1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Getting over 550mhz core on GTO2

Discussion in 'Graphics Cards' started by hb904460, Mar 2, 2006.

  1. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    10 Year Member at TPU
  2. amiiam New Member

    Joined:
    Feb 23, 2006
    Messages:
    22 (0.01/day)
    Thanks Received:
    0
    some r480 can go up to ~700 mhz
    depends on your luck
    voltmodding should help you
     
    10 Year Member at TPU
  3. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    550 is an okay overclock for the gto2.. u want more u will have to up the voltages.. what causes it is simple.. its reached its limits without higher voltages..

    trog
     
    10 Year Member at TPU
  4. JC316

    JC316 Knows what makes you tick

    Joined:
    Jan 24, 2006
    Messages:
    9,366 (2.19/day)
    Thanks Received:
    993
    Really? My X850 Pro can go to 594 before I get artifacts. It seems like the R480 should go higher than that.
     
    10 Year Member at TPU
  5. dolf New Member

    Joined:
    Mar 16, 2005
    Messages:
    2,058 (0.45/day)
    Thanks Received:
    30
    Normally less pipes = less heat output = higher frequency.
     
    10 Year Member at TPU
  6. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    some 480s got hand picked and stuck on X850 XT PEs.. the rest kinda found their way onto GTO2s.. he he..

    trog
     
    10 Year Member at TPU
  7. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    Why are there no artifacts in ATiTool if the core is being pushed so hard? Normally there would be errors before the OC that pushes the core over the edge so that it crashes.
    Think this is more likely to be a block in the bios or an error in the drivers/AtiTool than the core not being able to take the speed.
     
    10 Year Member at TPU
  8. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    my experience tells me cores when pushed can just crash.. memory tends to show artifacts..

    but the botom line here is that not all 480 cores are equal.. cores guaranteed to work at 540 out of the box (X850 XT PE) would carry a safey margin of around 10% on average.. so they could be expected to make say 600..

    cores fitted to cards guaranteed to run 400 out of the box dont need such a safety margin..

    being as the PE card is/was a flagship card.. they run em as fast as they safely can.. only the best chips would have got used..

    the rest.. well many thousands of em ended up elsewhere.. i have one and so do u.. mine makes around 550 before it crashes so does yours.. ??

    your assumption that any old 480 core should go well over 550 is simply wrong.. many of them wont.. not on standard voltages at least..

    trog
     
    Last edited: Mar 3, 2006
    10 Year Member at TPU
  9. yogurt_21

    yogurt_21

    Joined:
    Feb 18, 2006
    Messages:
    5,073 (1.19/day)
    Thanks Received:
    1,201
    Location:
    AZ
    you'll have to note that he did no bios modding in the burstfire review meaning he's runniing on the same amount of voltage used for the stock 400MHZ core. now i bet if he flashed it to an x850xtpe bios he'd be able to break the 550 barrier easily. seeing as when my freind flashed his gto fireblade to an x850pro he gets to 600/600(1200).
     
    10 Year Member at TPU
  10. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    its possible but your friend is just one guy who got a better than average result and as such is pretty meaningless..

    550 is a good stable overclock for GTO2.. some wont even make that..

    trog
     
    10 Year Member at TPU
  11. yogurt_21

    yogurt_21

    Joined:
    Feb 18, 2006
    Messages:
    5,073 (1.19/day)
    Thanks Received:
    1,201
    Location:
    AZ
    it was a mere example, i've seen plenty r480 core break the 550 barrier, on the same logic the few here who can't are just a few so they become pretty meaningless, just a few that can't. perhaps the gto2 have a limit, perhaps no, but the few examples here prove nothing. granted 550 is a fine overclock, but it doens't change the fact that any x850xt will hit 600MHZ easily.
     
    10 Year Member at TPU
  12. largon

    largon

    Joined:
    May 6, 2005
    Messages:
    2,786 (0.61/day)
    Thanks Received:
    435
    Location:
    Tre, Suomi Finland
    Apparently R423 and R480 tend to clock to a certain frequency without any artifacts, then suddenly the card just flips and freezes. For example my C3D GTO (R423) w/ 16P goes all the way upto 674MHz Atitool artifact free but 675MHz freezes instantly - actually there are some single pixel artifacts but for some reason Atitool won't detect them.

    3DMark05 behaves even more strangely:
    All GPU tests pass @ 650MHz, but currently at anything above 645MHz the CPU tests freeze the system.
     
    10 Year Member at TPU
  13. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    my 550 is an okay overclock for gto2 480 core is based on much reading on the subect..

    not just what i read on here.. i have one so i take the trouble to find out.. the information is all out there.. the average speed of a gto2 480 core is less than the average speed of an X850 XT PE 480 core.. cores fitted to Platinum Edition cards are specially selected ones for obvious reasons.. 480 cores fitted to a gto2 card can be any old crap that works..

    trog

    ps.. we are or should be talking about gto2 480 cores.. not other 480 cores.. thats where the over optimistic expectations come from..
     
    Last edited: Mar 3, 2006
    10 Year Member at TPU
  14. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    Sorry trog100 in my case you are wrong. I just manually set the core to 562mhz in ATItool and ran 3DMark03 (1500 point increase) and played COD2 for 45 minutes and the card runs great with the stock cooler. Have managed to manually set the core up to 588mhz before the card crashed.
    This shows the core can hack the speed so the software is at fault. Most probably an error in the find max core in ATiTool.
     
    10 Year Member at TPU
  15. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    u misread what i am saying HB.. i have run a 565/630.. 3Dmark 2005 with my gto2 480.. that is right on the edge with my card thow not what i would consider gaming stable.. just a benching setting..

    as for find the max core in atitool.. thats for dweebs.. doing it maually is the best way of finding the max i wouldnt trust atitool to do it and never have..

    with a slight memory timing tweak my card scores just over 7100 in 2005

    no way will my core make 588..

    565 is its absolute max (on a good day) and thats with an artic vga cooler.. running nice and cool..

    so u have slightly better one than me.. other folks might have worse.. i think yours is noticeably above average for a gto2 480 core..

    trog

    ps.. at 540/600 i get about 6850 in 2005.. and playing cod2 for 45 minutes dont mean your card is gaming stable.. it just means it can play cod 2 for 45 minutes.. he he.. what happened after the 45 minutes.. did it crash.. or did something more pressing call u away from cod2..

    and u also cant say there is something wrong with atitool if it sets your core speed XX percentage less than its crash speed.. i stick to my original statement 550 is a good gaming stable speed for a gto2 480 core its unreasonble to expect much more..
     
    Last edited: Mar 4, 2006
    10 Year Member at TPU
  16. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    trog, you are missing the point, the initial question for this thread was why does my card crash when it hits 550mhz in ATiTool when using find max and is there a way round this? Your response was that the core couldnt handle the speed. Further testing shows that it however can handle the speed so you were indeed wrong.
    I answered my own question to a certain extent, instead of using ATiTools find max (which many people use to find a safe OC as it detects artifacts before the human eye can) i should have manually clocked the core and checked for stability or graphical errors.
    COD2 is very graphically demanding (see latest SLI vs Crossfire test in CustomPC) and any card that can play at high settings and resolution for 45 minutes is as stable as its going to get. (Had to go play football, won 2-1 thanks!)
     
    10 Year Member at TPU
  17. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    "Wondering if anyone else has experienced the core 550mhz barrier with a GTO2? Whenever i try to overclock the core as soon as it hits 550mhz the card crashes. There are no artifacts in ATi tool at 549mhz or below and the card runs great.
    I see it is mentioned in the MadMoxx GTO2 review by W1zzard."

    ####

    i simply assumed from your first post u were expecting noticably more than a stable 550 overclock from the gto2 core.. to be honest my first reply answered the question..

    there is no "absolute" 550 barrier with the gto2 480 core.. where did this idea come from..

    and dude just cos u can manage 45 minutes of cod 2 on a certain day at a certain time at a certain temp does not mean your card is stable.. when u can play several (most) games for several hours at a time at any ambient temp without problems then u can assume its stable at 562.. until u can dont claim it is..

    there is no 550 barrier.. and a 550 stable overclock is a good one for a gto2 card.. he he he

    of course your last post does contradict your first one.. in one it crashes at 550.. in another it gets to 588.. would i be correct in assuming the low one would be useing atitools find max core and the latter one doing it manually..??

    when i first tried the find max core thing.. it just crashed my entire system and i had no idea what core speed it reached before the crash.. which is why i did mine manually right from the start.. i couldnt see much point in a tool that simply crashed the system leaving me no idea at which point..

    and anyone that says articfacts will allways show up before the crash point is wrong.. even if they do write for CustomPC.. he he.. cores can simply just crash with no warning whatsoever.. which does put atitools find max core at somewhat of a disadvantage.. cos it does rely on the odd glitch before the actual crash..

    memory will artifact.. cores can just crash..


    trog
     
    10 Year Member at TPU
  18. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    "there is no "absolute" 550 barrier with the gto2 480 core.. where did this idea come from.. "
    I think you made that up. I said that i had problems finding max core in ATiTool, when i hit 550mhz the system crashed. As was shown in the Madmoxx review and my own experience with ATiTool.
    You are still missing the point. The core can handle the speed so why does it crash at 550mhz when find max is used?
    Some sort of constructive answer would help.
     
