Depends on the cards. On my FEs and 480s/580s (8GB), 18.3.4 was fastest for Bittube.
470\480\570\580 8gb car work fine on 18.3.4 and 1.6.7
1.6.8 or new drivers low rate
for heavy\bittube algo
Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 119. (Read 237247 times)
October 24, 2018, 04:12:21 PM
470\480\570\580 8gb car work fine on 18.3.4 and 1.6.7
1.6.8 or new drivers low rate
for heavy\bittube algo
October 24, 2018, 04:07:04 PM
default intensity is to high for
Initializing OpenCL devices, please wait..
GPU0: AMD Radeon (TM) R7 370 Series [pitcairn] [4096 MB][Intensity 56.0][W: 8][T: 2][K: 2][BUS: 1]
GPU1: AMD Radeon (TM) R7 370 Series [pitcairn] [4096 MB][Intensity 56.0][W: 8][T: 2][K: 2][BUS: 3]
it will fail , 55 works
Initializing OpenCL devices, please wait..
GPU0: AMD Radeon (TM) R7 370 Series [pitcairn] [4096 MB][Intensity 56.0][W: 8][T: 2][K: 2][BUS: 1]
GPU1: AMD Radeon (TM) R7 370 Series [pitcairn] [4096 MB][Intensity 56.0][W: 8][T: 2][K: 2][BUS: 3]
it will fail , 55 works
October 24, 2018, 04:05:10 PM
Depends on the cards. On my FEs and 480s/580s (8GB), 18.3.4 was fastest for Bittube.
October 24, 2018, 03:58:28 PM
Someone would tell me how much changes the speed between algorithms Heavy vs haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
Heavy and Haven are practically identical in terms of speed from my experience. CN Saber (Bittubev2) can vary depending on the driver used.
what driver version need for Bittubev2 ?
October 24, 2018, 03:55:35 PM
Someone would tell me how much changes the speed between algorithms Heavy vs haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
Heavy and Haven are practically identical in terms of speed from my experience. CN Saber (Bittubev2) can vary depending on the driver used.
October 24, 2018, 03:22:48 PM
Someone would tell me how much changes the speed between algorithms Heavy vs haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
For example:
RX 580 4 GB - Heavy 700 h/s and what would be the speed in haven?
October 24, 2018, 03:15:45 PM
How many h/s do you get with your vega56 at v8?
And what is the power use?
I biosmod my vega56 to the 64 and have 1955h/s per card with v8 x4 at 840watt.
Thanks for your reply.
And what is the power use?
I biosmod my vega56 to the 64 and have 1955h/s per card with v8 x4 at 840watt.
Thanks for your reply.
October 24, 2018, 02:53:12 PM
How did you guys get the 18.6.1 drivers installed on your Vega FE? I have the 18.Q3.1 Pro driver installed but I can only install 18.4.1 or 18.7.1 and 18.8.1. There are no other options for the driver switch.
Here is link to the official site https://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Adrenalin-Edition-18.6.1-Release-Notes.aspx
just choose version for your OS and download. Then make clean installation.
I have the drivers thanks, but when I install them they default to the Pro drivers and you can't use overclocking/underclocking with them.
Why not just force the overclock/undervolt through PPT table and call it a day?
October 24, 2018, 02:02:44 PM
How did you guys get the 18.6.1 drivers installed on your Vega FE? I have the 18.Q3.1 Pro driver installed but I can only install 18.4.1 or 18.7.1 and 18.8.1. There are no other options for the driver switch.
Here is link to the official site https://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Adrenalin-Edition-18.6.1-Release-Notes.aspx
just choose version for your OS and download. Then make clean installation.
I have the drivers thanks, but when I install them they default to the Pro drivers and you can't use overclocking/underclocking with them.
October 24, 2018, 01:06:08 PM
How did you guys get the 18.6.1 drivers installed on your Vega FE? I have the 18.Q3.1 Pro driver installed but I can only install 18.4.1 or 18.7.1 and 18.8.1. There are no other options for the driver switch.
Here is link to the official site https://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Adrenalin-Edition-18.6.1-Release-Notes.aspx
just choose version for your OS and download. Then make clean installation.
October 24, 2018, 12:33:47 PM
How did you guys get the 18.6.1 drivers installed on your Vega FE? I have the 18.Q3.1 Pro driver installed but I can only install 18.4.1 or 18.7.1 and 18.8.1. There are no other options for the driver switch.
October 24, 2018, 11:17:20 AM
hello, who can help with my problem, i have lower hashrate on my vega's 64, even i cant reach more than 1900...
i'm using 18.6.1 driver with atikmdagpatcher applied, 1.6.8 srb miner, computemode on, all settings with overdrive tools with average common settings to 64 vegas.
and still i can see this hashrates on 2 my ferms:
https://i96.fastpic.ru/big/2018/1024/16/cef64ead1b2db259d8adec5d98e9fa16.jpg
https://i96.fastpic.ru/big/2018/1024/58/003b33adf3f9ad30b606cb3644ee9d58.jpg
what i'm doing wrong??
i'm using 18.6.1 driver with atikmdagpatcher applied, 1.6.8 srb miner, computemode on, all settings with overdrive tools with average common settings to 64 vegas.
and still i can see this hashrates on 2 my ferms:
https://i96.fastpic.ru/big/2018/1024/16/cef64ead1b2db259d8adec5d98e9fa16.jpg
https://i96.fastpic.ru/big/2018/1024/58/003b33adf3f9ad30b606cb3644ee9d58.jpg
what i'm doing wrong??
October 24, 2018, 11:15:05 AM
@doktor83
I just want to verify is this a bug in SRBMIner V1.6.8 or not? After 3 days running, my rig suddenly stopped working. And I saw this on the screen:
The messages stated:
Critical shutdown temperature reached (80C)!!! (I did set "shutdown_temperature" : 80 in my config files)
Shutting down system!
However, if you look at the screen status, none of my GPU reach over 55C
Thank you!
I just want to verify is this a bug in SRBMIner V1.6.8 or not? After 3 days running, my rig suddenly stopped working. And I saw this on the screen:
The messages stated:
Critical shutdown temperature reached (80C)!!! (I did set "shutdown_temperature" : 80 in my config files)
Shutting down system!
However, if you look at the screen status, none of my GPU reach over 55C
Thank you!
If it is because of a bug then it is something in the driver, in ADL part, because miner gets the temperature data from it, and then just makes a simple compare (if adl value > user set value = problem).
This gives me an idea, i will add to the log the value read from ADL if this happens.
October 24, 2018, 09:46:15 AM
Force Compute Mode + Disable Crossfire & ULPS in one click [ just copy into note pad and save as a *.reg file, make sure to adjust reg locations for gpu's, i use this on all the rigs not just vega's]
Code:
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0001]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0002]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0003]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0004]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0005]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0006]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0007]
"KMD_EnableInternalLargePage"=dword:00000002
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0001]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0002]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0003]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0004]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0005]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0006]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0007]
"EnableCrossFireAutoLink"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0001]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0002]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0003]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0004]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0005]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0006]
"EnableUlps"=dword:00000000
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0007]
"EnableUlps"=dword:00000000
The --setcomputemode also sets these 3 parameters per gpu, so no need to edit/run a registry file
Damn brotha! you so thoughtful, thank you!
October 24, 2018, 09:33:07 AM
I think you'd have to provide a manual value. So yeah, it was an idea. Doesn't have to be implemented.
At worst it reminded people that profitability is not merely about income. If you make $10 out of mining XMR and 9$ out of mining XHV, the latter choice would still highly likely net you more as that $1 is more than offset by the 40% higher power cost.
At worst it reminded people that profitability is not merely about income. If you make $10 out of mining XMR and 9$ out of mining XHV, the latter choice would still highly likely net you more as that $1 is more than offset by the 40% higher power cost.
October 24, 2018, 09:30:13 AM
Dok, one suggestion for the next release:
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
If the power measurement works as it's in HWInfo64 then there would be a lot of error. I have cards showing 40w and 80w as chip power while actual numbers would be 100w+ for both.
Increasing worksize from 16 to 32 reduces power consumption but also reduces hashrates. However the efficiency numbers might increase especially for polaris cards.
October 24, 2018, 09:24:01 AM
Dok, one suggestion for the next release:
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
I think you're on the right track, but IMO this isn't work Dok should be doing. Everyone uses different PPTs and is targeting different things. Everyone more than likely has different $/kWh parameters. you've already measured power pull; I'd suggest you simply input different hashrates into the algo file to accommodate power pull. This could be a scale factor, such as simply dividing hashrate by 1.2 or so, even though honestly it's more complicated than that, because people with cheap power won't mind the extra power very much on a profitable algo.
October 24, 2018, 09:11:49 AM
I tried a couple of pools and the payment is 20-30% less than the 2 calculators show (whattomine and https://monerotools.github.io/calculator/). I am yet to try nanopool - is it better than the others?
Yeah same here payments are low, my pool hashrates match what i see on srb tho October 24, 2018, 09:10:46 AM
Dok, one suggestion for the next release:
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
Incorporate wattage per algo in the algo file, not just hashrate. Switching pools like MoneroOcean will not pay attention to the power usage of said algo in their profitability calculations. It only looks at the hashrate you've specified per algo and then sets you on what gives you more $.
However, if for example an algo like CNv2 draws 40% more power than CN-Heavy (and it does), this should be included in the profitability calcs. So currently one has two choices:
1/ ignore it an let MO mine whatever makes the highest turnover
2/ tweak your estimated hashrates per algo to approximate real earnings
A little system that would allow you to specify rig power draw per algo would be nice to have so it would automatically adjust profitability and provide weighted hashrate estimations to MoneroOcean.
Makes sense?
October 24, 2018, 08:22:38 AM
Dang this v8 is a completely different animal from v7, keep having to make adjustments. It stresses the cards more than I thought, actually had to add 5 mV to mem voltage. Wondering why my cards were underperforming with pool, realized it's bad to roll with variable difficulty on XMR. Static difficulty tuned to find a share every ~45 sec. is definitely better.
Jump to: