[ANN] TeamRedMiner v0.10.10 - Ironfish/Kaspa/ZIL/Kawpow/Etchash and More

[kerney666]

TMR is hands down the fastest Vega CN8 out there, that negates increased power consumption that CN8 brought to us. It is well worth to try. Dev fee of 2.5 % may seem high but it is worth the price, the devs are active all the time and try to help in every way.

There are 2 things that I would like to see.
1. Miner uptime. Maybe on some of the lines the information about miner uptime can be squeezed into.
2. Watchdog. Once i got my rigs stable, i didn't experience single GPU dropout. however, sometimes it happens that miner starts and initilazies all the cards, but hashes are 0. Some simple watchdog would be very useful for such occasion.

Hey, thanks for the kind words and the time you've invested in this to get it running smoothly on all your rigs!

Miner uptime - just put it on the todo list. Watchdog as well, will be added shortly.

[kerney666]

Windows 7 ?
does the miner work?

Sorry so say we haven't tested on Win7 at all ourselves, and all reports point to no, doesn't work . If we ever get some time over in the near future, I'll try to get a Win7 env going and see what's going on.

[kerney666]

I'm looking at some of my stats being show in the tool.   I have a rig with 9 GPUs and I see that the number of results for each GPU varies quite a bit.  I'm at 300 results and one of the 9 GPUs shows 46/0 while the lowest shows 23/0.  The cn v8 hashrate avg is 2.178khs and 2.102 khs so they are pretty close in speed which leads me to think the lowest GPU isnt getting enough work.

Are the jobs balanced across the GPUs evenly or is there some other mechanism?

Also, how long should it take before the reported pool speed matches to what I see in supportXMR or MoneroOcean? I know the dev fee will take a small amount of the totals but I am seeing a pretty large difference between what the GPU average is and the pool rate.  Almost 1khs.  There arent any reported failures.

I have thought a number of times that I should write a primer on the math involved in the mining process so people can plug in numbers and verify certain situations. A super quick back-of-the-envelope calculation for you finding 300 shares when you maybe should have found 306 shares gives that this would happen with a little less than 40% probability, so really nothing out of the ordinary.

For the individual gpus, it's much too few trials to get a good read.

How is the poolside stats looking now, a few hours later?

[ku4eto]

hello, hope this helps someone:

my bat:
teamredminer.exe -a cnv8 -o stratum+tcp://xmr-eu1.nanopool.org:14444 -u mywallet -p x -d 0,1,2,3,4,5,6,7,8,9,10,11 --cn_config 7+7,7+7,7+7,8+8,8+8,8+8,7+7,8+8,8+8,8+8,7+7,8+8

my cards
core is for all cards 1200/900mV, mem is or each cards below:

1   XTR 8G Hynix 2200 H5GQ8H24MJR
2   XTR 8G Hynix 75,4 2225 H5GQ8H24MJR

3   GIGABYTE RX 580 GAMING 4 GB  Hynix 76.1 2000 H5GC4H24AJR
4   XFX RX 580 GTS XXX 8 GB Samsung 73.8 2025 K4G80325FB

5   GIGABYTE RX 580 GAMING 8 GB Hynix 74.4 2200 H5GC8H24MJR
6   GIGABYTE RX 580 GAMING 8 GB Hynix 78.2 2200 H5GC8H24MJR

7   GIGABYTE RX 580 GAMING 4 GB Hynix 73.9 2000 H5GC4H24AJR
8   GIGABYTE RX 580 GAMING 4 GB Hynix 71,5 2000 H5GC4H24AJR
9   GIGABYTE RX 580 GAMING 4 GB Hynix 75.8 2000 H5GC4H24AJR
10 GIGABYTE RX 580 GAMING 4 GB Hynix 77.2 1975 H5GC4H24AJR
 
11 Sapphire PULSE RX 580 8 GB Samsung 76.3 2050 K4G80325FB
12 GIGABYTE AORUS RX 580 8 GB Hynix 71,3 2100 H5GC8H24MJR

this is my results:
855
871
912
953
858
860
912
755
910
894
1032
822

i dont understand why i cant get above 900 for almost all cards..any advice?
P.S. ignore cn_config order as it is not the same, just 4GB has 7+7 and 8gb has 8+8 results and my cards match,thats important.

