[Cast XMR] high speed XMR/CryptoNight miner for RX Vega GPUs (2 KHash/s)

[Marvell2]

My vega Fes start out high at 2100 with -I 12 on CN heavy

but after a few mins they are down to 1900-1800

my ref 56 flashed to 64 all do 1230 max

they start at 1500 too with -I 10

THings are even lower with default intensities

Any ideas ?

I have power play settings etc, large pages and all set up

Are you using --ratewatchdog option? If yes, delete it from .bat file.

my fans are set to 3000 min , is that too low ?

what should i set the min fan speed for the FEs ?  the temps on cast say 60-70 degrees but I imagine that is just core.

No I don't have --ratewatchdog set

[1714103682]

1714103682

[1714103682]

1714103682

[1714103682]

1714103682

[1714103682]

1714103682

[1714103682]

1714103682

[Bazzaar]

My vega Fes start out high at 2100 with -I 12 on CN heavy

but after a few mins they are down to 1900-1800

my ref 56 flashed to 64 all do 1230 max

they start at 1500 too with -I 10

THings are even lower with default intensities

I think thats an indication that intensity is too high, try stepping down 1 each time and see if you get a setting that gives you nearly what you are seeing max, but doesnt fall off with time.
With my Vega56s I get best hash with I 8. 9 or 10 looks better to start with but falls away, 8 doesn't.

Baz

[Marvell2]

My vega Fes start out high at 2100 with -I 12 on CN heavy

but after a few mins they are down to 1900-1800

my ref 56 flashed to 64 all do 1230 max

they start at 1500 too with -I 10

THings are even lower with default intensities

I think thats an indication that intensity is too high, try stepping down 1 each time and see if you get a setting that gives you nearly what you are seeing max, but doesnt fall off with time.
With my Vega56s I get best hash with I 8. 9 or 10 looks better to start with but falls away, 8 doesn't.

Baz

hhm ill give that a shot thanks

[CryptoPeter]

This program keeps freezing on my vega rigs, anyone else have this problem or have a fix? on v100 and v110

same here, miner just stops hashing for me but if i come back and hit enter/space or whatever itll spit out that it has old shares or w/e and shit and then get back to mining current shares or whatever

https[Suspicious link removed]

seems it started acting up again and stops hashing

any other suggestions?

never mind looks like my risers were falling off somehow, i flipped that switch to keep them locked in place now so that should be why miner was crashing on me lol

[treanski]

My vega Fes start out high at 2100 with -I 12 on CN heavy

but after a few mins they are down to 1900-1800

my ref 56 flashed to 64 all do 1230 max

they start at 1500 too with -I 10

THings are even lower with default intensities

Any ideas ?

I have power play settings etc, large pages and all set up

Are you using --ratewatchdog option? If yes, delete it from .bat file.

my fans are set to 3000 min , is that too low ?

what should i set the min fan speed for the FEs ?  the temps on cast say 60-70 degrees but I imagine that is just core.

No I don't have --ratewatchdog set

yes its just the core, you can expect memory is 5-10°C hotter & to hot memory can cause hashrate drops or bad shares on vega

[Jo82]

Just tried the latest version to mine Haven V3 on my 3 Vega 56 rig.
2 cards hash 1300h/s the one with the monitor plugged in only does 300h/s.
Already tried lowering intesity on the card...
who has a solution for this?

thnx for the advice.

[bumbumzz]

Hello,

How to set fail-over server? Like the second stratum when the first one is down. Try to find an example but cannot find anything.

[Marvell2]

My vega Fes start out high at 2100 with -I 12 on CN heavy

but after a few mins they are down to 1900-1800

my ref 56 flashed to 64 all do 1230 max

they start at 1500 too with -I 10

THings are even lower with default intensities

Any ideas ?

I have power play settings etc, large pages and all set up

Are you using --ratewatchdog option? If yes, delete it from .bat file.

my fans are set to 3000 min , is that too low ?

what should i set the min fan speed for the FEs ?  the temps on cast say 60-70 degrees but I imagine that is just core.

No I don't have --ratewatchdog set

There are multiple possible reasons for hash drop so it's just something to try but once I set my Vega 56 fans to run at 4900 at all times, I stopped getting hash drops. Lowering voltages also helps reduce heat but of course you can't go too far down or else it'll crash the system

the fans on the fes run high, at like 4800 so thats not the. issue, i
can raise the fans on the 56s , this weeken ill try a new lower voltage powerplau for the cards
for some reason i cant change any settings on the powercolor red devil 56 cards and the vega fes
in overdriven tool, but i can change the settings with a powerplay reg mod using the spread sheet

[CryptoPeter]

This program keeps freezing on my vega rigs, anyone else have this problem or have a fix? on v100 and v110

same here, miner just stops hashing for me but if i come back and hit enter/space or whatever itll spit out that it has old shares or w/e and shit and then get back to mining current shares or whatever

https[Suspicious link removed]

well miner still gets stuck for me and stops hashing ;(

[zeef]

This program keeps freezing on my vega rigs, anyone else have this problem or have a fix? on v100 and v110

same here, miner just stops hashing for me but if i come back and hit enter/space or whatever itll spit out that it has old shares or w/e and shit and then get back to mining current shares or whatever

https[Suspicious link removed]

well miner still gets stuck for me and stops hashing ;(

Same here, and i have de edition mode off in cmd options..

[nemzy]

need HELP
after re-instal windows and blockchain drivers miner stacking on Initializing GPU , loading kernel ...
Crossfire - OFF
Virtual - 50000MB
             64000MB
was working perfect before windows update
with adrenaline drivers it's restarting all the time. I've done all from the beginning - re-instal drivers with DDU, disable Crossfire - same story

[BogdanCo]

need HELP
after re-instal windows and blockchain drivers miner stacking on Initializing GPU , loading kernel ...
Crossfire - OFF
Virtual - 50000MB
             64000MB
was working perfect before windows update
with adrenaline drivers it's restarting all the time. I've done all from the beginning - re-instal drivers with DDU, disable Crossfire - same story

Try this: https://bitcointalk.org/index.php?topic=1433925.msg39173647#msg39173647

Don't forget to disable ULPS for ALL cards installed on your rig.

[nemzy]

need HELP
after re-instal windows and blockchain drivers miner stacking on Initializing GPU , loading kernel ...
Crossfire - OFF
Virtual - 50000MB
             64000MB
was working perfect before windows update
with adrenaline drivers it's restarting all the time. I've done all from the beginning - re-instal drivers with DDU, disable Crossfire - same story

Try this: https://bitcointalk.org/index.php?topic=1433925.msg39173647#msg39173647

Don't forget to disable ULPS for ALL cards installed on your rig.

Done this before , same thing ... cards are Vegas 56 with Bios from 64 and 1-64

[huntingthesnark]

Just tried the latest version to mine Haven V3 on my 3 Vega 56 rig.
2 cards hash 1300h/s the one with the monitor plugged in only does 300h/s.
Already tried lowering intesity on the card...
who has a solution for this?

thnx for the advice.

Yeah, me too. Had to start another instance of Cast, back to the bad old days. Have two rigs that just won't deal with the newest version at all - new intensity settings just fry everything, crashes. Previous version is fine... (Know you can change intensity, but if I back it off hashrate plummets. dial it up and crash. Not a great combo)

[MinermanNC]

Hey gang, just jumping in here with a question  

BT250 with 4 x RX 480 8/GB cards.

Just started using Cast about 4 hours ago.  Its hashing fine, but I am using 18.6.1 latest driver atm. ( IK that may not be best)

I'm getting avg 635 hs with stock intensity etc, But when I enable "Compute mode" (Radeon Settings) and run the software again, my hash rate doubles to around 1280 per card h's total all 4 cards 5,110 h's

But, its also throwing lines and lines of an error message something about calling memory CL etc... I should have wrote it down, but it does still hash at the extremely high doubled hash rate lol

Ok, whats the trick I may be missing here. I am not done tweaking and testing but a quick fix for enabling Compute would help THX    Running out of time to mess with it anymore tonight  

EDIT: this is the error. I see others in the thread have had this problem. will try some of the things they checked. And yes I do get 1280 Hs per card when its throwing this error but its hashing lol

Code:

[13:51:27] Error CL_MEM_OBJECT_ALLOCATION_FAILURE calling clEnqueueNDRangeKernel for kernel 0.

[SpceGhst]

Hey gang, just jumping in here with a question  

BT250 with 4 x RX 480 8/GB cards.

Just started using Cast about 4 hours ago.  Its hashing fine, but I am using 18.6.1 latest driver atm. ( IK that may not be best)

I'm getting avg 635 hs with stock intensity etc, But when I enable "Compute mode" (Radeon Settings) and run the software again, my hash rate doubles to around 1280 per card h's total all 4 cards 5,110 h's

But, its also throwing lines and lines of an error message something about calling memory CL etc... I should have wrote it down, but it does still hash at the extremely high doubled hash rate lol

Ok, whats the trick I may be missing here. I am not done tweaking and testing but a quick fix for enabling Compute would help THX    Running out of time to mess with it anymore tonight  

Lower the intensity setting.  If cards are OC’d, lower mem clock by 25 and try with default intensity.

[MinermanNC]

Hey gang, just jumping in here with a question  

BT250 with 4 x RX 480 8/GB cards.

Just started using Cast about 4 hours ago.  Its hashing fine, but I am using 18.6.1 latest driver atm. ( IK that may not be best)

I'm getting avg 635 hs with stock intensity etc, But when I enable "Compute mode" (Radeon Settings) and run the software again, my hash rate doubles to around 1280 per card h's total all 4 cards 5,110 h's

But, its also throwing lines and lines of an error message something about calling memory CL etc... I should have wrote it down, but it does still hash at the extremely high doubled hash rate lol

Ok, whats the trick I may be missing here. I am not done tweaking and testing but a quick fix for enabling Compute would help THX    Running out of time to mess with it anymore tonight  

Lower the intensity setting.  If cards are OC’d, lower mem clock by 25 and try with default intensity.

Hey thanks, ill take a look at that in the morning  they were not OC'd however.

[Tarv]

Thanks for fixing, issue resolved in 1.2.0/0.9.5.

Also seems to have fixed FAN speed not showing up, which I honestly thought was just a problem with my drivers from day one (and kind of forgot about).

Thanks.

Hi,

thanks for the info. Seems like a solid bug, will hammer it out in the next version.

Regards,
glph3k

Hey, just noticing a UI error with switch-radeon-gpu.

Looks like it's mixing up the names.  4,5 are in fact the VEGA cards.  Numbered order matches CLINFO as well.

[MinermanNC]

Hey gang, just jumping in here with a question  

BT250 with 4 x RX 480 8/GB cards.

Just started using Cast about 4 hours ago.  Its hashing fine, but I am using 18.6.1 latest driver atm. ( IK that may not be best)

I'm getting avg 635 hs with stock intensity etc, But when I enable "Compute mode" (Radeon Settings) and run the software again, my hash rate doubles to around 1280 per card h's total all 4 cards 5,110 h's

But, its also throwing lines and lines of an error message something about calling memory CL etc... I should have wrote it down, but it does still hash at the extremely high doubled hash rate lol

Ok, whats the trick I may be missing here. I am not done tweaking and testing but a quick fix for enabling Compute would help THX    Running out of time to mess with it anymore tonight  

Lower the intensity setting.  If cards are OC’d, lower mem clock by 25 and try with default intensity.

Hey thanks, ill take a look at that in the morning   they were not OC'd however.

Ok, got it hashing fine with "Compute" on. Not sure how much this made the difference...but I increased virtual memory from 16000 to 64000. Turned on Compute and all seemed fine, but than within a minute I noticed only 3 of 4 cards were hashing then only 2 ..THEN I get a windows(10) message saying: Windows has blocked the application to the graphics cards...  eventually shutting it down.

So I went ahead and changed the properties of the cast_xmr-Vega exe in the mining folder to "run as administrator" and that seemed to resolve the windows blocking.

In the end I am getting and avg of 705 h's on each 480 now, up from 635 without the "compute". So I'm pretty happy with that. During my attempts to get right I even tried the blockchain drivers but that didn't go to well lol

PS: I am running 18.6.1 drivers again.... and mining V7

[Afshin]

Driver 18.5.2  Vs. 18.6.1

5 x VEGA 64
Win10, Claymore.CryptoNote.AMD.GPU.Miner.v11.3

9237 H/s   >    9040 H/s