 Show Posts
|
|
Pages: [1]
|
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.
|
|
|
|
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. Detected OpenCL Platform: OpenCL 2.1 AMD-APP (2580.6)
GPU0: Radeon RX Vega (gfx804) ... no changes.
GPU1: Radeon RX Vega (gfx804) ... no changes.
GPU2: Radeon RX 550 Series (gfx804) ... no changes.
GPU3: Radeon RX 550 Series (gfx804) ... no changes.
GPU4: Radeon RX 550 Series (gfx900) ... no changes.
GPU5: Radeon RX 550 Series (gfx900) ... no changes.
GPU0: Radeon RX Vega (gfx804)
GPU1: Radeon RX Vega (gfx804)
GPU2: Radeon RX 550 Series (gfx804)
GPU3: Radeon RX 550 Series (gfx804)
GPU4: Radeon RX 550 Series (gfx900) | HBCC: OFF | Large Pages: ON
GPU5: Radeon RX 550 Series (gfx900) | HBCC: OFF | Large Pages: ON |
|
|
|
|
Went from 0.8.1 to 0.8.5. I notice a very slight bump in hash rate around +50mh/s, and also a increase in temperature from of about +2°C.
I'm running these cards in low power, aiming for a hash around 1850mh/s, so I don't see any stability changes.
|
|
|
|
Hey all. I took a bit of a different approach to the hash drops and wrote a script to check and restart Cast_XMR as needed. Basically you set it up as a scheduled task and it will look at the current hash from Cast_XMR. If the hash rates are lower than what you set, it will kill the miner and restart it. I posted it on GitHub if anyone wants to have a look or try it out. It's written in VBS so it's pretty straight forward to step your way through the code and see what it's doing. https://github.com/tarvcode/Check-Cast_XMRLet me know if you have any comments or suggestions, I'm willing to spend some time tweaking this if anyone finds it useful. Best to PM me directly or comment on GitHub. |
|
|
|
|
Request: Add option to cast_xmr to set a timer to stop and restart the miner gracefully. Alternatively trigger a restart if hash rate is below a threshold for an amount of time.
I've read and seen the scripts to automate this, however would like to request this be added to the miner directly if possible. Like others, I'm able to get 1900+ h/s per VEGA, however it will drop to around 1550 h/s after some time and stays there. Lots of things cause this drop and I have found restarting the miner is necessary. On my setup, I never seem to have to reset (enable/disable) the cards, only restart the miner, to get back to full hash rate.
I am running multiple cards on this RIG. 2x VEGA's with Cast_XMR. 4x RX550's using XMR-STAK. 2x R9 290's and 4x GTX 1060's using Claymore. Also running XMRig for CPU mining.
I must start all the other miners BEFORE Cast_XMR to ensure full hash. Restarting another miner will cause Cast_XMR to drop the hash rate instantly.
In lieu of solving what seems to be a driver issue, would like the miner to handle this hash drop internally, either with a timer or a hash rate monitor.
Hope that all makes sense. Thanks for the work on the miner dev!
|
|
|
|
Greetings, I am having trouble with one card in a rig with 6 cards. Two R9 280 and four Gigabyte GTX 1060 3GB Mini cards. Issue is with one of the 1060's. Just added these 1060's yesterday, brand new. The three cards are mining fine with overclocks at around 19.5mh/s. The fourth card fails during mining. I have it at stock settings and it will mine at 17.5mh/s. After about 5 minutes, it will drop to 1.7mh/s. At this point the system will become unresponsive and MSI afterburner will freeze. If I wait, the card will eventually go back to 17.5mh/s and then again drop to 1.7mh/s If I remove the problem card from the startup.bat, the rig is fine. I have tried replacing the riser and cable, no difference. Any suggestions? Is this card shot? EthDcrMiner64.exe -epool eth-us-east1.nanopool.org:9999 -ewal XXX -epsw X -mode 1 -ftime 10 -ethi 8,8,8,8,8,8 -dcri 30,30,20,20,20,20 -allpools 1 -gser 2 -di 036897 Problem card is #7 Underclocking -400/-500 didn't help either. EDIT: Narrowed it down that the issue was happening with any card plugged into a certain PCIE slot on the motherboard. Moved the connections to not use that PCIE slot and everything seems to be working now. Using the ASRock H110 Pro BTC+ motherboard and the problem slot was the second one from the top labelled PCIE1. |
|
|
|
Maybe I need older nVidia drivers and windows 10 with anniversary update?
You need to update to Win10 Anniversary Update or better. Try the Creators update build 15063. Latest nVidia driver will be fine. |
|
|
|
I have one rig with 2x 1060 and 1x 970 and when I try to dual mine eth and sia the 970 eth hashrate is 3 mh.. Can i disable the one 970 card? Or any other fixes? I already tried older drivers..problem not fixed  Same problem on my second rig with 3x 970. I can only get 3mh with 1 card. Use -di to specify which cards to use, Card ID will show when you start. If the bad 970 is card 2, use -di 01 Try -dcri to set intensity of the SIA hash. Default is 30, try lower maybe. -dcri 30 (for all cards) or -dcri 30,30,30 (specify each card separately) Good luck! |
|
|
|
It depends on GPU, settings, selected coin, it's price and net diff. For example, for SIA on stock 1070 ccminer shows 1400MH/s. In dual mode in v9.7 I get 650MH/s SIA (so 45% of pure SIA speed) with only 2% ETH lost (about 25.8MH/s instead of 26.4MH/s).
DCR speed is not so good, agreed.
After seeing this comment I spent some time today testing my 1070's with different -dcri settings. I wasn't getting anywhere near that SC hash. About 30.6mh/s eth and 309mh/s sia with default -dcri 30. While testing on my main PC, I came to the conclusion that a -dcri 60 was about the best trade off for loss of ETH hash vs. gains of SC hash. I was now getting around 29.5mh/s eth and 660mh/s sia. I moved these new settings to the mining rig and the ETH hash just tanked. Checking around I noticed that the cards wanted to draw more power with this new -dcri setting. The 1070's in my rig are overclocked and the power was decreased. Originally when I set up the rig, I tested the miner with ETH only and found I could get away with 75% power on the cards with no change to hash. Increasing the -dcri seems to require more power. I tried pushing the power to 112% and the hash rates just flew. 29.6mh/s and 693mh/s respectively. Seems like quite a lot more power for that extra SIA hash. Just posting as it's something people might want to consider. TL;DR -dcri 30, 75% power = 30.6mh/s eth, 307mh/s sia -dcri 30, 112% power = 30.9mh/s eth, 309mh/s sia -dcri 60, 75% power = 24.5mh/s eth, 490mh/s sia -dcri 60, 112% power = 29.6mh/s eth, 693mh/s sia Did you try experimenting with core clocks while dual mining? I found 75dcri with a higher core clock for my 1070's gave me higher results for both ETH/SIA at the cost of even more power. In the end I just decided dual mining wasnt worth the added power, heat, and stress along with the 2% eth tax. I didn't this time, but as before I found that core didn't really effect ETH hash. I had my core underclocked to reduce power/heat. Good point about the extra tax, especially since I was running with low SIA hash. Will have to think about it some more. On the other hand, mining on my main PC I don't mind power/heat so I'll be playing with my OC's a bit more. |
|
|
|
It depends on GPU, settings, selected coin, it's price and net diff. For example, for SIA on stock 1070 ccminer shows 1400MH/s. In dual mode in v9.7 I get 650MH/s SIA (so 45% of pure SIA speed) with only 2% ETH lost (about 25.8MH/s instead of 26.4MH/s).
DCR speed is not so good, agreed.
After seeing this comment I spent some time today testing my 1070's with different -dcri settings. I wasn't getting anywhere near that SC hash. About 30.6mh/s eth and 309mh/s sia with default -dcri 30. While testing on my main PC, I came to the conclusion that a -dcri 60 was about the best trade off for loss of ETH hash vs. gains of SC hash. I was now getting around 29.5mh/s eth and 660mh/s sia. I moved these new settings to the mining rig and the ETH hash just tanked. Checking around I noticed that the cards wanted to draw more power with this new -dcri setting. The 1070's in my rig are overclocked and the power was decreased. Originally when I set up the rig, I tested the miner with ETH only and found I could get away with 75% power on the cards with no change to hash. Increasing the -dcri seems to require more power. I tried pushing the power to 112% and the hash rates just flew. 29.6mh/s and 693mh/s respectively. Seems like quite a lot more power for that extra SIA hash. Just posting as it's something people might want to consider. TL;DR -dcri 30, 75% power = 30.6mh/s eth, 307mh/s sia -dcri 30, 112% power = 30.9mh/s eth, 309mh/s sia -dcri 60, 75% power = 24.5mh/s eth, 490mh/s sia -dcri 60, 112% power = 29.6mh/s eth, 693mh/s sia |
|
|
|
|
For the EPG, do we have to add a delay to offset the time, or just take it as is?
|
|
|
|
|