Increase core clocks maybe? Or your memory straps are probably bad. Or both. At 1300/2100, a 580 4GB Samsung does ~1070h/s.



Also, the reported stats from the API... i kinda do not understand one thing from it:

Code:

ku4eto@KU4ETO:/mnt/c/Users/ku4eto/Desktop$ echo -n "gpu|1" | nc 192.168.0.106 4028
STATUS=S,When=1541273696,Code=17,Msg=GPU1,Description=TeamRedMiner 0.3.5|GPU=1,Enabled=Y,Status=Alive,Temperature=0.00,Fan Speed=0,Fan Percent=0,GPU Clock=0,Memory Clock=0,GPU Voltage=0.000,GPU Activity=0,Powertune=0,MHS av=0.001071,MHS 30s=0.001073,KHS av=1.071,KHS 30s=1.073,Accepted=3513,Rejected=2,Hardware Errors=0,Utility=2.115,Intensity=20,XIntensity=0,RawIntensity=0,Last Share Pool=0,Last Share Time=1541273664,Total MH=106737536.0000,Diff1 Work=105460424.693228,Difficulty Accepted=105390424.41133447,Difficulty Rejected=60000.24162330,Last Share Difficulty=30000.12081165,Last Valid Work=1541273664,Device Hardware%=0.0000,Device Rejected%=0.0569,Device Elapsed=99665|

Why the Share size is not an Integer number, instead, its Decimal O_o. This is, considering that i am using static difficulty of 30000. The Rejected Difficulty shows 60k, which is 2x30k shares, but there are some decimals appended. Same with Last Share Difficulty.

[keksik]

hello, hope this helps someone:

my bat:
teamredminer.exe -a cnv8 -o stratum+tcp://xmr-eu1.nanopool.org:14444 -u mywallet -p x -d 0,1,2,3,4,5,6,7,8,9,10,11 --cn_config 7+7,7+7,7+7,8+8,8+8,8+8,7+7,8+8,8+8,8+8,7+7,8+8

my cards
core is for all cards 1200/900mV, mem is or each cards below:

1   XTR 8G Hynix 2200 H5GQ8H24MJR
2   XTR 8G Hynix 75,4 2225 H5GQ8H24MJR

3   GIGABYTE RX 580 GAMING 4 GB  Hynix 76.1 2000 H5GC4H24AJR
4   XFX RX 580 GTS XXX 8 GB Samsung 73.8 2025 K4G80325FB

5   GIGABYTE RX 580 GAMING 8 GB Hynix 74.4 2200 H5GC8H24MJR
6   GIGABYTE RX 580 GAMING 8 GB Hynix 78.2 2200 H5GC8H24MJR

7   GIGABYTE RX 580 GAMING 4 GB Hynix 73.9 2000 H5GC4H24AJR
8   GIGABYTE RX 580 GAMING 4 GB Hynix 71,5 2000 H5GC4H24AJR
9   GIGABYTE RX 580 GAMING 4 GB Hynix 75.8 2000 H5GC4H24AJR
10 GIGABYTE RX 580 GAMING 4 GB Hynix 77.2 1975 H5GC4H24AJR
 
11 Sapphire PULSE RX 580 8 GB Samsung 76.3 2050 K4G80325FB
12 GIGABYTE AORUS RX 580 8 GB Hynix 71,3 2100 H5GC8H24MJR

this is my results:
855
871
912
953
858
860
912
755
910
894
1032
822

i dont understand why i cant get above 900 for almost all cards..any advice?
P.S. ignore cn_config order as it is not the same, just 4GB has 7+7 and 8gb has 8+8 results and my cards match,thats important.

Increase core clocks maybe? Or your memory straps are probably bad. Or both. At 1300/2100, a 580 4GB Samsung does ~1070h/s.



Also, the reported stats from the API... i kinda do not understand one thing from it:

Code:

ku4eto@KU4ETO:/mnt/c/Users/ku4eto/Desktop$ echo -n "gpu|1" | nc 192.168.0.106 4028
STATUS=S,When=1541273696,Code=17,Msg=GPU1,Description=TeamRedMiner 0.3.5|GPU=1,Enabled=Y,Status=Alive,Temperature=0.00,Fan Speed=0,Fan Percent=0,GPU Clock=0,Memory Clock=0,GPU Voltage=0.000,GPU Activity=0,Powertune=0,MHS av=0.001071,MHS 30s=0.001073,KHS av=1.071,KHS 30s=1.073,Accepted=3513,Rejected=2,Hardware Errors=0,Utility=2.115,Intensity=20,XIntensity=0,RawIntensity=0,Last Share Pool=0,Last Share Time=1541273664,Total MH=106737536.0000,Diff1 Work=105460424.693228,Difficulty Accepted=105390424.41133447,Difficulty Rejected=60000.24162330,Last Share Difficulty=30000.12081165,Last Valid Work=1541273664,Device Hardware%=0.0000,Device Rejected%=0.0569,Device Elapsed=99665|

Why the Share size is not an Integer number, instead, its Decimal O_o. This is, considering that i am using static difficulty of 30000. The Rejected Difficulty shows 60k, which is 2x30k shares, but there are some decimals appended. Same with Last Share Difficulty.

u were correct man, thank you for tip. i tried use different timings from srb and i was able to increase hash for my rx 580 4gb from 912 to 1015! still playing with 8gb cards.
btw do you know how to load bios for card more than 10? i have 11 and 12 cards too and cant go with atiflash above 0-9

EDIT: please can someone share with me timings for 8GB hynix H5GQ8H24MJR and H5GC8H24MJR?
i used 777000000000000022AA1C00B56A6D46C0551017BE8E060C006AE6000C081420EA8900AB0300000 01B162C31C0313F17 from pbe 1.6.9
and now using 999000000000000022559D0052626C48A0551214BC0D460B0048C4007D0714204A8900A00200712 419123138B42D3D17 from SRBpolaris v3.5
in first case 1250/2200 got 855
in second case i can go only 2150 and getting 878

in first article here, someone mentioned for rx 580 8gb hynix at 1250/2000 getting 1000+

[kerney666]

u were correct man, thank you for tip. i tried use different timings from srb and i was able to increase hash for my rx 580 4gb from 912 to 1015! still playing with 8gb cards.
btw do you know how to load bios for card more than 10? i have 11 and 12 cards too and cant go with atiflash above 0-9

EDIT: please can someone share with me timings for 8GB hynix H5GQ8H24MJR and H5GC8H24MJR?
i used 777000000000000022AA1C00B56A6D46C0551017BE8E060C006AE6000C081420EA8900AB0300000 01B162C31C0313F17 from pbe 1.6.9
and now using 999000000000000022559D0052626C48A0551214BC0D460B0048C4007D0714204A8900A00200712 419123138B42D3D17 from SRBpolaris v3.5
in first case 1250/2200 got 855
in second case i can go only 2150 and getting 878

in first article here, someone mentioned for rx 580 8gb hynix at 1250/2000 getting 1000+

The 580 8GB Hynix is one of our testing 580 8GB cards. Need to check the straps, and also make 100% sure it's Hynix so I wasn't lying . When you're at 1250 cclk and still at 855 h/s with a 580 8GB using 8+8 for config, it's clearly mem bound and better straps is the only thing you should look for. Hynix usually produces > 1000 h/s though, we've had more problems with Elpida.

Not sure if it works, but if you e.g. disable two other cards in the device manager, is card 11+12 moved down to 9+10 and visible in atiflash?

[kerney666]

Also, the reported stats from the API... i kinda do not understand one thing from it:

Code:

ku4eto@KU4ETO:/mnt/c/Users/ku4eto/Desktop$ echo -n "gpu|1" | nc 192.168.0.106 4028
STATUS=S,When=1541273696,Code=17,Msg=GPU1,Description=TeamRedMiner 0.3.5|GPU=1,Enabled=Y,Status=Alive,Temperature=0.00,Fan Speed=0,Fan Percent=0,GPU Clock=0,Memory Clock=0,GPU Voltage=0.000,GPU Activity=0,Powertune=0,MHS av=0.001071,MHS 30s=0.001073,KHS av=1.071,KHS 30s=1.073,Accepted=3513,Rejected=2,Hardware Errors=0,Utility=2.115,Intensity=20,XIntensity=0,RawIntensity=0,Last Share Pool=0,Last Share Time=1541273664,Total MH=106737536.0000,Diff1 Work=105460424.693228,Difficulty Accepted=105390424.41133447,Difficulty Rejected=60000.24162330,Last Share Difficulty=30000.12081165,Last Valid Work=1541273664,Device Hardware%=0.0000,Device Rejected%=0.0569,Device Elapsed=99665|

Why the Share size is not an Integer number, instead, its Decimal O_o. This is, considering that i am using static difficulty of 30000. The Rejected Difficulty shows 60k, which is 2x30k shares, but there are some decimals appended. Same with Last Share Difficulty.

The decimal representation of the difficulties are inherited from cgminer/sgminer, we wanted to keep it 100% compatible.

Next, the decimals come from a rounding effect. TL;DR is that everything is correct and the problem is that 30,000 isn't representable as a target _integer_, and what you're really mining against is pool diff 30,000.1208. The details follow below, might be a bit technical but you seem to be savvy enough to understand it!

So, when the pool says diff 30000, what it means is that (for CN) on average, every 30,000th hash should be eligible to submit to the pool. A hash is just a big random 256-bit number. To decide if a hash is good enough, we need another 256-bit number to compare it to. If a hash is lower than this number, we submit it to the pool.

Hence, we can't just compare a hash to the pool diff 30000, we need to convert it to a 256-bit integer first. Typically, you only compare the upper 32 or 64 (most common) bits of the hash. To produce a 32-bit number, we take 2^32 / 30000 = 143165.577. Here is the issue: this is not an integer, and we need to floor it rather than round it upwards or we will accept shares with diff 29999.999. So, the 32-bit number you compare with is 143165. This is the _true_ difficulty target we're comparing hashes against.

This number (143165) does not represent difficulty 30000. To convert it back, we do X = 2^32 / 143165 = 30,000.1208, which is the number you're seeing. The end result is that you are mining against diff 30,000.1208, because mining against exactly pool diff 30,000 is impossible, and that is why you're seeing this number everywhere in the api numbers.

Hope that helps!

[keksik]

u were correct man, thank you for tip. i tried use different timings from srb and i was able to increase hash for my rx 580 4gb from 912 to 1015! still playing with 8gb cards.
btw do you know how to load bios for card more than 10? i have 11 and 12 cards too and cant go with atiflash above 0-9

EDIT: please can someone share with me timings for 8GB hynix H5GQ8H24MJR and H5GC8H24MJR?
i used 777000000000000022AA1C00B56A6D46C0551017BE8E060C006AE6000C081420EA8900AB0300000 01B162C31C0313F17 from pbe 1.6.9
and now using 999000000000000022559D0052626C48A0551214BC0D460B0048C4007D0714204A8900A00200712 419123138B42D3D17 from SRBpolaris v3.5
in first case 1250/2200 got 855
in second case i can go only 2150 and getting 878

in first article here, someone mentioned for rx 580 8gb hynix at 1250/2000 getting 1000+

The 580 8GB Hynix is one of our testing 580 8GB cards. Need to check the straps, and also make 100% sure it's Hynix so I wasn't lying . When you're at 1250 cclk and still at 855 h/s with a 580 8GB using 8+8 for config, it's clearly mem bound and better straps is the only thing you should look for. Hynix usually produces > 1000 h/s though, we've had more problems with Elpida.

Not sure if it works, but if you e.g. disable two other cards in the device manager, is card 11+12 moved down to 9+10 and visible in atiflash?

yes im sure i have hynix and types i described above. most probably it is caused by the straps as im still around 850h/s with 8gb cards...4gb hash around 1000 and im using 1250core.
can u share with us some tips/timings/numbers u had with elpida? i have 4gb elpida cards, so curious about numbers and problems to avoid.

that might work,will try, but these are 8gb hynix,so will wait for new info about the strap used in your test scenario:)

[kerney666]

yes im sure i have hynix and types i described above. most probably it is caused by the straps as im still around 850h/s with 8gb cards...4gb hash around 1000 and im using 1250core.
can u share with us some tips/timings/numbers u had with elpida? i have 4gb elpida cards, so curious about numbers and problems to avoid.

that might work,will try, but these are 8gb hynix,so will wait for new info about the strap used in your test scenario:)

You know, I'm def not a mem strap wizard. For this particular 580, I'm quite sure it was a one click Pimp My Straps in the latest version of SRB PBE (the free one without any donation delays). I will check though, no worries!

