Hit my old 1700 with some mean voltage by accident, need to know if it will be ok

Discussion in 'AMD Processors' started by Kato1144, Mar 21, 2019.

  1. Kato1144

    Kato1144 Limp Gawd

    Messages:
    324
    Joined:
    Oct 18, 2007
    Some of you might be aware that im building a PC for my brother with a Gigabyte AB350 Gaming 5 and my old 1700, So I got it built tonight and was doing some overclocking to get it to around 3.8GHz with a 3x3 configuration. Well I didnt quite understand how gigabyte was doing there stupid voltage offset (Why i could not just input the voltages is beyond me) and there is nothing in the manual on how that works and surprising little online about it either so I had to just assume and assumed it was going off the the lowest base like 1.1v for the core and 0.9v for the soc (the last board that i had that only did offsets did it like that) and so with that i gave the vcore a offest of +0.250 for a vcore of 1.35 and a vsoc of +0.200 for a vsoc of 1.1v or so i thought, I used HWinfo and look at the voltages and Vcore red 1.3v and Vsoc read 1.15v so all looked good and i stared stress testing.. about a hour in i noticed my temps were out of control on the CPU and VRM, so i did some more research and some one had listing the offset and the output voltages and fuck me I was running the Vcore at 1.5v and the Vsoc at 1.3v, and of coures i was looking at the voltage readout according to the CPU not the MOBD on HWinfo and sure shit I was just pounding the poor guy with way way to much voltae and it was under a stress test for over a hour by the time i realized.....

    So I did set everything correct a Vcore of +048.0 and a Vsoc of +0.024 and so far so good, she is doing stress test right now but do you guys think I might have really hurt the chip, it was a short time to run it like that but i was way beyond the safety zone and every one says never go over 1.2Vsoc well i did for a while and now im worried, very worried.... I will see if the stress test pass and if they do I will be lest worried but for now quite upset that it happend and quite mad that Gigabyte made it this way and then did nothing to inform you on how it worked, like if they had a reference voltage listed next to the option so you knew what to add your offset too that would have been great but there is nothing, just stupid :(


    BTW before you bash me for not doing my research before hand I did and I found mostly reddit threads of ppl taking about there overclocks and I think some of them were just as misinformed as me because abunch of there guys were running there 1600 at a 0.200 vcore offset witch is 1.5vcore but they did not seem to know that and that is part of the reason i started my offsets there thinking that was normal, I'm really dumbfounded to why there was nothing in the manual besides "Advanced Voltage Settings: This sub menu allows you to set CPU, chipset and memory voltages" that is word for word from the manual and would only be an acceptable explanation if i could directly input the voltages. Anyway thats my gripe but not why im here, let me know how bad my dumbdumb was, in the mean time im going to start pounding my head against the wall to make the pain stop
     
  2. mda

    mda [H]ard|Gawd

    Messages:
    1,410
    Joined:
    Mar 23, 2011
    If it's still alive and kicking, then you haven't fried it.

    In all likelihood, you probably docked a year off it's 20 year lifespan.
     
    Kato1144 likes this.
  3. sirmonkey1985

    sirmonkey1985 [H]ard|DCer of the Month - July 2010

    Messages:
    21,068
    Joined:
    Sep 13, 2008
    honestly wouldn't really worry about it.. if it was 1.3v vSOC over a long period of time then yeah there most likely will be issues later on down the road. pretty much agree with mda except that the expected/designed lifespan for a cpu is 10 years not 20 ;) .
     
    Kato1144 likes this.
  4. Kato1144

    Kato1144 Limp Gawd

    Messages:
    324
    Joined:
    Oct 18, 2007
    kk thanks quite worried after i had done something so dumb
     
  5. Legendary Gamer

    Legendary Gamer Limp Gawd

    Messages:
    497
    Joined:
    Jan 14, 2012
    I had my 1700 at over 1.4 when I was pushing it at 4,100 MHz. It's been pretty bullet proof. I expect it to run forever now since I turned it into a file server and down clocked it to 1.5 GHz per core . You should be fine.

    I eventually had to bleed off the clocks and voltage on it over a years time until I got a stable OC of 3.9 GHz. I haven't had a single issue with mine and I beat the hell out of it.
     
    Sulphademus and Kato1144 like this.
  6. Rvenger

    Rvenger [H]ard|Gawd

    Messages:
    1,658
    Joined:
    Sep 12, 2012
    You are fine. They are pretty damn durable chips.
     
    Kato1144 likes this.
  7. funkydmunky

    funkydmunky 2[H]4U

    Messages:
    2,218
    Joined:
    Aug 28, 2008
    What is a 3x3 configuration?
     
  8. Nenu

    Nenu [H]ardened

    Messages:
    18,579
    Joined:
    Apr 28, 2007
    I did similar to a 2500K when the motherboard got a bios update with LLC.
    On the 2nd test boot I accidentally left LLC on max and did some stress tests.
    After about 1/2hr I noticed temps were really high and the chip wasnt the same from then on. I'd given it almost 1.5V for that 1/2hr when it was previously on 1.42V.
    It immediately dropped from 4.6GHz stable to 4.5GHz and then slowly degraded to 4.2GHz while I had it, and is now doing Dad duty at 3.7GHz.

    That isnt to say the same will happen to you, but its worth bracing yourself.
     
    Kato1144 likes this.
  9. DTN107

    DTN107 [H]ardness Supreme

    Messages:
    4,686
    Joined:
    Jun 30, 2008
    It's game over man.

    Now sell it to me for $20.
     
    Hakaba, Legendary Gamer and Kato1144 like this.
  10. MyNameIsAlex

    MyNameIsAlex Limp Gawd

    Messages:
    209
    Joined:
    Mar 10, 2019
    I've seen procs run 24/7 for 25 years. The proc in question is still doing fine.
     
  11. MyNameIsAlex

    MyNameIsAlex Limp Gawd

    Messages:
    209
    Joined:
    Mar 10, 2019
    I've seen procs run 24/7 for 25 years. The proc in question is still doing fine.
     
  12. Kato1144

    Kato1144 Limp Gawd

    Messages:
    324
    Joined:
    Oct 18, 2007
    Dam dude, hopefully it does not degrade but i have it on a pretty light OC its 3.825GHz 1.30Vcore and 1.122Vsoc, I would trim the vcore down more but I dont have a load line calibration option so I can really change the way the MOBD gives the CPU a boost under load and i dont think it would be to stable any lower but I am running 2 less cores so the power demand should be somewhat lower when it was running 3.835GHz on all 8cores in my main rig, i should try trimming to voltages more down the road but this setup is 100% stable as far as i can tell

    you can basically turn off cores in the CCX modules 3x3 is how the cores are arranged in the CCX in a 6core mode, this being a 8core chip it can do 1x1 2x2 3x3 4x0

    I made it 3x3 because the AB350 does not have really enough power delivery to do a 8core CPU, infact when we were updating windows after the fresh install i left the bios settings default and we had 2 crashes, turned on the 3x3 not one crash since


    Thanks man I needed the laugh :)


    I also wanted to thank everyone else, the reply to this were very helpful, I knew the Zen chips are pretty tough but I never seen anyone push the Vsoc to such stupid levels and I knew 1.5vcore was not good either, so thanks for the assurances, im going to keep a eye on her alittel bit maybe see how low I can push the voltages but it should be very happy were it is now just need to 8-10 hours on prim95 blend test to make sure the CPU and MEM are in harmony.

    BTW does anyone still do prime95 stress test in the 8 to 10 hour range or am i just overkilling it, it what i was taught in the core2quad days but back then overclocking was much more of a science and really easy to screw up, there was core ratios man... ratios that required a spreadsheet to figure out.

    anyway thats enough from me thanks again for all the help :)
     
  13. sirmonkey1985

    sirmonkey1985 [H]ard|DCer of the Month - July 2010

    Messages:
    21,068
    Joined:
    Sep 13, 2008
    oh definitely, hell i have an old socket A 1700+ overclocked to a 2000+ still running strong. but they're still only guaranteed for 10 years even if they'll survive longer than that if they're well maintained.

    definitely could see that happening with the 2500k, those things do not like high voltage unless watercooling/LN2. wasn't recommended max like 1.4v for that thing anyways?
     
  14. Legendary Gamer

    Legendary Gamer Limp Gawd

    Messages:
    497
    Joined:
    Jan 14, 2012
    FYI I was using and still am an MSI Tomahawk with my 1700 and I had it dialed into 3.9 on all 8 cores, stable.

    Anyone would be hard pressed to do much better on an X370 board and you're only likely to get 100 to 200 MHz based off the silicon lottery if lucky, on a first gen Ryzen. The need for a high end board with the Ryzen parts is really overrated. I'm on a 100-120 dollar motherboard running my memory at 3200 MHz (Corsair designed for Intel, non Samsung B die) though it took MSI a good long while to get the BIOS microcode updates right .

    That AB350 might be a little less robust, but I bet you could still get all the CCX'es rocking if you needed em. An A320 will do it at stock, so unless you've already verified your silicon lottery is poo, you might dial it back up and enjoy.

    As a file server at a down clocked core of 1.5 , this thing is an absolute beast. It sustains a maximum transfer rate to my other computers of 113 Megs a second (which I believe is about the max for my gig network at home). Internally it's rocking up to about 200 Meg's a sec on the array and I don't even think that's the limit because I have that older Adaptec controller in a 4x slot instead of rocking full 8. Really need to swap that crappy AMD 3250 down to the 4x slot and see if the system posts (moving the Adaptec to the 16x slot). One of these days when I get around to fussing with it again. Til then it's fine for anything I throw at it .

    Edit: I was wrong... Max transfer rate on 1000 Mbps Ethernet is 125/Megs a sec. Looks like I need to do that experimentation and move that Adaptec controller up a slot or... It's just a limitation of my Ethernet controller on that board. I've never seen above 113 megs to clients... Internally (on the server) much faster.
     
    Last edited: Mar 22, 2019
  15. Nenu

    Nenu [H]ardened

    Messages:
    18,579
    Joined:
    Apr 28, 2007
    I agree, it was on custom water.