Author

Topic: PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux) - page 219. (Read 784965 times)

full member
Activity: 357
Merit: 101
IMPORTANT! All owners of AMD cards with 4 GB RAM must upgrade to PhoenixMiner 5.2b or later version to continue mining after DAG epoch 373 (set -daglim 1, or see bellow how to use the -daglim option; you also need to add -dagrestart 1 to allow maxium possible value of -daglim). Additionally, here are some important tips for longest possible usage of 4 GB AMD cards with PhoenixMiner 5.2a and later:

  • After the epoch 373 the hashrate will drop dramatically (to just a few MH/s) because the DAG no longer fits in the card's VRAM and part of it is stored in the system memory. Use the option -daglim 1 to use only 4006 MB for the DAG buffer. This will allow you to mine for several more weeks with gradually lower hashrate for each passing DAG epoch. If the speed is too low (less than 10 MH/s for epoch 374), you need to use -daglim 4000, or even lower value - keep trying with lower and lower -daglim values until the speed jumps higher. Note that this value can be different for each card but as a general rule of thumb, if you are able to mine without problems and speed loss on epoch 373, -daglim 1 should work fine. To allow higher -daglim values, you also need to add -dagrestart 1
  • Note that -daglim will only work on Polaris GPUs (RX470, RX480, RX570, RX580, RX590), and the PhoenixMiner 5.2b is still in alpha stage of testing
  • It is also important to use auto-tune (do not specify -gt values) because the optimal -gt value may change with each new epoch, and will definitely be different than before.
  • You can also try the new option -rxboost 1 for additional increase of hashrate

IMPORTANT! All owners of AMD cards with 6 GB or 8 GB RAM must either keep drivers 20.4.x or lower (do not upgrade to 20.5.1 or later), or upgrade to PhoenixMiner 5.2b or later version to continue mining after DAG epoch 384 (ETC has already passed DAG epoch 384, and ETH will pass it in less than two months).
newbie
Activity: 84
Merit: 0
guys.

all 4gb cards can't mining anymore

Mine are all running just fine...on eth...on linux...30 some odd 4gb 570s
Mine rigs are on 5.0e and one is 5.1c.

So appears soon i will need to update the miner. 

Linux ETH can mine till when ?
full member
Activity: 1275
Merit: 141
guys.

all 4gb cards can't mining anymore

Mine are all running just fine...on eth...on linux...30 some odd 4gb 570s
Mine rigs are on 5.0e and one is 5.1c.

So appears soon i will need to update the miner. 
newbie
Activity: 84
Merit: 0
guys.

all 4gb cards can't mining anymore
newbie
Activity: 3
Merit: 0
newbie
Activity: 3
Merit: 0
3 cards
5700XT, 5700, 480 4GB

PhoenixMiner 5.2a ~19 Mhs with -daglim 3967, any greater daglim produces a hashrate drop to 0-1 Mhs

If I use PhoenixMiner for the 5700 and Lolminer with 480 it makes 30,31Mhs

Driver 20.8.1, Windows.

Any suggestion?

Same with RX 580 4 GB! Works only -daglim 3967 ~19 Mhs. Driver 19.12.3.
sr. member
Activity: 1484
Merit: 253
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

This test was 2 part... Mainly was about how the console hash comes is pretty much always going to be 3.5% higher than at the pool.... There will be days where you actually get it (as you can see in my test, but odds are going to be very slim) Also, TRM produced a better over hashrate, over the course of the year, that .75-1% will add up quickly.

Power on both miners were so close that it was marginal.
I didn't know about your system, but for me PM uses about 1-2% less power only from sensors (overall power economy from wall can be more). So 0.75-1% less efficiency but 1-2% less power.
Until lol and TR didn't lower power draw by their miners for me choise is obvious.

Software power draw means 0.... I have a 5700xt that's showing software power draw at 77w and 80w!!!!!!

