping test 5 without errors I will be very impressed .

    Nice job anyway, especially with PC2700 mem .

    PiLsY.

Page 3 of 6 FirstFirst 123456 LastLast

Bookmarks

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •