Still an ongoing struggle, rig has been down more than its up, I am considering breaking the rig up and resurrecting the FX990-GAMING motherboard, I still have a 450W psu from that, all I need is a 4GB stick of DDR3 RAM, but I would need to buy another windows license or run linux on it ?
So I'll try to run 4 GPU's on the MSI motherboard and 1200 W Roseville PSU with Windows 10, and 3 GPU's on the Gigabyte motherboard with 450W PSU with Ubuntu.
There's no information on equihash (the algorithm behind ZEC) and very little cryptonite (the alg used by monero).
http://whattomine.com lists the most common GPU's - you can enter 1 next to each card and then click the card to turn it on/off and click calculate to see the default hashrates.
Of the cards listed, the Nvidia GTX 1070 has the highest equihash rate, 410 h/s
The Nvidia cards don't do as well with ethereum and especially dual mining if you wanted to use Claymore's dual miner.
Also important is the energy consumption of the cards, as far as I know the AMD RX 470 has the highest hashrate for the lowest energy use, in other words you'll keep more of your mined profit and spend less on electricity.
There's no information on equihash (the algorithm behind ZEC) and very little cryptonite (the alg used by monero).
http://whattomine.com lists the most common GPU's - you can enter 1 next to each card and then click the card to turn it on/off and click calculate to see the default hashrates.
Of the cards listed, the Nvidia GTX 1070 has the highest equihash rate, 410 h/s
The Nvidia cards don't do as well with ethereum and especially dual mining if you wanted to use Claymore's dual miner.
Also important is the energy consumption of the cards, as far as I know the AMD RX 470 has the highest hashrate for the lowest energy use, in other words you'll keep more of your mined profit and spend less on electricity.
I always use whattomine.com for calculations (coins/algorithms). Fortunately I don't have to worry about power comsumption because in my country the energy is almost free (of course, you have constant energy outages and problems...epicfail)
The rig worked for approx 1.5 days mining at 162.5 mh/s, then last night abruptly two cards stopped with the opencl error and this morning I struggled for several hours to get it working again.
I have a new theory, I think some components, be it the GPU's the capitors in the risers, whatever overheats or something and when the rig is shutdown for a while and restarted, it just works again ?
My replacement GPU arrived and while the rig was down this morning, I added the 7th GPU.
I first unplugged all the others and put the new GPU in slot 0, flashed the bios, replugged all the GPU's in order and then went thru trixx and made sure all the cards were at 1100/1950, some at 1925, GPU at 1875, then restarted and it magically worked, so the rig is now hashing at 190 mh/s
Anybody getting a new batch of RX 470 4GB with samsung chip?
I have 30 PowerColor Red Dragon RX470 4GB and all of them are Samsung chip
as usual, im using 1500 timing > 2000
my setting now is
1130Core @ 930mV 2100Mem @ 930mv
Getting 29.2Mh/s per card
Mind to share yours ?
I have 23 , 24 MH/s on my powercolor red dragon rx 470 with 4GB samsung memory with modified bios 1500 timing to all higher. what i do wrong, how can i make 28 or 27 mh/s with this card?
Anybody getting a new batch of RX 470 4GB with samsung chip?
I have 30 PowerColor Red Dragon RX470 4GB and all of them are Samsung chip
as usual, im using 1500 timing > 2000
my setting now is
1130Core @ 930mV 2100Mem @ 930mv
Getting 29.2Mh/s per card
Mind to share yours ?
I have 23 , 24 MH/s on my powercolor red dragon rx 470 with 4GB samsung memory with modified bios 1500 timing to all higher. what i do wrong, how can i make 28 or 27 mh/s with this card?
Anybody getting a new batch of RX 470 4GB with samsung chip?
I have 30 PowerColor Red Dragon RX470 4GB and all of them are Samsung chip
as usual, im using 1500 timing > 2000
my setting now is
1130Core @ 930mV 2100Mem @ 930mv
Getting 29.2Mh/s per card
Mind to share yours ?
I have 23 , 24 MH/s on my powercolor red dragon rx 470 with 4GB samsung memory with modified bios 1500 timing to all higher. what i do wrong, how can i make 28 or 27 mh/s with this card?
Some cards need more power. If you use watttool bump the power to 10 or 15 and see if your mh/s improves.
My replacement GPU arrived and while the rig was down this morning, I added the 7th GPU.
I first unplugged all the others and put the new GPU in slot 0, flashed the bios, replugged all the GPU's in order and then went thru trixx and made sure all the cards were at 1100/1950, some at 1925, GPU at 1875, then restarted and it magically worked, so the rig is now hashing at 190 mh/s
Lets see how long it stays up this time.
which motherboard for 7 GPU's? You connected them with 2 PSU-s? Which type/model?
I recently got a pair of Sapphire RX 470 Nitro (Samsung memory), I reflashed the bios memory time straps from 1750 to 2000 for memory and adjusted speeds as 1074core/2000mem.
In EthOS (Debian or Ubuntu fork) the AMDGPU driver and 400 series as regards to the power and voltage is a bit different then usual. I tried adjusting the voltage but the best I've been able to accomplish is setting the DPM (Dynamic Power Management) to 2 as an attempt to undervolt which doesn't do much and setting vlt to .93 doesn't do much either. The card runs as ~120watts ETH and dual mining jacks the wattage up to ~200watts eth/dcr.
I've tested other combinations of time straps and core/mem with mixed results.
these are my 3 Sapphire rx 470 oc+ 4gb. I flashed them with same bios. shifting dram timing from 1500 to 1625 and 2000. i can't get a decent power consumption and stability. by the way i reached 29 mhs setting ram to 2070.
Buying a Gigabyte Radeon RX 470 4GB G1 Gaming tomorrow for my new rig. Will let you guys know of the results! Not much information here about those cards.
My replacement GPU arrived and while the rig was down this morning, I added the 7th GPU.
I first unplugged all the others and put the new GPU in slot 0, flashed the bios, replugged all the GPU's in order and then went thru trixx and made sure all the cards were at 1100/1950, some at 1925, GPU at 1875, then restarted and it magically worked, so the rig is now hashing at 190 mh/s
Lets see how long it stays up this time.
which motherboard for 7 GPU's? You connected them with 2 PSU-s? Which type/model?
MANNNNNNNN CRANK THOSE FANS UP!!!!! Fans are easily replaceable, GPU's arent
When running claymore, it gives about 26,7MH/s. I'm okay with it. It already needs +2 voltage offset in Watt Tool to give this MH/s stable. My problem is that after an hour I still have 0 shares and about 4-5 'incorrect shares'. Also, when moving my mouse cursor it is really slow. And when using some programs (firefox) it gives artifacts on screen. If I give more voltage or change core / memory my MH/s tanks down to 10 or I get a unresponsive computer.
Edit: temps are 62-65 on core (with 35% fam speed) and 78-80 on vrm.
When running claymore, it gives about 26,7MH/s. I'm okay with it. It already needs +2 voltage offset in Watt Tool to give this MH/s stable. My problem is that after an hour I still have 0 shares and about 4-5 'incorrect shares'. Also, when moving my mouse cursor it is really slow. And when using some programs (firefox) it gives artifacts on screen. If I give more voltage or change core / memory my MH/s tanks down to 10 or I get a unresponsive computer.
Edit: temps are 62-65 on core (with 35% fam speed) and 78-80 on vrm.
What can i try next?
Have you tried 1750 straps ? 1130/2100 , 16.9.1, -120mV, 29.xx Mh/s.. at least this is what i get
When running claymore, it gives about 26,7MH/s. I'm okay with it. It already needs +2 voltage offset in Watt Tool to give this MH/s stable. My problem is that after an hour I still have 0 shares and about 4-5 'incorrect shares'. Also, when moving my mouse cursor it is really slow. And when using some programs (firefox) it gives artifacts on screen. If I give more voltage or change core / memory my MH/s tanks down to 10 or I get a unresponsive computer.
Edit: temps are 62-65 on core (with 35% fam speed) and 78-80 on vrm.
What can i try next?
I recommend 1750 strap for samsung mem. are you sure your mining to a valid address? what is your claymore .bat file look like
Thanks for the replies. I now modded my own vbios instead of taking one from forums. I took the original one from my card and copied the 1750 straps to 2000 as well. My .bat file is:
Of course I change my walletadress/name/email in the real .bat file. I do get shares now btw!!! After modding my own bios i also installed 16.11.2. Claymore still running 26.4MH/s-ish.. But when Claymore is running the pc can hardly be used anymore. So many stutters! Guess it's less stable now?
I am considering adding another RX470 to my mining rig. I am not sure which is better, XFX RX470 Triple X or PowerColor RX470 Red Devil? Any advice? My main concern is to optimise the card for power efficiency since I am paying slightly higher than $0.15 per kWH for electricity.
@charliebridge82 What GPUs are currently the best mining ZEC and Monero?
Honestly I don't know, because I'm only mining ETH right now, but I heard you can mine ZEC at good hashrates with "old" AMD cards (R7 and R9 series)
I am getting 300 sol/s for my RX 480 8GB Red Devil and 240 sol/s on my HD7950 3GB Vapor-X. It is true that the older AMD cards are performing well enough for ZEC.