By the way; the software power draw for PM on these cards (with these settings) showed 84w.
I didn't know about 5700xt. I use rx 580 8Gb rigs. On Polarises PM use less power.
jr. member
Activity: 195
Merit: 4
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

This test was 2 part... Mainly was about how the console hash comes is pretty much always going to be 3.5% higher than at the pool.... There will be days where you actually get it (as you can see in my test, but odds are going to be very slim) Also, TRM produced a better over hashrate, over the course of the year, that .75-1% will add up quickly.

Power on both miners were so close that it was marginal.
I didn't know about your system, but for me PM uses about 1-2% less power only from sensors (overall power economy from wall can be more). So 0.75-1% less efficiency but 1-2% less power.
Until lol and TR didn't lower power draw by their miners for me choise is obvious.

Software power draw means 0.... I have a 5700xt that's showing software power draw at 77w and 80w!!!!!!

By the way; the software power draw for PM on these cards (with these settings) showed 84w.
jr. member
Activity: 200
Merit: 3
-cdmport is not working in the new version...

It is working fine here (read-only access).
Huh

CMD-Line used:
PhoenixMiner.exe -amd -eres 1 -mi 12 -pool DAGGERHASHIMOTO.eu.nicehash.com:3353 -wal ************************.********** -pass x -proto 4 -log 0 -wdog 0 -cdmport 4001 -gpus 2
jr. member
Activity: 76
Merit: 1
3 cards
5700XT, 5700, 480 4GB

PhoenixMiner 5.2a ~19 Mhs with -daglim 3967, any greater daglim produces a hashrate drop to 0-1 Mhs

If I use PhoenixMiner for the 5700 and Lolminer with 480 it makes 30,31Mhs

Driver 20.8.1, Windows.

Any suggestion?
sr. member
Activity: 1484
Merit: 253
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

This test was 2 part... Mainly was about how the console hash comes is pretty much always going to be 3.5% higher than at the pool.... There will be days where you actually get it (as you can see in my test, but odds are going to be very slim) Also, TRM produced a better over hashrate, over the course of the year, that .75-1% will add up quickly.

Power on both miners were so close that it was marginal.
I didn't know about your system, but for me PM uses about 1-2% less power only from sensors (overall power economy from wall can be more). So 0.75-1% less efficiency but 1-2% less power.
Until lol and TR didn't lower power draw by their miners for me choise is obvious.
jr. member
Activity: 195
Merit: 4
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

This test was 2 part... Mainly was about how the console hash comes is pretty much always going to be 3.5% higher than at the pool.... There will be days where you actually get it (as you can see in my test, but odds are going to be very slim) Also, TRM produced a better over hashrate, over the course of the year, that .75-1% will add up quickly.

Power on both miners were so close that it was marginal.
jr. member
Activity: 195
Merit: 4
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

I certainly could write a kernel using way less power (actually one for 8Gs is on the way) ... but pushing that too far would need me to fake the hashrate to get users on it... ahm ... wait...

Sorry Lolli that your miner wasn't included but it was after everything already started... Perhaps I will do another one after the 1st of the year for some others that I had missed Smiley
member
Activity: 433
Merit: 48
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!

I certainly could write a kernel using way less power (actually one for 8Gs is on the way) ... but pushing that too far would need me to fake the hashrate to get users on it... ahm ... wait...
sr. member
Activity: 1484
Merit: 253
You forgot about PM uses less power than any other miner like lol or TR. So his loses in effectivity about some part of percent compensates by power economy in some percent!
member
Activity: 221
Merit: 12
-cdmport is not working in the new version...
newbie
Activity: 32
Merit: 0
I try new version PM 5.2a on my gtx 1070 rig. So far I don't notice any changes compared to the version 5.0e
newbie
Activity: 74
Merit: 0
Who ran 6 GB on 1060, what did they write in the batch file, what is the speed ETH? Thanks
newbie
Activity: 16
Merit: 0
@Phoenixminer: the 5.2a will compatible with the ETCHASH?
Jump to: