RX 580 Hashrate

1246711

Comments

  • eireeire Member Posts: 43
    when somebody posts these figures:

    1130/2150 -96mv

    1130 is core?
    2150 is mem?
    -96mv is the voltage setting?

    Do they leave the straps alone too or edit them also?
  • memsmems Member Posts: 3
    eire said:

    when somebody posts these figures:

    1130/2150 -96mv

    1130 is core?
    2150 is mem?
    -96mv is the voltage setting?

    Do they leave the straps alone too or edit them also?

    all correct (-96mv is core voltage offset) & usually memory straps are bios edited.
  • technologytechnology Member Posts: 3
    edited June 2017

    Same memory manufacturer for your 4gb and 8gb card? Check both with gpuZ and show us your results.

    Found out that there are 2 versions from gigabyte. One with samsung and one with hynix memory and i got the hynix memory. Copied from 1750 to 2000 and 2250 but i think 27mh/s is the end while overclocking to 2150 on memory.
  • Wolf0Wolf0 Member Posts: 329 ✭✭✭

    Same memory manufacturer for your 4gb and 8gb card? Check both with gpuZ and show us your results.

    Found out that there are 2 versions from gigabyte. One with samsung and one with hynix memory and i got the hynix memory. Copied from 1750 to 2000 and 2250 but i think 27mh/s is the end while overclocking to 2150 on memory.
    The Hynix memory you speak of (H5GQ8H24MJR) is actually - surprisingly - worse than the Hynix rated for 1750Mhz actual (H5GC4H24AJR for the 4GB) - even though it's rated for 2000Mhz actual!

    However, that doesn't mean it's hopeless... it's tempermental and very touchy, but if you treat it right, it does do over 30MH/s: https://ottrbutt.com/tmp/ethwolf-06032017.jpg -- that's of Naomi with my Sapphire Pulse 580 (Hynix 8GB) in.
  • jetpackjetpack Member Posts: 88
    Using the hynix 580 and at 2100mhz doing 32.3 bump to 2200mhz im at 33.5 so saying samsung memory is better is not true. 120watt 74C btw
  • PierrePierre Member Posts: 2
    Not possible on Gigabyte with Hynix mem. only getting 26.5 at 2100 before errors starts.
  • Wolf0Wolf0 Member Posts: 329 ✭✭✭
    jetpack said:

    Using the hynix 580 and at 2100mhz doing 32.3 bump to 2200mhz im at 33.5 so saying samsung memory is better is not true. 120watt 74C btw

    If you're saying you get 33.5MH/s out of H5GQ8H24MJR (8GB Hynix rated for 2000Mhz actual) I'm gonna have to call bullshit.
  • zerf2004zerf2004 Member Posts: 1
    inmortan said:


    I got 29.9 with Gigabyte Radeon RX 580 Gaming 4G 4GB GDDR coping 1500 to 1750 and 2000 in bios.

    What type of memory? I have hynix memory and I am struggling to get past 27.
  • jetpackjetpack Member Posts: 88
    Wolf0 said:

    jetpack said:

    Using the hynix 580 and at 2100mhz doing 32.3 bump to 2200mhz im at 33.5 so saying samsung memory is better is not true. 120watt 74C btw

    If you're saying you get 33.5MH/s out of H5GQ8H24MJR (8GB Hynix rated for 2000Mhz actual) I'm gonna have to call bullshit.
    Sure call bs, doesnt bother me. If you get to 37 lmk what you did and we can combine our potentiall different tweaks to hit 45. I took the pic, what should I do to prove it haha. Heres another one. and more info. It could be a flaw I guess? hynix 580, c
    ustom bios, win7 ultimate, FX cpu with mods.. actually doing 37mhs on it now. at 2050mhz. going for 40mhs unless you had the same computer all the settings do not apply each app is different so many variables. How can you confirm a mhs like if im doing 81 shares a hour at 71mhs on two cards, claymore. so 1.14 shares per min is a good rate? 113watt per card. Thats all i got :)
  • jetpackjetpack Member Posts: 88
    edited June 2017
    What kind of red flags do we have to look out for when chasing for 50mhs on a 580? Is temp the only thing that can damage the cards if you keep volts below standard settings?

    Also is there a chance it can produce garbage hashes? Even with this pic I was only able to grab 69 shares in 80min. At 55(98%)/1(2%)/00 on ethermine tho, just kinda testing these settings. 72 Current 67 Reported 53 Average right now.

    So with the info I have told you is my mhs 36 or are my numbers looking like 31mhs?
  • grchinagrchina Member Posts: 7
    jetpack said:

    What kind of red flags do we have to look out for when chasing for 50mhs on a 580? Is temp the only thing that can damage the cards if you keep volts below standard settings?

