Topic: [Cast XMR] high speed XMR/CryptoNight miner for RX Vega GPUs (2 KHash/s) - page 59. (Read 206425 times)
April 02, 2018, 01:06:44 PM
Is 0.9.2 ready for the Sumo fork happening in the next few days?
April 02, 2018, 10:02:08 AM
Yeah... so I got a little too excited too soon. Definitely have my overclocks well within previous working ranges, and yet after mining for a while cards will randomly start hashing ~100-200 H/s. Very strange... I can exit the miner without it hanging and having to restart, but still, can't be around to monitor the machines.
Will probably pull out the old JJHashRateMonitor modified for CAST and see if that can take care of it, since --ratewatchdog can't fix this problem. We're getting close though!
Will probably pull out the old JJHashRateMonitor modified for CAST and see if that can take care of it, since --ratewatchdog can't fix this problem. We're getting close though!
I had a similar experience. 2 RX Vega 64s on 18.3.4 driver, PCIE set to GEN1, power play tables etc. Rates are good (maybe about 2% better than with cast 0.9.0) and temps are maybe 1 deg C lower. After a few hours though both cards dropped down to ~ 400-500 H/s. I saw there were kernel resets in the miner, but they were unable to restore the full speed (~3900 H/s).
Was able to exit gracefully and restart. Now running without the --ratewatchdog option. My suspicion is that the kernel resets are themselves responsible for the strange rates - these are hashrates (~400-500 H/s per card) I've never had in the past. It could be in the past the miner would hang/system would crash prompted by whatever prompted the initial fall and kernel resets however. More to be learned. Regardless, the implementation of (an eventually working native) option like ratewatchdog is sorely needed, so this initial development is applauded! Hope it continues...
I removed the -ratewatchdog and everything went back to normal.
I made a couple of tests: with and without. Not using the -ratewatchdog is definitively working better for me.
One thing to note is that the developer stated that the -ratewatchdog was experimental.
April 02, 2018, 06:34:22 AM
this support for SSL port?
April 02, 2018, 05:05:21 AM
Yeah... so I got a little too excited too soon. Definitely have my overclocks well within previous working ranges, and yet after mining for a while cards will randomly start hashing ~100-200 H/s. Very strange... I can exit the miner without it hanging and having to restart, but still, can't be around to monitor the machines.
Will probably pull out the old JJHashRateMonitor modified for CAST and see if that can take care of it, since --ratewatchdog can't fix this problem. We're getting close though!
Will probably pull out the old JJHashRateMonitor modified for CAST and see if that can take care of it, since --ratewatchdog can't fix this problem. We're getting close though!
I had a similar experience. 2 RX Vega 64s on 18.3.4 driver, PCIE set to GEN1, power play tables etc. Rates are good (maybe about 2% better than with cast 0.9.0) and temps are maybe 1 deg C lower. After a few hours though both cards dropped down to ~ 400-500 H/s. I saw there were kernel resets in the miner, but they were unable to restore the full speed (~3900 H/s).
Was able to exit gracefully and restart. Now running without the --ratewatchdog option. My suspicion is that the kernel resets are themselves responsible for the strange rates - these are hashrates (~400-500 H/s per card) I've never had in the past. It could be in the past the miner would hang/system would crash prompted by whatever prompted the initial fall and kernel resets however. More to be learned. Regardless, the implementation of (an eventually working native) option like ratewatchdog is sorely needed, so this initial development is applauded! Hope it continues...
April 01, 2018, 11:10:39 PM
should m using dummy to IGPU? when m mining with rx 570 ? like vega rig...
^ Set output display to IGFX, don't need dongle. April 01, 2018, 10:42:19 PM
should m using dummy to IGPU? when m mining with rx 570 ? like vega rig...
^ Set output display to IGFX, don't need dongle. April 01, 2018, 09:31:10 PM
should m using dummy to IGPU? when m mining with rx 570 ? like vega rig...
March 31, 2018, 09:08:41 PM
When do we have new update for April 6th fork?
Read page 1
Already updated!!
I think no..
March 31, 2018, 09:04:11 PM
When do we have new update for April 6th fork?
Read page 1
Already updated!!
March 31, 2018, 09:01:02 PM
When do we have new update for April 6th fork?
Read page 1
Already updated!!
March 31, 2018, 08:59:49 PM
When do we have new update for April 6th fork?
March 31, 2018, 08:00:07 PM
i wish you could include RX550 in you next update . i bought one to fill my pcie and it doesnt work with cast so i had to switch to stak but my vegas give me less hasrate ... please could you ? thanks
March 31, 2018, 05:08:49 PM
When I run cast xmr it just closes, how to fix this?
I have 6x vega 56, win10 64bit. Also in amd radeon setting cards 3 and 6 won't save settings on compute mode, how to solve that?
I have 6x vega 56, win10 64bit. Also in amd radeon setting cards 3 and 6 won't save settings on compute mode, how to solve that?
you most likely have an error in your config. If you're running a bat file then add "pause" at end of the script that should keep the cmd window open and you may see the error message.
March 31, 2018, 02:06:58 PM
First off, thanks so much glph3k! I've been using your miner since the beginning, it's come a long way since then!
Like many vega owners/miners, I've been waiting for the day when using the August BC drivers is no longer necessary. Seems that day is here!
I'm getting full speeds on all 56/64 cards with your updated miner and the new 18.3.4 Adrenalin drivers. No more disable/enable needed!
I'm also hoping to get my Vega FEs switched over soon as well, but Overdriven Tool does not support Vega FE yet on the new drivers.
I will say that I did notice the same issue that RAJNIFICENT mentioned on several of my rigs. I've reduced some clocks and added some voltage, fingers crossed that that solves it.
Also wanted to say I'm liking your --ratewatchdog argument. From what I've seen, it picks up the slight drops in hashrate pretty well. Vega mining is getting close to plug and play!
Like many vega owners/miners, I've been waiting for the day when using the August BC drivers is no longer necessary. Seems that day is here!
I'm getting full speeds on all 56/64 cards with your updated miner and the new 18.3.4 Adrenalin drivers. No more disable/enable needed!
I'm also hoping to get my Vega FEs switched over soon as well, but Overdriven Tool does not support Vega FE yet on the new drivers.
I will say that I did notice the same issue that RAJNIFICENT mentioned on several of my rigs. I've reduced some clocks and added some voltage, fingers crossed that that solves it.
Also wanted to say I'm liking your --ratewatchdog argument. From what I've seen, it picks up the slight drops in hashrate pretty well. Vega mining is getting close to plug and play!
Hi, please share your success / failures. I'm trying to reduce my overclock, however my hashrates on one rig continues to drop. If I can get rate watch to work as designed, dude my life would be complete.
Hi,
Looks like I have same issue on 1st day mining.
1 - Vega FE
4 - Vega 64
Hicehash
567 - Accepted (75%)
187 - Rejected by pool (25%)
I have never rejected by pool until this version. Same OC as on previous versions.
Not sure why you're getting the rejects... particularly if you never got them before, I'm not having this issue. Are your cards all Samsung memory? Only time I've run into rejects is when I overclock Hynix mem too high. Only other instance I could foresee is if you're not using the right algo on some shares... Why don't you try forcing --algo 0 in your bat file and report back?
March 31, 2018, 01:47:13 PM
When I run cast xmr it just closes, how to fix this?
I have 6x vega 56, win10 64bit. Also in amd radeon setting cards 3 and 6 won't save settings on compute mode, how to solve that?
I have 6x vega 56, win10 64bit. Also in amd radeon setting cards 3 and 6 won't save settings on compute mode, how to solve that?
March 31, 2018, 08:06:42 AM
It's not working quite right for me 
* Used DUU to uninstall drivers, and installed new Adrenaline drives.
* Disabled Crossfire and set Profiles to custom and custom voltage to on.
* Wrote Soft PP tables to registry.
* Restarted.
So everything works great at the start - 2000H/s and lower temperatures, same voltages and power consumption.
But after about 20min it just stops working entirely, if I press q to quit it just hangs, and I have to restart the system.
Back on blockchain drivers
Check your mother board Bios setting of the PCIE slots. They should be set to GEN1 or GEN2. Normally the MB are delivered with AUTO settings.* Used DUU to uninstall drivers, and installed new Adrenaline drives.
* Disabled Crossfire and set Profiles to custom and custom voltage to on.
* Wrote Soft PP tables to registry.
* Restarted.
So everything works great at the start - 2000H/s and lower temperatures, same voltages and power consumption.
But after about 20min it just stops working entirely, if I press q to quit it just hangs, and I have to restart the system.
Back on blockchain drivers
Normally switching to this settings take care of the "after 20 minutes" stops.
I had them as well.
Great thanks trying this now, will report back if it work! Thanks for the help
Dont' want to jinx it, but seems to be working great now!! I'm getting similar hashrates and noticably lower temps (>5 deg cooler), though it seems that the drivers are clever since it's holding the gpu speed at 1380Hz (I have default to 1407Hz) probably due to the voltage limitations, so probably get a few degrees drop from that. Hash rates are similar if not marginally faster. Will test playing around with the voltages.
March 31, 2018, 07:24:07 AM
Guys, any different Watt consume Aug23 Blockchain driver and Adrenalin 18.3.4 driver ? 
i don;t have kill-a-watt, sad.
i don;t have kill-a-watt, sad. March 31, 2018, 07:18:36 AM
It's not working quite right for me
* Used DUU to uninstall drivers, and installed new Adrenaline drives.
* Disabled Crossfire and set Profiles to custom and custom voltage to on.
* Wrote Soft PP tables to registry.
* Restarted.
So everything works great at the start - 2000H/s and lower temperatures, same voltages and power consumption.
But after about 20min it just stops working entirely, if I press q to quit it just hangs, and I have to restart the system.
Back on blockchain drivers
Check your mother board Bios setting of the PCIE slots. They should be set to GEN1 or GEN2. Normally the MB are delivered with AUTO settings.* Used DUU to uninstall drivers, and installed new Adrenaline drives.
* Disabled Crossfire and set Profiles to custom and custom voltage to on.
* Wrote Soft PP tables to registry.
* Restarted.
So everything works great at the start - 2000H/s and lower temperatures, same voltages and power consumption.
But after about 20min it just stops working entirely, if I press q to quit it just hangs, and I have to restart the system.
Back on blockchain drivers
Normally switching to this settings take care of the "after 20 minutes" stops.
I had them as well.
Great thanks trying this now, will report back if it work! Thanks for the help
March 31, 2018, 06:59:30 AM
First off, thanks so much glph3k! I've been using your miner since the beginning, it's come a long way since then!
Like many vega owners/miners, I've been waiting for the day when using the August BC drivers is no longer necessary. Seems that day is here!
I'm getting full speeds on all 56/64 cards with your updated miner and the new 18.3.4 Adrenalin drivers. No more disable/enable needed!
I'm also hoping to get my Vega FEs switched over soon as well, but Overdriven Tool does not support Vega FE yet on the new drivers.
I will say that I did notice the same issue that RAJNIFICENT mentioned on several of my rigs. I've reduced some clocks and added some voltage, fingers crossed that that solves it.
Also wanted to say I'm liking your --ratewatchdog argument. From what I've seen, it picks up the slight drops in hashrate pretty well. Vega mining is getting close to plug and play!
Like many vega owners/miners, I've been waiting for the day when using the August BC drivers is no longer necessary. Seems that day is here!
I'm getting full speeds on all 56/64 cards with your updated miner and the new 18.3.4 Adrenalin drivers. No more disable/enable needed!
I'm also hoping to get my Vega FEs switched over soon as well, but Overdriven Tool does not support Vega FE yet on the new drivers.
I will say that I did notice the same issue that RAJNIFICENT mentioned on several of my rigs. I've reduced some clocks and added some voltage, fingers crossed that that solves it.
Also wanted to say I'm liking your --ratewatchdog argument. From what I've seen, it picks up the slight drops in hashrate pretty well. Vega mining is getting close to plug and play!
Hi, please share your success / failures. I'm trying to reduce my overclock, however my hashrates on one rig continues to drop. If I can get rate watch to work as designed, dude my life would be complete.
Hi,
Looks like I have same issue on 1st day mining.
1 - Vega FE
4 - Vega 64
Hicehash
567 - Accepted (75%)
187 - Rejected by pool (25%)
I have never rejected by pool until this version. Same OC as on previous versions.
March 31, 2018, 04:26:13 AM
Thanks for your efforts and information!
Will Cast-xmr (Vega) also support Sumokoins "cryptonight heavy" fork?
BR gar
Will Cast-xmr (Vega) also support Sumokoins "cryptonight heavy" fork?
BR gar
Jump to: