|
PovertyByte (OP)
|
 |
February 07, 2017, 02:03:00 AM |
|
This is hopefully the thread to end all of my RX series questions post and responding in other threads
How efficient are you guys able to run these cards on what coins, and is this a wall or software sensor measurement? And of course what variation of the 470 and 480 do you have and with which memory if you know it?
|
|
|
|
|
|
|
|
If you see garbage posts (off-topic, trolling, spam, no point, etc.), use the "report to moderator" links. All reports are investigated, though you will rarely be contacted about your reports.
|
|
|
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction.
|
|
|
|
|
|
|
|
|
|
PovertyByte (OP)
|
|
February 07, 2017, 03:13:36 AM |
|
Really nice on the details, very much appreciated. What's the hash rate difference on the two different sets when mining XMR? I'm going to assume 29-31 MH/s on ETH |
|
|
|
|
|
Cubirez
|
|
February 07, 2017, 07:36:34 AM |
|
Really nice on the details, very much appreciated. What's the hash rate difference on the two different sets when mining XMR? I'm going to assume 29-31 MH/s on ETH Got 6x470 (all sapphire Nitro 4GB, different mem chips though : 3 samsung, 2 elpida, 1 hynix) mining xmr @ ~ 4400 H/s. About 640W at the wall for the whole system so that's around 100W/card. No big difference between them in terms of hashing except for the hynix which is a bit lower (but it's the gpu on display). Best choice regarding H/W and H/$ would be currently RX470 4G. However if you can find a 480 4GB at the same price, go for the 480 |
|
|
|
|
|
Walrusbonzo
|
|
February 07, 2017, 07:59:19 AM
Last edit: February 07, 2017, 08:13:28 AM by Walrusbonzo |
|
I've got 7x RX480 Sapphire Nitro+ pulling 204mh/s mining Ethereum for 930w at the wall.
And a another 5x RX480 Sapphire Nitro+ with 1x Ref RX480 pulling 179mh/s for 810w at the wall.
All cards have Samsung RAM.
Serious GPU undervolting is needed for this, I.E. around 0.89v. Also, using Windows 10, as mining is consistently slower with Linux.
|
|
|
|
|
|
Pennywis3
|
|
February 07, 2017, 08:15:13 AM |
|
Got 6x470 (all sapphire Nitro 4GB, different mem chips though : 3 samsung, 2 elpida, 1 hynix) mining xmr @ ~ 4400 H/s. About 640W at the wall for the whole system so that's around 100W/card. No big difference between them in terms of hashing except for the hynix which is a bit lower (but it's the gpu on display).
Best choice regarding H/W and H/$ would be currently RX470 4G. However if you can find a 480 4GB at the same price, go for the 480
Those are some impresive results dude, only 640W for 4400H/s is just wow O.o My server CPU's burn more wattage hehe. |
|
|
|
|
|
un4given
|
|
February 07, 2017, 08:24:11 AM |
|
Really nice on the details, very much appreciated. What's the hash rate difference on the two different sets when mining XMR? I'm going to assume 29-31 MH/s on ETH Got 6x470 (all sapphire Nitro 4GB, different mem chips though : 3 samsung, 2 elpida, 1 hynix) mining xmr @ ~ 4400 H/s. About 640W at the wall for the whole system so that's around 100W/card. No big difference between them in terms of hashing except for the hynix which is a bit lower (but it's the gpu on display). Best choice regarding H/W and H/$ would be currently RX470 4G. However if you can find a 480 4GB at the same price, go for the 480 what core/mem settings r u using for the RX 470 rig? 1150/2000? 640W ist very low for the whole rig. the measurement is from the wall? |
|
|
|
|
|
Cubirez
|
|
February 07, 2017, 08:33:10 AM |
|
Got 6x470 (all sapphire Nitro 4GB, different mem chips though : 3 samsung, 2 elpida, 1 hynix) mining xmr @ ~ 4400 H/s. About 640W at the wall for the whole system so that's around 100W/card. No big difference between them in terms of hashing except for the hynix which is a bit lower (but it's the gpu on display).
