Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux)

[Er0l]

Simplemining just implemented Rxboost on linux  

I am running Rxboost on ethOS (Linux), RX474 32.1mh/s , before 31.4mh/s

Finally got 33 mh/s with RX470 - with rxboost

Please share wit us command line options and GPU OC settings ?

I have a lot of different 474's and cant get more then 30.8 with -rxboost 1 and -strap 1

I am running rx470 4Gb 1180 core 2150 mem 32.7mh/s with phoenixminer + rxboost on Linux

[1714871059]

1714871059

[1714871059]

1714871059

[1714871059]

1714871059

[UnclWish]

Does rxboost require the Claymore driver install? I tried to run it last night and got a bunch of errors, but I hadn't run the driver install setting but wanted to check if I needed the driver before I tried other troubleshooting.

Yes.

[UnclWish]

Anyone test the power consumption difference with "rxboost" command? Also for those who are having trouble keeping it "rxboost" stable on windows, try -rxboost 30 .... seems to be the most stable for Windows and SMOS (rxboost -i 30) rigs without crashing systems. I've attached a chart for a mini farm we operate (SMOS - 4.0 Gh/s), and we noticed a 3.38% overall gain withing the farm (approx. 135.74 Mh/s Increase overall). I attached a chart with details.

Again our main concern is power consumption increasing (hopefully decrease?  ) , as we currently don't have access to power readings due to farm being remotely located. We will test the power levels in person soon, but we would like to know a range for expectation purposes, if anyone already tested.


Chart Showing Gains from 'rxboost -i 30' (SMOS):

https://ibb.co/THyqpGP



Disclosure: Rig 11 and 13 seem to have been 'underclocked' prior so try to refrain from using those as a we also fine tuned some timings on those on top of 'rxboost'



https://www.RigSpec.io

Memory doesn't consumes so much power... Memory clock up to 2250 MHz on GDDR5 increases it a bit. But memory timings not so important to power, i think.

You are correct that the memory chips don't consume that much power with better straps but the GPU compute part of the graphic card actually does use more power overall.

Think of it this way. The GPU compute is actually idle alot because there is a bottle-neck in the memory, when the memory is faster the GPU compute is idle less and overall there is more power consumption. This is why back in 2017, it was very popular to dual mine coins with ETH because most of the time the memory is busy while the GPU compute part of the video card stays inactive and does nothing.

Hmmm, maybe you're right. I don't think this way...

[RAKOT]

Simplemining just implemented Rxboost on linux  

I am running Rxboost on ethOS (Linux), RX474 32.1mh/s , before 31.4mh/s

Finally got 33 mh/s with RX470 - with rxboost

Please share wit us command line options and GPU OC settings ?

I have a lot of different 474's and cant get more then 30.8 with -rxboost 1 and -strap 1

I am running rx470 4Gb 1180 core 2150 mem 32.7mh/s with phoenixminer + rxboost on Linux

Thanks for info , Im on W10 m maybee i should try Linux-based environment.

[big_daddy]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

[big_daddy]

Latest one, i personally use 19.3.2 with compute mode switcher and atikmdag patcher.

are your cards stock or modded bios and what are your results?

[Er0l]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

[RAKOT]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

Wait a minute , ist rxboost only ClaymoreMiner option ?

[Er0l]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

Wait a minute , ist rxboost only ClaymoreMiner option ?

No for sure not :-)

[big_daddy]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

what rxboost are you combining here? the rxboost is on Clay or what am I missing?

[Er0l]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

what rxboost are you combining here? the rxboost is on Clay or what am I missing?

No i use the rxboost tool for linux from simplemining os, best you switch  to simpleminingOS run your own straps and run with phoenixminer... for rx400/500

[Eliovp]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

what rxboost are you combining here? the rxboost is on Clay or what am I missing?

No i use the rxboost tool for linux from simplemining os, best you switch  to simpleminingOS run your own straps and run with phoenixminer... for rx400/500

You can use Amd Memory Tweak and use "REF" flag, it's the same thing, free to use

Cheers!

[gedeonodn]

Finally got 33 mh/s with RX470 - with rxboost

I have some rx470's Samsung, Elpida and Hynix, using rxboost my rigs are crashing
what driver are you using and are your cards bios modded or stock?

I use bios modded straps, not the freaky  on the fly, looks like some of the gpus already bricked with this when i go back some pages here.
Elpida and hynix the same, 32.7 to 32.9mh/s with rxboost and phoenixminer.

what rxboost are you combining here? the rxboost is on Clay or what am I missing?

No i use the rxboost tool for linux from simplemining os, best you switch  to simpleminingOS run your own straps and run with phoenixminer... for rx400/500

You can use Amd Memory Tweak and use "REF" flag, it's the same thing, free to use

Cheers!

I wrote about the same few posts ago, people not read before ask

Would you share some infomation? What is correct interval of values for --ref, if you have some stats, what is most used values for GPUs, if possible by models?

[Bojcha]

Latest one, i personally use 19.3.2 with compute mode switcher and atikmdag patcher.

are your cards stock or modded bios and what are your results?

Yes all cards uses modded bios since my straps still kinda better then straps in miner. So i only use rxboost on top.
btw if you change rxboost number or GPU clock its good to always pre-tune dcri at asm2,

[Smokeyw]

still no big boost for amd vii ??

[rigspec]

Here are some results from another farm we manage and operate running on Windows.

Specs:

OS: Windows
Claymore Version: 14.5

Farm (MH/s) BEFORE Update: 11,942.107 MH/s
Farm (MH/s) AFTER Update: 12,034.358 MH/s

Change MH/s: +92.251 Mh/s (Overall Farm)
Change %: +0.772% (Overall Farm)

Upon updating to v14.5 we noticed that running the miner alone WITHOUT the -rxboost command gave us the best results on this particular farm running Windows. Running either configuration of the following commands, -rxboost 1; -rxboost 30; -rxboost 50; -rxboost 75; -rxboost 100 - actually gave us 1-2 Mh/s decrease hashrate (Overall in Rig; ie: 356Mh/s (NO rxboost) --> 354 Mh/s Using -rxboost + (various parameters mentioned))

We would like to know the what other miners have experienced using this rxboost feature, and if they had similar results on WINDOWS.

Disclosure: ALL GPU's on this farm are RX570/580 (4GB+8GB), and have been individually tuned by RigSpec with custom vBios straps and memory+core timings have been fine tuned, so they were already maxed out prior to the v14.5 update for the most part; hence why the increases are not as great as results posted previously on the SMOS farm we manage.

Detailed Chart - Individual Rig's:

https://ibb.co/sWXkSDs


https://www.RigSpec.io

[laik2]

Here are some results from another farm we manage and operate running on Windows.

Specs:

OS: Windows
Claymore Version: 14.5

Farm (MH/s) BEFORE Update: 11,942.107 MH/s
Farm (MH/s) AFTER Update: 12,034.358 MH/s

Change MH/s: +92.251 Mh/s (Overall Farm)
Change %: +0.772% (Overall Farm)

Upon updating to v14.5 we noticed that running the miner alone WITHOUT the -rxboost command gave us the best results on this particular farm running Windows. Running either configuration of the following commands, -rxboost 1; -rxboost 30; -rxboost 50; -rxboost 75; -rxboost 100 - actually gave us 1-2 Mh/s decrease hashrate (Overall in Rig; ie: 356Mh/s (NO rxboost) --> 354 Mh/s Using -rxboost + (various parameters mentioned))

We would like to know the what other miners have experienced using this rxboost feature, and if they had similar results on WINDOWS.

Disclosure: ALL GPU's on this farm are RX570/580 (4GB+8GB), and have been individually tuned by RigSpec with custom vBios straps and memory+core timings have been fine tuned, so they were already maxed out prior to the v14.5 update for the most part; hence why the increases are not as great as results posted previously on the SMOS farm we manage.

Detailed Chart - Individual Rig's:

https://ibb.co/sWXkSDs


https://www.RigSpec.io

Claymore already revealed what rxboost does, gave entire specification of "Memory refresh period". In order for "-rxboost" to work as what is meant for you need to keep core a bit higher than usual and clock latency tighter.

[rigspec]

Here are some results from another farm we manage and operate running on Windows.

Specs:

OS: Windows
Claymore Version: 14.5

Farm (MH/s) BEFORE Update: 11,942.107 MH/s
Farm (MH/s) AFTER Update: 12,034.358 MH/s

Change MH/s: +92.251 Mh/s (Overall Farm)
Change %: +0.772% (Overall Farm)

Upon updating to v14.5 we noticed that running the miner alone WITHOUT the -rxboost command gave us the best results on this particular farm running Windows. Running either configuration of the following commands, -rxboost 1; -rxboost 30; -rxboost 50; -rxboost 75; -rxboost 100 - actually gave us 1-2 Mh/s decrease hashrate (Overall in Rig; ie: 356Mh/s (NO rxboost) --> 354 Mh/s Using -rxboost + (various parameters mentioned))

We would like to know the what other miners have experienced using this rxboost feature, and if they had similar results on WINDOWS.

Disclosure: ALL GPU's on this farm are RX570/580 (4GB+8GB), and have been individually tuned by RigSpec with custom vBios straps and memory+core timings have been fine tuned, so they were already maxed out prior to the v14.5 update for the most part; hence why the increases are not as great as results posted previously on the SMOS farm we manage.

Detailed Chart - Individual Rig's:

https://ibb.co/sWXkSDs


https://www.RigSpec.io

Claymore already revealed what rxboost does, gave entire specification of "Memory refresh period". In order for "-rxboost" to work as what is meant for you need to keep core a bit higher than usual and clock latency tighter.

We have our cores maxed out on all rigs at 1100-1150 (any further crashes rigs) ... the rigs in this particular farm have been fine tuned to the max possible settings via multiple trial test. Still no changes with rxboost, but v14.5 is definitely more efficient than prior versions, and as you can see has brought higher hashrates overall by simply updating

[laik2]

We have our cores maxed out on all rigs at 1100-1150 (any further crashes rigs) ... the rigs in this particular farm have been fine tuned to the max possible settings via multiple trial test. Still no changes with rxboost, but v14.5 is definitely more efficient than prior versions, and as you can see has brought higher hashrates overall by simply updating

I don't deny your results. What I am saying is a bit more technical probably for you to understand.
To provide more understandable meaning:
Elpida[EDW4032BAB], SK Hynix[H5GC4H24AJ/H5GC4H24AJR] will get the most of -rxboost, Micron[MT51J256M3],Samsung[K4G80325FB] and least SK Hynix[H5GQ8H24MJ/H5GQ8H24MJR].
Only exception might be Samsung[K4G41325FE] which is somewhere in the middle between SK Hynix[H5GC4H24AJ/H5GC4H24AJR] and Samsung[K4G80325FB].

[rigspec]

We have our cores maxed out on all rigs at 1100-1150 (any further crashes rigs) ... the rigs in this particular farm have been fine tuned to the max possible settings via multiple trial test. Still no changes with rxboost, but v14.5 is definitely more efficient than prior versions, and as you can see has brought higher hashrates overall by simply updating

I don't deny your results. What I am saying is a bit more technical probably for you to understand.
To provide more understandable meaning:
Elpida[EDW4032BAB], SK Hynix[H5GC4H24AJ/H5GC4H24AJR] will get the most of -rxboost, Micron[MT51J256M3],Samsung[K4G80325FB] and least SK Hynix[H5GQ8H24MJ/H5GQ8H24MJR].
Only exception might be Samsung[K4G41325FE] which is somewhere in the middle between SK Hynix[H5GC4H24AJ/H5GC4H24AJR] and Samsung[K4G80325FB].

Thanks for clarifying.... we actually built this farm in a way that all rigs had GPUs of similar memory manufacturers together (ie: Samsung, Micron, Elpida, Hynix).... we just need to figure out now which one gets the best performance with the rxboost with our tests and configurations