    Also is there a chance it can produce garbage hashes? Even with this pic I was only able to grab 69 shares in 80min. At 55(98%)/1(2%)/00 on ethermine tho, just kinda testing these settings. 72 Current 67 Reported 53 Average right now.

    So with the info I have told you is my mhs 36 or are my numbers looking like 31mhs?

    Try installing hw info and look for gpu memory errors,as long as it shows small number of errors you are good
  • PierrePierre Member Posts: 2
    So wolf0 what did you change on the Gigabyte rx 580 gaming 8g bios.
  • Wolf0Wolf0 Member Posts: 329 ✭✭✭
    Pierre said:

    So wolf0 what did you change on the Gigabyte rx 580 gaming 8g bios.

    I own a Gigabyte RX 580 G1 Gaming 8G, but haven't actually played with it much yet... what made you ask about Gigabyte? None of the GPUs in my screenshot were Gigabyte...
  • andreas911andreas911 Member Posts: 23
    edited June 2017


    So I have constant 28 Mh/s with Powercolor RX580 Red Devil Golden Sample

    But I don't have a Core or Memory Clock and no temperature
  • MEng95MEng95 Member Posts: 28
    edited June 2017
    grchina said:

    Try installing hw info and look for gpu memory errors,as long as it shows small number of errors you are good

    What number is a small number of errors? I never realised having a high hash rate wasn't the total end game. So 30MH/s could potentially not be as effective as 27MH/s for example?

  • grchinagrchina Member Posts: 7
    MEng95 said:

    grchina said:

    Try installing hw info and look for gpu memory errors,as long as it shows small number of errors you are good

    What number is a small number of errors? I never realised having a high hash rate wasn't the total end game. So 30MH/s could potentially not be as effective as 27MH/s for example?

    i try to stay in 2 digits or less but I have seen people getting millions
  • lablettlablett Member Posts: 333 ✭✭
    jetpack said:

    Wolf0 said:

    jetpack said:

    Using the hynix 580 and at 2100mhz doing 32.3 bump to 2200mhz im at 33.5 so saying samsung memory is better is not true. 120watt 74C btw

    If you're saying you get 33.5MH/s out of H5GQ8H24MJR (8GB Hynix rated for 2000Mhz actual) I'm gonna have to call bullshit.
    Sure call bs, doesnt bother me. If you get to 37 lmk what you did and we can combine our potentiall different tweaks to hit 45. I took the pic, what should I do to prove it haha. Heres another one. and more info. It could be a flaw I guess? hynix 580, c
    ustom bios, win7 ultimate, FX cpu with mods.. actually doing 37mhs on it now. at 2050mhz. going for 40mhs unless you had the same computer all the settings do not apply each app is different so many variables. How can you confirm a mhs like if im doing 81 shares a hour at 71mhs on two cards, claymore. so 1.14 shares per min is a good rate? 113watt per card. Thats all i got :)
    What strategy did you use to optmise the straps ?
  • HendrickOrtizHendrickOrtiz Member Posts: 5
    edited June 2017
    jetpack said:

    Wolf0 said:

    jetpack said:

    Using the hynix 580 and at 2100mhz doing 32.3 bump to 2200mhz im at 33.5 so saying samsung memory is better is not true. 120watt 74C btw

    If you're saying you get 33.5MH/s out of H5GQ8H24MJR (8GB Hynix rated for 2000Mhz actual) I'm gonna have to call bullshit.
    Sure call bs, doesnt bother me. If you get to 37 lmk what you did and we can combine our potentiall different tweaks to hit 45. I took the pic, what should I do to prove it haha. Heres another one. and more info. It could be a flaw I guess? hynix 580, c
    ustom bios, win7 ultimate, FX cpu with mods.. actually doing 37mhs on it now. at 2050mhz. going for 40mhs unless you had the same computer all the settings do not apply each app is different so many variables. How can you confirm a mhs like if im doing 81 shares a hour at 71mhs on two cards, claymore. so 1.14 shares per min is a good rate? 113watt per card. Thats all i got :)
    Any memory errors on those card?
  • xxharryxxharry Member Posts: 25
    edited June 2017
    Hey guys, I have just picked up a msi rx580 8gb samsung memory and tried to mod the bios. I copied the 1500 settings to 1625,1750 and 2000 like in this video When I load the ROM, the computer boots up fine and shows video but after running etherminer for few seconds, it crashed. for those who are getting 30MH/s, can you please share your BIOS. thank you so much

    edit: my system consumed 220watt with rx480 in it and 28MH/sec but with the rx480 plus rx580 its at 400 watt. are those figures right? or there is something wrong with my 580?
    Post edited by xxharry on
  • botarnoldbotarnold Member Posts: 16
    xxharry said:

    Hey guys, I have just picked up a msi rx580 8gb samsung memory and tried to mod the bios. I copied the 1500 settings to 1625,1750 and 2000 like in this video When I load the ROM, the computer boots up fine and shows video but after running etherminer for few seconds, it crashed. for those who are getting 30MH/s, can you please share your BIOS. thank you so much

    edit: my system consumed 220watt with rx480 in it and 28MH/sec but with the rx480 plus rx580 its at 400 watt. are those figures right? or there is something wrong with my 580?

    Its just as simple as copying 1750 strap to 2000. leave the 1500 as it is.
  • xxharryxxharry Member Posts: 25
    done that bro but I struggle to get it past 28MH/s. I dont mind it but the power draw by the rx580 is 140 watt. if i try to put the memory clock past 2150 in wattman or try to undervolt even a little, I get a crash.
  • Mike_Nguyen53Mike_Nguyen53 Member Posts: 12



    Thanks everyone for this threat. I have learnt a lot from you guys to build my very first rig. I got 29.3Mh for Ether and 528Mh for Siacoin (dual mining). Below are the details.

    6 * Rx 580 Gaming 4G Gigabyte, memory Hynix (H5GC4H24AJR).
    Mod Bios: Strap 1500 to 1625 1750 2000.
    Clocks: 1100/2050
    Voltage: -cvddc 850 -mvddc 850
    GPU-z: 77.6w (single card)
    From wall: 922w

    The rig has been running 12 hours stable till now. Any comments or questions are more than welcome :)

    Cheers!
  • eltimmoeltimmo Member Posts: 4
    eltimmo said:

    Has any one used this configuration it's for Sapphire RX580 8GB hynix

    http://1stminingrig.com/best-bios-rom-sapphire-nitro-rx-580-8gb-oc-hynix-memory-30-mhs/#disqus_thread

    I've tried this and it's working well, all cards hashing at pretty much bang on 30 Mh/s and drawing around 75w in GPU-z.
  • xxharryxxharry Member Posts: 25




    Thanks everyone for this threat. I have learnt a lot from you guys to build my very first rig. I got 29.3Mh for Ether and 528Mh for Siacoin (dual mining). Below are the details.

    6 * Rx 580 Gaming 4G Gigabyte, memory Hynix (H5GC4H24AJR).
    Mod Bios: Strap 1500 to 1625 1750 2000.
    Clocks: 1100/2050
    Voltage: -cvddc 850 -mvddc 850
    GPU-z: 77.6w (single card)
    From wall: 922w

    The rig has been running 12 hours stable till now. Any comments or questions are more than welcome :)

    Cheers!

    I try and copy the Strap 1500 to 1625 1750 2000 on my msi rx580 8gb. is there any other setting you changed in the BIOS apart from that?
  • AjninDoCAjninDoC Member Posts: 2
    I have a RX580 Armor 4gb OC hynix mem. I use the 1500 straps up to 2000 for both 1 and 2 straps. core is 1120 mem is 1950, tdp is 100. Powertune is 3 in linux claymore. It's good for 7 days 100% stable, 514 watts for 6 gpus @ 28-28.5 mh. not the promised land 30mh. But it's stable. Hope this helps someone, because it was hell for me finding this out on my own.
  • JoeZenerJoeZener Member Posts: 13
    Wolf0 said:

    Same memory manufacturer for your 4gb and 8gb card? Check both with gpuZ and show us your results.

    Found out that there are 2 versions from gigabyte. One with samsung and one with hynix memory and i got the hynix memory. Copied from 1750 to 2000 and 2250 but i think 27mh/s is the end while overclocking to 2150 on memory.
    The Hynix memory you speak of (H5GQ8H24MJR) is actually - surprisingly - worse than the Hynix rated for 1750Mhz actual (H5GC4H24AJR for the 4GB) - even though it's rated for 2000Mhz actual!

    However, that doesn't mean it's hopeless... it's tempermental and very touchy, but if you treat it right, it does do over 30MH/s: https://ottrbutt.com/tmp/ethwolf-06032017.jpg -- that's of Naomi with my Sapphire Pulse 580 (Hynix 8GB) in.
    Wolf.... the other GPUs in Naomi are Samsung... would you share what are your mods?
  • rhombusrhombus Member Posts: 13
    AjninDoC said:

    514 watts for 6 gpus @ 28-28.5 mh.

    514 at the wall? if that's the case what's your secret?

  • kriGERkriGER Member Posts: 3
    Strange figures from AjninDoC. I got rx 580 pulse 4G hynix, moded, and 30.5 Mh /s per card. For 5 cards 850 from wall. I know i can get down maybe even for 200W with right setup but 514W with 6 cards no way.
    Share us your setup AjninDoC pls
  • ecs87ecs87 Dekalb, ILMember Posts: 339 ✭✭✭
    He's obviously not measuring at the wall. At the wall is probably double that so ~1000 watts.
Sign In or Register to comment.