Best choice regarding H/W and H/$ would be currently RX470 4G. However if you can find a 480 4GB at the same price, go for the 480
Those are some impresive results dude, only 640W for 4400H/s is just wow O.o My server CPU's burn more wattage hehe. The guys with mem tuning skills like Wolf0 or Eliovp will manage to get at the very least 10% more with the same wattage using their custom bioses. As for me I only copied the mem straps (the 1750Mhz for samsung, the 1500 MHz for Elpida and Hynix), overclocked the mem, and slightly underclocked and undervolted core. The big difference, when using Claymore comes with the -a 2 mode (I'm not sure about the numbers but it's something like -15% power usage with -3% hashrate). I haven't played enough with Wolf0's miner and sgminer GM but i might try them this week-end when i have some time. |
|
|
|
|
|
Cubirez
|
|
February 07, 2017, 08:38:40 AM |
|
Really nice on the details, very much appreciated. What's the hash rate difference on the two different sets when mining XMR? I'm going to assume 29-31 MH/s on ETH Got 6x470 (all sapphire Nitro 4GB, different mem chips though : 3 samsung, 2 elpida, 1 hynix) mining xmr @ ~ 4400 H/s. About 640W at the wall for the whole system so that's around 100W/card. No big difference between them in terms of hashing except for the hynix which is a bit lower (but it's the gpu on display). Best choice regarding H/W and H/$ would be currently RX470 4G. However if you can find a 480 4GB at the same price, go for the 480 what core/mem settings r u using for the RX 470 rig? 1150/2000? 640W ist very low for the whole rig. the measurement is from the wall? Mesurement is from the wall. Got a 1000W gold PSU@220V so that should be around 90% efficiency. For now core is at 1200 and mem between 1950 and 2100 depending on the GPU. The undervolt is light (between 0,95V and 0,975V). Will try to decrease core & undervolt to see if it's worth it. But since my rig is stable i got a bit lazy <_< The big difference in Power draw came from using -a 2 setting in Claymore. |
|
|
|
|
TheKoolaider
Member
 Offline
Activity: 84
Merit: 10
|
|
February 07, 2017, 01:23:14 PM |
|
Really nice on the details, very much appreciated. What's the hash rate difference on the two different sets when mining XMR? I'm going to assume 29-31 MH/s on ETH Unfortunately not. I haven't done any modded bios. So for ETH, both rigs hash at 25Mh/s not overclocked. The MSI rig, if I OC the Memory to 2200, the cards hash rate increase to 27-28. For XMR, both rigs hash at ~640 per card. I have not successfully overclocked XMR ever, it's so hard for me. I see people doing it and I honestly halfway don't believe them. I just need to do more research. |
|
|
|
|
|
IOTUSA
|
|
February 07, 2017, 03:58:08 PM |
|
Does anyone have a modded Bios that works for Rx 480 4/8GB and is stable for Zcash/Eth/XMR? I would prefer being able to algo-switch without re-flashing?
|
|
|
|
|
LowByter
|
|
February 07, 2017, 04:47:55 PM |
|
Just an info on old 380 series...after one slot died (and the card) 9 cards on two rigs dig 0,8 Eth per day and when electricity bill is payed you're left with 80 euros per month... lowsy investment  |
|
|
|
|
Truthchanter
|
|
February 08, 2017, 05:01:03 PM |
|
In the latest claymore ethereum miner I don't see an option for -a 2 or any -a at all, although i see a low intensity more -li (could use -li 2) is that what you mean by -a?
|
|
|
|
|
|
Cubirez
|
|
February 09, 2017, 03:59:19 AM |
|
In the latest claymore ethereum miner I don't see an option for -a 2 or any -a at all, although i see a low intensity more -li (could use -li 2) is that what you mean by -a?
I'm not using the eth miner so i don't know about -li, but it might be similar, you should test. -a is an option for Claymore's XMR miner :
-a Select algorithm:
"-a 0" (default) means autoselection.
"-a 1" fastest mining mode for 290/290x and 270/270x cards.
"-a 2" low power consumption mode for 290/290x cards. A bit less speed (1%) but less power consumption (10-15%).
"-a 3" fastest mining for old cards (6xxx and 5xxx).
"-a 4" fastest mining for recent cards (280x, 270/270x etc).
You can also specify mode for every card separately, for example, "-a 1234".
It is very important for good mining speed to select correct "-a" and "-h" values for your cards.
For old cards 5xxx and 6xxx use "-a 3" mode only, otherwise you can get "LowDifficultyShare" errors from pool. |
|
|
|
|
|