I recently got a pair of Sapphire RX 470 Nitro (Samsung memory), I reflashed the bios memory time straps from 1750 to 2000 for memory and adjusted speeds as 1074core/2000mem.
In EthOS (Debian or Ubuntu fork) the AMDGPU driver and 400 series as regards to the power and voltage is a bit different then usual. I tried adjusting the voltage but the best I've been able to accomplish is setting the DPM (Dynamic Power Management) to 2 as an attempt to undervolt which doesn't do much and setting vlt to .93 doesn't do much either. The card runs as ~120watts ETH and dual mining jacks the wattage up to ~200watts eth/dcr.
I've tested other combinations of time straps and core/mem with mixed results.
Anyone have a way lowering the wattage?
Adding a global offset to Loop 1 of the voltage controller will do it on Linux - some Sapphire cards have the command already, in others you have to add it, which is a major pain.
Comments
So I'll try to run 4 GPU's on the MSI motherboard and 1200 W Roseville PSU with Windows 10, and 3 GPU's on the Gigabyte motherboard with 450W PSU with Ubuntu.
Thoughts ?
There's no information on equihash (the algorithm behind ZEC) and very little cryptonite (the alg used by monero).
http://whattomine.com lists the most common GPU's - you can enter 1 next to each card and then click the card to turn it on/off and click calculate to see the default hashrates.
Of the cards listed, the Nvidia GTX 1070 has the highest equihash rate, 410 h/s
The Nvidia cards don't do as well with ethereum and especially dual mining if you wanted to use Claymore's dual miner.
Also important is the energy consumption of the cards, as far as I know the AMD RX 470 has the highest hashrate for the lowest energy use, in other words you'll keep more of your mined profit and spend less on electricity.
I have a new theory, I think some components, be it the GPU's the capitors in the risers, whatever overheats or something and when the rig is shutdown for a while and restarted, it just works again ?
I first unplugged all the others and put the new GPU in slot 0, flashed the bios, replugged all the GPU's in order and then went thru trixx and made sure all the cards were at 1100/1950, some at 1925, GPU at 1875, then restarted and it magically worked, so the rig is now hashing at 190 mh/s
Lets see how long it stays up this time.
what i do wrong, how can i make 28 or 27 mh/s with this card?
EDIT: I should mention that's on linux - should be higher on Win.
which motherboard for 7 GPU's? You connected them with 2 PSU-s? Which type/model?
In EthOS (Debian or Ubuntu fork) the AMDGPU driver and 400 series as regards to the power and voltage is a bit different then usual. I tried adjusting the voltage but the best I've been able to accomplish is setting the DPM (Dynamic Power Management) to 2 as an attempt to undervolt which doesn't do much and setting vlt to .93 doesn't do much either. The card runs as ~120watts ETH and dual mining jacks the wattage up to ~200watts eth/dcr.
I've tested other combinations of time straps and core/mem with mixed results.
Anyone have a way lowering the wattage?
these are my 3 Sapphire rx 470 oc+ 4gb. I flashed them with same bios. shifting dram timing from 1500 to 1625 and 2000.
i can't get a decent power consumption and stability. by the way i reached 29 mhs setting ram to 2070.
MANNNNNNNN CRANK THOSE FANS UP!!!!! Fans are easily replaceable, GPU's arent
MSI gaming x RX 470 4g (SAMSUNG memory, 64,6% ASIC qual)
BIOS: 1500 straps copied > 1500
Driver: 16.6
Core Clock: 1130mhz
Memory Clock: 1900mhz
Miner: claymore 9.0
Pool: Nanopool
Tweaking: MSI Afterburner & Watt Tool
When running claymore, it gives about 26,7MH/s. I'm okay with it. It already needs +2 voltage offset in Watt Tool to give this MH/s stable. My problem is that after an hour I still have 0 shares and about 4-5 'incorrect shares'. Also, when moving my mouse cursor it is really slow. And when using some programs (firefox) it gives artifacts on screen. If I give more voltage or change core / memory my MH/s tanks down to 10 or I get a unresponsive computer.
Edit: temps are 62-65 on core (with 35% fam speed) and 78-80 on vrm.
What can i try next?
setx GPU_FORCE_64BIT_PTR 0
setx GPU_MAX_HEAP_SIZE 100
setx GPU_USE_SYNC_OBJECTS 1
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_SINGLE_ALLOC_PERCENT 100
EthDcrMiner64.exe -epool eth-eu1.nanopool.org:9999 -ewal walletadress/name/email
Of course I change my walletadress/name/email in the real .bat file. I do get shares now btw!!! After modding my own bios i also installed 16.11.2. Claymore still running 26.4MH/s-ish.. But when Claymore is running the pc can hardly be used anymore. So many stutters! Guess it's less stable now?
Edit: monitor screenshot: