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

[kerney666]

strange issue on my side...but dont know if it´s version related

upgraded my 8xVega Rig with 2 Vegas

using 18.6.1 drivers and everything else like with 8 cards

so starting TRMiner with 8 cards...everything is good
starting TRMiner with 9 cards..BSOD with "Video Scheduler Internal Error" - it dont matter which GPU i use...i can mixe all gpus and it runs fine with only 8 cards in the batch. After adding a 9. card - BSOD


with SRBMiner i can use all 10 GPUs without error...any hint for me what can cause this problem with TRMiner?

I doubt this is it, but just in case... did you increase your windows page file size after adding the other 2 cards (or make sure it was big enough before)?

it was/is 96.000

i dont know what to do now...8 cards perfect running...9 cards dont :/

Hi! Sorry to say I only have a 8 x Vega rig, need to cram another riser in there to test. What happens if you run two miner instances, one with -d 0,1,2,3 and another with -d 4,5,6,7,8?

[nordmann666]

strange issue on my side...but dont know if it´s version related

upgraded my 8xVega Rig with 2 Vegas

using 18.6.1 drivers and everything else like with 8 cards

so starting TRMiner with 8 cards...everything is good
starting TRMiner with 9 cards..BSOD with "Video Scheduler Internal Error" - it dont matter which GPU i use...i can mixe all gpus and it runs fine with only 8 cards in the batch. After adding a 9. card - BSOD


with SRBMiner i can use all 10 GPUs without error...any hint for me what can cause this problem with TRMiner?

I doubt this is it, but just in case... did you increase your windows page file size after adding the other 2 cards (or make sure it was big enough before)?

it was/is 96.000

i dont know what to do now...8 cards perfect running...9 cards dont :/

Hi! Sorry to say I only have a 8 x Vega rig, need to cram another riser in there to test. What happens if you run two miner instances, one with -d 0,1,2,3 and another with -d 4,5,6,7,8?

just the same...crash with video sheduler...i will buy another 2 cards and go back to 2x6 card rig

[dragonmike]

If I remember correctly, one cannot run more than 8 AMD cards in one rig... At least not in Windows. Look up the reviews of the 18 slot mining mobos and they'll mention just that thing.

Iirc bitsbetrippin made a video as well, look it up.

EDIT: but then again what do I know, that info might very well be outdated! 

[misterkit]

This isnt true anymore. 

If I remember correctly, one cannot run more than 8 AMD cards in one rig... At least not in Windows. Look up the reviews of the 18 slot mining mobos and they'll mention just that thing.

Iirc bitsbetrippin made a video as well, look it up.

[osirus466]

I have an 8 card rig getting 8k/s total.

5- MSI RX470 8GB mining edition
3- Nitro RX570 4GB

Im getting 980 h/s from the 470s and 1k/s from the 570s but at 125 watts per card.  Total rig power at the wall is 1000watts.  Is there any way I van lower the power draw???

Hi! I daresay this miner is the most power efficient miner for all the variants we support. The tools at your disposal are bios and clocks tweaks. There are other people here that have trimmed their Polaris rigs much more than I have, hopefully then can pitch in.

As a start, can you tell us a little more about your setup? What clocks do you run the cards at currently?

I have the RX570s at 1250/900 core 2050/900 mem  and the RX470s at 1250/900 core and 2100/900 mem.  Even lower settings it doest save much energy at all.  I have them modded with the pimp my straps mod.  And Im using the blockchain drivers.

[seefatlow]

strange issue on my side...but dont know if it´s version related

upgraded my 8xVega Rig with 2 Vegas

using 18.6.1 drivers and everything else like with 8 cards

so starting TRMiner with 8 cards...everything is good
starting TRMiner with 9 cards..BSOD with "Video Scheduler Internal Error" - it dont matter which GPU i use...i can mixe all gpus and it runs fine with only 8 cards in the batch. After adding a 9. card - BSOD


with SRBMiner i can use all 10 GPUs without error...any hint for me what can cause this problem with TRMiner?

I doubt this is it, but just in case... did you increase your windows page file size after adding the other 2 cards (or make sure it was big enough before)?

it was/is 96.000

i dont know what to do now...8 cards perfect running...9 cards dont :/

Hi! Sorry to say I only have a 8 x Vega rig, need to cram another riser in there to test. What happens if you run two miner instances, one with -d 0,1,2,3 and another with -d 4,5,6,7,8?

just the same...crash with video sheduler...i will buy another 2 cards and go back to 2x6 card rig

@nordmann666 I can vouch the miner is able to work with more than 8 cards under win10. I have a 12 card rig that uses this miner. Whatever problem you may have might be either due to drivers or pagefile size. Try increasing your pagefile to 80GB or more and if it still doesn't work, try re-installing the drivers. Check your riser cards as well.

[seefatlow]

Low power consumption TRTL mining is a hit or miss thingy and seems to depend very much on silicon lottery.

This can't be the case... I have 8 nitro 64s on a rig pulling 1112w at the wall - maybe ~35w for system, so that's <135w per GPU on average, each doing 18.5 kh/s (850/1107/818/L18+18.)  I also have a ref 64 and a flashed ref 56 (850/1107/818/L22+22) doing the same on another rig.  These were all bought over a period of many months during the vega stock shortages, so they certainly weren't all binned together.

Granted I'm on platinum PSUs and a 240v line, so your power needs could be slightly higher, but prob not more than 5%-10% max.  I'm also on linux, but my tests on windows (18.12.x drivers iirc) showed similar results.

Really, memory manufacturer is the only thing I see impacting performance - mine are all samsung, and i expect based on widespread reports, and my experience w/ polaris, that hynix would underperform.  Drivers are the other possibility - though I have yet to see any versions since blockchain significantly impact performance for any algo (other than on xmr-stack derived miners.)

8 nitros 64 pulling only 1112W at the wall is mighty impressive!! I think you might be the only person being able to pull that off

The problem with my rig is none of my Vega64 nor Vega56 cards is able to run at 1200/1100/<880mv/L28+28(L24+24) without crashing. Same goes for CNr algo. Can't go below 870mV. So, I am really baffled how you are able to pull off that power draw. Wonder if temperature have a role in the stability that you are getting with that extreme undervolt. It's hot where I live and my GPU temp is 65C.

[pbfarmer]

Low power consumption TRTL mining is a hit or miss thingy and seems to depend very much on silicon lottery.

This can't be the case... I have 8 nitro 64s on a rig pulling 1112w at the wall - maybe ~35w for system, so that's <135w per GPU on average, each doing 18.5 kh/s (850/1107/818/L18+18.)  I also have a ref 64 and a flashed ref 56 (850/1107/818/L22+22) doing the same on another rig.  These were all bought over a period of many months during the vega stock shortages, so they certainly weren't all binned together.

Granted I'm on platinum PSUs and a 240v line, so your power needs could be slightly higher, but prob not more than 5%-10% max.  I'm also on linux, but my tests on windows (18.12.x drivers iirc) showed similar results.

Really, memory manufacturer is the only thing I see impacting performance - mine are all samsung, and i expect based on widespread reports, and my experience w/ polaris, that hynix would underperform.  Drivers are the other possibility - though I have yet to see any versions since blockchain significantly impact performance for any algo (other than on xmr-stack derived miners.)

8 nitros 64 pulling only 1112W at the wall is mighty impressive!! I think you might be the only person being able to pull that off

The problem with my rig is none of my Vega64 nor Vega56 cards is able to run at 1200/1100/<880mv/L28+28(L24+24) without crashing. Same goes for CNr algo. Can't go below 870mV. So, I am really baffled how you are able to pull off that power draw. Wonder if temperature have a role in the stability that you are getting with that extreme undervolt. It's hot where I live and my GPU temp is 65C.

The 850 cclock and cn_config=L18+18 are key.  IIRC, for the same exact voltage setting, I saw a 15w difference in power consumption just increasing cn_config to L28+28. 

I also had a ton of trouble trying to find stable settings for the 1950 h/s range - at the end of the day, the time spent hunting for stable settings, plus the extra power required, just weren't worth the extra 1kh/s.

That being said, if you can't get < 880mv, you likely have a problem somewhere else, possibly your ppt - try disabling all other states, or make sure you don't have out-of-order voltages (since 19.x, many people are forgoing ppts, which is a mistake given that you still can't change the mem-p2 voltage setting.)  And while higher temps do require more power, it's also all the more reason to get your voltage down...

[QuirkSilver]

I have an 8 card rig getting 8k/s total.

5- MSI RX470 8GB mining edition
3- Nitro RX570 4GB

Im getting 980 h/s from the 470s and 1k/s from the 570s but at 125 watts per card.  Total rig power at the wall is 1000watts.  Is there any way I van lower the power draw???

Hi! I daresay this miner is the most power efficient miner for all the variants we support. The tools at your disposal are bios and clocks tweaks. There are other people here that have trimmed their Polaris rigs much more than I have, hopefully then can pitch in.

As a start, can you tell us a little more about your setup? What clocks do you run the cards at currently?

I have the RX570s at 1250/900 core 2050/900 mem  and the RX470s at 1250/900 core andcore. 0/900 mem.  Even lower settings it doest save much energy at all.  I have them modded with the pimp my straps mod.  And Im using the blockchain drivers.

Hey, when I first tried this miner (v8) I have read the readme guide, this miner needs bit higher undervolt (windowa 10 here), but unlike other mines, even 1150 mhz core gives you still near max speed. I think you have an issue with speeds, see when core increasing starts to give deminishing returns.... going from 1150 to 1250 for +30 h/s is so inoptimum.  I wouldn't use 8gb ever to mine monero. Bittubev2 and heavy variants, speed is 1200+ with proper timings and 1250 core. The 4gb is surely fine. Notice 8gb filled of gddr5 filled with data conaumes more than 4gb.. so I would use those for an algo that rewards (most heavy variants) and keep 4gb 570 to monero. (560 2gbs I found even more efficient, more than vega if for one reason)

[okanist]

Hello, do you have any plans for adding support of CN-Conceal algorithm? If you are interested, please let me know and I'll share the contacts.

[dragonmike]

Low power consumption TRTL mining is a hit or miss thingy and seems to depend very much on silicon lottery.

This can't be the case... I have 8 nitro 64s on a rig pulling 1112w at the wall - maybe ~35w for system, so that's <135w per GPU on average, each doing 18.5 kh/s (850/1107/818/L18+18.)  I also have a ref 64 and a flashed ref 56 (850/1107/818/L22+22) doing the same on another rig.  These were all bought over a period of many months during the vega stock shortages, so they certainly weren't all binned together.

Granted I'm on platinum PSUs and a 240v line, so your power needs could be slightly higher, but prob not more than 5%-10% max.  I'm also on linux, but my tests on windows (18.12.x drivers iirc) showed similar results.

Really, memory manufacturer is the only thing I see impacting performance - mine are all samsung, and i expect based on widespread reports, and my experience w/ polaris, that hynix would underperform.  Drivers are the other possibility - though I have yet to see any versions since blockchain significantly impact performance for any algo (other than on xmr-stack derived miners.)

8 nitros 64 pulling only 1112W at the wall is mighty impressive!! I think you might be the only person being able to pull that off

The problem with my rig is none of my Vega64 nor Vega56 cards is able to run at 1200/1100/<880mv/L28+28(L24+24) without crashing. Same goes for CNr algo. Can't go below 870mV. So, I am really baffled how you are able to pull off that power draw. Wonder if temperature have a role in the stability that you are getting with that extreme undervolt. It's hot where I live and my GPU temp is 65C.

The 850 cclock and cn_config=L18+18 are key.  IIRC, for the same exact voltage setting, I saw a 15w difference in power consumption just increasing cn_config to L28+28. 

I also had a ton of trouble trying to find stable settings for the 1950 h/s range - at the end of the day, the time spent hunting for stable settings, plus the extra power required, just weren't worth the extra 1kh/s.

That being said, if you can't get < 880mv, you likely have a problem somewhere else, possibly your ppt - try disabling all other states, or make sure you don't have out-of-order voltages (since 19.x, many people are forgoing ppts, which is a mistake given that you still can't change the mem-p2 voltage setting.)  And while higher temps do require more power, it's also all the more reason to get your voltage down...

When you say 850 cclock do you mean effective or setting in overdriventool?
850 in overdriventool is closer to 800 effective and that'd be super low...
Would L18+18 work for flashed 56's just as well?

[pbfarmer]

8 nitros 64 pulling only 1112W at the wall is mighty impressive!! I think you might be the only person being able to pull that off

The problem with my rig is none of my Vega64 nor Vega56 cards is able to run at 1200/1100/<880mv/L28+28(L24+24) without crashing. Same goes for CNr algo. Can't go below 870mV. So, I am really baffled how you are able to pull off that power draw. Wonder if temperature have a role in the stability that you are getting with that extreme undervolt. It's hot where I live and my GPU temp is 65C.

The 850 cclock and cn_config=L18+18 are key.  IIRC, for the same exact voltage setting, I saw a 15w difference in power consumption just increasing cn_config to L28+28. 

I also had a ton of trouble trying to find stable settings for the 1950 h/s range - at the end of the day, the time spent hunting for stable settings, plus the extra power required, just weren't worth the extra 1kh/s.

That being said, if you can't get < 880mv, you likely have a problem somewhere else, possibly your ppt - try disabling all other states, or make sure you don't have out-of-order voltages (since 19.x, many people are forgoing ppts, which is a mistake given that you still can't change the mem-p2 voltage setting.)  And while higher temps do require more power, it's also all the more reason to get your voltage down...

When you say 850 cclock do you mean effective or setting in overdriventool?
850 in overdriventool is closer to 800 effective and that'd be super low...
Would L18+18 work for flashed 56's just as well?

i'm using core-p0, so ACG is disabled -- I only see a droop to 849 effective vs 852 setting. 

My flashed 56 likes L22+22 best.

[dragonmike]

8 nitros 64 pulling only 1112W at the wall is mighty impressive!! I think you might be the only person being able to pull that off

The problem with my rig is none of my Vega64 nor Vega56 cards is able to run at 1200/1100/<880mv/L28+28(L24+24) without crashing. Same goes for CNr algo. Can't go below 870mV. So, I am really baffled how you are able to pull off that power draw. Wonder if temperature have a role in the stability that you are getting with that extreme undervolt. It's hot where I live and my GPU temp is 65C.

The 850 cclock and cn_config=L18+18 are key.  IIRC, for the same exact voltage setting, I saw a 15w difference in power consumption just increasing cn_config to L28+28.  

I also had a ton of trouble trying to find stable settings for the 1950 h/s range - at the end of the day, the time spent hunting for stable settings, plus the extra power required, just weren't worth the extra 1kh/s.

That being said, if you can't get < 880mv, you likely have a problem somewhere else, possibly your ppt - try disabling all other states, or make sure you don't have out-of-order voltages (since 19.x, many people are forgoing ppts, which is a mistake given that you still can't change the mem-p2 voltage setting.)  And while higher temps do require more power, it's also all the more reason to get your voltage down...

When you say 850 cclock do you mean effective or setting in overdriventool?
850 in overdriventool is closer to 800 effective and that'd be super low...
Would L18+18 work for flashed 56's just as well?

i'm using core-p0, so ACG is disabled -- I only see a droop to 849 effective vs 852 setting.  

My flashed 56 likes L22+22 best.

I had to set my p0 cclock a little higher (900) or else a GPU would drop... but with L22+22, 825mV on 4 Vegas and 850mV on the two weaker cards, I'm now hashing at just under 16 kh/s per card, so approx 95 kh/s for the rig at a power draw at the wall of 875W (which is indeed ~135W per GPU). That's close to 110 h/W. Pretty amazing in itw own right, but still far from your 18.5 kh/s. Not sure how you manage to get these hashrates at these voltages...

[ETHETCSIA]

Dear todxx

I have a question. Could you please answer it?

I added --log_interval=1 to the command line trm043 for my 6 x RX470 4GB rig..

the dos console shows me 400H/s and 810 H/s repeatedly.. 3 seconds about 400 and then 2 seconds about 800.

What is the reason of this?

Why does the hash rate fluctuate?

[kerney666]

Dear todxx

I have a question. Could you please answer it?

I added --log_interval=1 to the command line trm043 for my 6 x RX470 4GB rig..

the dos console shows me 400H/s and 810 H/s repeatedly.. 3 seconds about 400 and then 2 seconds about 800.

What is the reason of this?

Why does the hash rate fluctuate?

Hi! You have asked for hashrate logs every second with —log_interval=1. But, we only display hashes from rounds that have completed between two logs. Since a single hash round takes ~2 secs on Polaris cards, it will be random how many of the 6x2 threads that completed a round between each pair of log prints, and your output will vary a lot.

[ETHETCSIA]

Thank you. 3 or 4 second is ideal i think.

Dear todxx

I have a question. Could you please answer it?

I added --log_interval=1 to the command line trm043 for my 6 x RX470 4GB rig..

the dos console shows me 400H/s and 810 H/s repeatedly.. 3 seconds about 400 and then 2 seconds about 800.

What is the reason of this?

Why does the hash rate fluctuate?

Hi! You have asked for hashrate logs every second with —log_interval=1. But, we only display hashes from rounds that have completed between two logs. Since a single hash round takes ~2 secs on Polaris cards, it will be random how many of the 6x2 threads that completed a round between each pair of log prints, and your output will vary a lot.

[logicus]

After  testing TeamRedMiner 0.4.3 under MMP OS 2 (Ubuntu) ,the result is impressive:

TeamRedMiner 0.4.3 under MMP OS 2

 4xVega 56(Vega 64 bios) core 1050mHz/845mV memory 1100mHz/845mV  80kH/s @ 690W from wall

https://imgur.com/xBHYNJj

https://imgur.com/0Jjvpp3

https://imgur.com/eHv9Cxj


P.S.
Under win 10  my best result is 78kH/s @ 770W from wall

[dragonmike]

After  testing TeamRedMiner 0.4.3 under MMP OS 2 (Ubuntu) ,the result is impressive:

TeamRedMiner 0.4.3 under MMP OS 2

 4xVega 56(Vega 64 bios) core 1050mHz/845mV memory 1100mHz/845mV  80kH/s @ 690W from wall





https://imgur.com/eHv9Cxj


P.S.
Under win 10  my best result is 78kH/s @ 770W from wall

Interesting... Is there any type of mem timing tweaking going on in the background here? I assume it's that "smart tune" they refer to in their thread?...

[ZombieWorm]

Great stats there @dragonmike !

Hoping there is a heavy/haven update for this miner soon.

[logicus]

Yes,i set "Smart tune" from "off" to "auto" and set --cn_config L22+22