I believe dragonmike managed to get his 470 4GB (Elpida) going in a nice way in the end, see if you can find his posts earlier in this thread.

[keksik]

anybody with hasrates for 470 and / or 570 4GB?

ASUS RX570 Expedition OC (modded Elpida bios) clocked at 1225/1940 yields just under 1kh/s. cn_config 7+7. 8+7 seems to be slower.

You've got to give it a go and see for yourself which config will work best, but the above is a pretty safe bet.

hello,

please can u share with me your timing and what voltage u used for core/mem?

thank you in advance

[Bael]

Hi

First of all excuse my ignorance, I have only used Srbminer to mine cryptonight.

What software could I use to monitor the miner?

regards!! : )

[kerney666]

Hi

First of all excuse my ignorance, I have only used Srbminer to mine cryptonight.

What software could I use to monitor the miner?

regards!! : )

Haha, no worries mate, after all we only added CNv8 less than a week ago!

The miner comes with a sgminer/cgminer compatible API. It's not the standard in the CN world, and we will build a little adapter that exposes an xmr stak-like http api and gui so all CN miners can just continue with what they were using earlier. We're still missing fans and temps though, needs to be added, and we will.

There are a number of tools which supports the API: Awesome Miner and MinerStats are two that have confirmed integration.

Cheers,
K

[Bael]

Hi

First of all excuse my ignorance, I have only used Srbminer to mine cryptonight.

What software could I use to monitor the miner?

regards!! : )

Haha, no worries mate, after all we only added CNv8 less than a week ago!

The miner comes with a sgminer/cgminer compatible API. It's not the standard in the CN world, and we will build a little adapter that exposes an xmr stak-like http api and gui so all CN miners can just continue with what they were using earlier. We're still missing fans and temps though, needs to be added, and we will.

There are a number of tools which supports the API: Awesome Miner and MinerStats are two that have confirmed integration.

Cheers,
K

Well, I predict a great future for your miner : )

I will look awesome miner and minerstats (this one I have not tried)

Thank you!!

[Shuree Rida]

Has anyone got this working on HiveOS?

[TrM_TrM]

Miner hangs the RIG when trying to close it.
Same problem on two different RIG's.
1- G3900 / ram 4gb / 4xRX570 4GB + 3 RX470 8GB / LTSB 14393 / 18.6.1
2- G3900 / ram 4gb / 7xRX470 4GB /  / LTSB 14393 / 18.6.1

[peteris-apse]

So I am meesing arround with settings...

7x vega 56 reference cards rig. 1408/875 1100/875

16+16 just hangs at launch
16+15 starts to work, but restarts rig after few minutes
16+14 works, but restarts rig after 30 mins or so. which never happend with other miners.14.3khs 1250w from wall
14+14 gives a little lower hashrate, but seems to be stable, also zero lost shares. 14.1khs  1250w from wall

[ku4eto]

Miner hangs the RIG when trying to close it.
Same problem on two different RIG's.
1- G3900 / ram 4gb / 4xRX570 4GB + 3 RX470 8GB / LTSB 14393 / 18.6.1
2- G3900 / ram 4gb / 7xRX470 4GB /  / LTSB 14393 / 18.6.1

Are you closing it down with the "X"? If so, try using CTRL+C and then typing "Y" + Enter.

If it still does, maybe your cards are not happy with the straps.


Also, the miner needs a good exit way. The old sgminer had "q" for exiting. The allocation would need to be cleared (have not actually checked if its doing so) when exiting. This can prevent crashes when re-launching the miner multiple times.


EDIT:


Performance tests:

1280/1950 Elpida
1300/2100 Samsung
1300/1950 Elpida
1300/1925 Elpida
At 0.95V
Doing 4260h/s at ~530W at wall.

1280/1950
1280/2100
1280/1950
1280/1925
At 0.925V
Doing 4180h/s at ~515W at wall.

1250/1950
1250/2100
1250/1950
1250/1925
At 0.925V
Doing 4060h/s at ~500W at wall.

I did not really try 1250@0.9V, performance loss seems to be going 1:1 with power consumption reduction below 1280 for me.

Compared to XMR-Stak: First decribed clocks at same voltage were giving me 4090h/s@520W on CNv7, and 3870h/s@570W on CNv8.

EDIT2: I saw something weird. Fails to connect to the Devpool for some reason.



Even after restarting the miner, its the same. Starts with the same message.

[keksik]

Miner hangs the RIG when trying to close it.
Same problem on two different RIG's.
1- G3900 / ram 4gb / 4xRX570 4GB + 3 RX470 8GB / LTSB 14393 / 18.6.1
2- G3900 / ram 4gb / 7xRX470 4GB /  / LTSB 14393 / 18.6.1

Are you closing it down with the "X"? If so, try using CTRL+C and then typing "Y" + Enter.

If it still does, maybe your cards are not happy with the straps.


Also, the miner needs a good exit way. The old sgminer had "q" for exiting. The allocation would need to be cleared (have not actually checked if its doing so) when exiting. This can prevent crashes when re-launching the miner multiple times.


EDIT:


Performance tests:

1280/1950 Elpida
1300/2100 Samsung
1300/1950 Elpida
1300/1925 Elpida
At 0.95V
Doing 4260h/s at ~530W at wall.

1280/1950
1280/2100
1280/1950
1280/1925
At 0.925V
Doing 4180h/s at ~515W at wall.

1250/1950
1250/2100
1250/1950
1250/1925
At 0.925V
Doing 4060h/s at ~500W at wall.

I did not really try 1250@0.9V, performance loss seems to be going 1:1 with power consumption reduction below 1280 for me.

Compared to XMR-Stak: First decribed clocks at same voltage were giving me 4090h/s@520W on CNv7, and 3870h/s@570W on CNv8.

EDIT2: I saw something weird. Fails to connect to the Devpool for some reason.

hi man, nice results! please can u share with me what cards are u using for each number? also maybe timings for elpida and sasmung? im not able to go above 950h/s or so...u have 4gb or 8gb?

[ku4eto]

Miner hangs the RIG when trying to close it.
Same problem on two different RIG's.
1- G3900 / ram 4gb / 4xRX570 4GB + 3 RX470 8GB / LTSB 14393 / 18.6.1
2- G3900 / ram 4gb / 7xRX470 4GB /  / LTSB 14393 / 18.6.1

Are you closing it down with the "X"? If so, try using CTRL+C and then typing "Y" + Enter.

If it still does, maybe your cards are not happy with the straps.


Also, the miner needs a good exit way. The old sgminer had "q" for exiting. The allocation would need to be cleared (have not actually checked if its doing so) when exiting. This can prevent crashes when re-launching the miner multiple times.


EDIT:


Performance tests:

1280/1950 Elpida
1300/2100 Samsung
1300/1950 Elpida
1300/1925 Elpida
At 0.95V
Doing 4260h/s at ~530W at wall.

1280/1950
1280/2100
1280/1950
1280/1925
At 0.925V
Doing 4180h/s at ~515W at wall.

1250/1950
1250/2100
1250/1950
1250/1925
At 0.925V
Doing 4060h/s at ~500W at wall.

I did not really try 1250@0.9V, performance loss seems to be going 1:1 with power consumption reduction below 1280 for me.

Compared to XMR-Stak: First decribed clocks at same voltage were giving me 4090h/s@520W on CNv7, and 3870h/s@570W on CNv8.

EDIT2: I saw something weird. Fails to connect to the Devpool for some reason.

hi man, nice results! please can u share with me what cards are u using for each number? also maybe timings for elpida and sasmung? im not able to go above 950h/s or so...u have 4gb or 8gb?

All cards are 4GB.

Cards in order are:

RX 570 Gigabyte Aorus
RX 580 Red Dragon
RX 580 MSI Gaming X
RX 580 MSI Gaming X

I am core bottlenecked at around 1280-1300Mhz for all 4 cards.

The straps are not public.

Also, more on the Dev Pool connection.

After even restarting the rig, issue persists, along with new message.

[peteris-apse]

Have this error all the time. Is it messing with my hashrate?

Code:

[2018-11-04 16:08:26] Dev pool failed to connect.

I dont want to block anything. I am mining on nanopool with hiveos. Can we fix this? For me it seems that miner trys to connect to dev pool all the time therefore is messing with hashrate. It is fluctuating from 6 to 11mhs one rig.

edit: Also same errors at same time on Windows rig with vega. With the hashrate also power consumption is also fluctuating for all rigs mining monero at same time.

Am I right or is this considered to be normal?