    10 Year Member at TPU
  19. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    thats a fairly straight forward question then.. why does find max core appear to crash the core at a lower speed than manually doing it does.. ????

    assumming i aint missing the point yet again.. he he.. either i aint very clever at getting these "points" or u aint very clever at making em.. ??

    to be honest i cant see any logical reason why it should do this.. that could be construed as a none constructive "dont know"..

    finding max core speeds can mean two things.. the actual core crash point which can vary from test to test in my experience.. and finding a safe stable speed that a nooby could safely run his core at..

    the second being far more important than the last.. i think to be genuinely 24/7 stable a core should not be run too close to its crash point.. a gap of 20mhz minimum is required..

    if i designed an idiot proof auto overclocker i would build in a 10% safety margin and cleary state that going beyond this point was entirely at the users risk and should be done gradually and with much (in-game) testing before settling for true safe stable overclock speed..

    i am a long term tweaker and all i know is it can takes weeks to find genuinley stable settings when u go as close to the edge as u can.. its a lot easier to use the 10% factor..

    its hard finding a max point when that point results in a crash.. one way of doing would be to keep a log and simply state in said log that the last recorded point was xxx.. go back xxx beyond that point for safe running.. another way would simply be to report less than the actual speed and say u can use the lower reported speed..

    atitool is magic peice of kit but like most things it aint perfect.. and it is being used by folks who really lack the knowledeg to be doing such tweakings.. overclocking is just taken too much for granted these days.. it used to be an arcane science.. now its just something everybody does and expects to be able to do it.. he he

    trog
     
    10 Year Member at TPU
  20. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    trog so you dont know the answer and you havent experienced this particular problem yourself. Its probably best to go for a proven accurate answer rather than just guessing or trying to.
    I have been overclocking/modding my home built systems for over 5 years now and have been using ATiTool or powerstrip, and the omegadrivers to do the software side of this with great success. However this is the first time i have had an unexplained problem with the software. I am going to go with my own thinking on this as i know my system better than anyone else. When i establish for certain what is the cause i will update this thread!
     
    10 Year Member at TPU
  21. trog100

    Joined:
    Dec 18, 2005
    Messages:
    5,784 (1.34/day)
    Thanks Received:
    759
    dude u are asking the question.. i rather foolishly have tried to help by attempting to provide an answer to a badly put question..

    and i have been building systems for far longer than u have my friend .. and have had many an unexplained problem.. cured by nothing more than educated guesses.. what i really need to learn is not to waste my time trying to help arrogant ungratefull know it all assholes who really know bugger all..

    lots of folks have read this thread dude so i recon the definitive answer u seek might be harder to find that u think it is.. best of luck in your search thow.. he he

    trog..
     
    10 Year Member at TPU
  22. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    Thats the community spirit trog! keep up the good work! Ha Ha!
     
    10 Year Member at TPU
  23. Ketxxx

    Ketxxx Heedless Psychic

    Joined:
    Mar 4, 2006
    Messages:
    11,507 (2.71/day)
    Thanks Received:
    570
    Location:
    Kingdom of gods
    do it yourself, dont let ati tool do it, i tried it once just to see how effective it was - it wasnt, it found my max core speed, went up a bit more, found artifacts, then kept lowering the core speed, and lower, and lower, and lower....so yeah, dont let ati tool do it. and to cure any questions you have, just flash the thing and see if you can then go higher, im putting odds at 50\50 given what i encounterd, but dont hold your breath.
     
    10 Year Member at TPU
  24. hb904460

    Joined:
    Feb 26, 2006
    Messages:
    41 (0.01/day)
    Thanks Received:
    2
    Cheers Ketxxx! Get my Arctic Cooling Silencer 5 tomorrow and cant wait to have low noise overclocking fun!
     
    10 Year Member at TPU
  25. D-rock New Member

    Joined:
    Jan 26, 2006
    Messages:
    49 (0.01/day)
    Thanks Received:
    1
    You're right Trog. Not all R480 chips are equals. The best of the best are put in the XT PE, then the next best go into the XT's, then the ones that are left are put into the GTO's and GTO2's. They can be any old R480 because they only have to work at the advertised speed of 400/490. And not all R480's in GTO2's even reach the XT speeds of 520/540. Read throughout this forum and other video card forums on the internet and you'll see a lot of people complaining that their cards only OC up to around 500-510 on the core. I believe the majority of the R480's in the GTO2's will get up to XT and possibly XT PE speeds. I know mine runs very well at 540/590 stock PE speeds. After hours of gaming its steady at 65deg C core temp with stock cooling fan running at 100%. Thats why I havent tried to go higher until I get a better cooling fan.

    One thing I did notice is that as soon as I flashed the 16pipeline bios to the card it ran a lot hotter. Stock 12pipes: 35deg C idle, 55deg C max under load. After flashing with 16pipe bios from techpowerup website 47deg C idle 65deg C max under load. I know in the modded bios the dynamic fan speed table is not accurate, because the fan wont go over 54%. So in ATITool I went in and set it to go 100% fan on any temp over 50deg C and anything under that the fan is at 50%. Did your card run that much hotter just by flashing the bios even at the stock GTO2 speeds of 400/490??
     
    10 Year Member at TPU

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)