Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 1276.

d57heinz

legendary

Activity: 1453

Merit: 1011

Bitcoin Talks Bullshit Walks

Quote from: Claymore on June 10, 2016, 12:55:33 PM

Confirmed, today after reboot Windows 10 closed v4.5 (on one rig it even removed .exe). I disabled "Windows Update" service in past, but it didn't help, Windows downloaded some update anyway. Damn, how to stop it? It seems I can disable all services and it still downloads all that it wants.

I found this in my journeys

if you goto the search bar at the bottom next to the windows logo. lower left corner.. start typing in "command prompt" when it shows up in the list. Right click it and run it as administrator.

Then type this in the command box.

net stop wuauserv this is the update service

net stop bits this the background intelligent transfer service

edit i just encountered the issue with my 60 gb ssd filling up.. Come to find out its an issue since back in 2012 where the windows.edb file which is somethign to do with drive indexing will bloat to over 60 - 200 gb in size. Thus filling your hdd.. i downloaded winstatdir which is an awesome program and it pointed it out right away.. You have to shut off indexing which is on by default now in win 10 and then run a few commands to delete the .edb file.. mine was 47.6 GB!!!

Hope this helps you and others

Best Regards
d57heinz

nhando

full member

Activity: 196

Merit: 100

Anyone have issues with Windows 10 where the start file when run it just looks like it opens real quick and then close? My Claymore was working for a few days and now is acting up like this. Re-install the OS and it would work the first time and after reboot it would do the same. Driver version is 15.12 AMD R9 390

osnwt

sr. member

Activity: 353

Merit: 251

Quote from: Claymore on June 10, 2016, 01:28:24 PM

I just made Linux port by requests.

And that's great, thanks! Switched my rigs to Linux since stable 4.5 version (I like cmdline UI for administration).

Quote

2. Memory leak means that memory grows all the time. How can I see it?

Simplest method is the 'top' utility that refreshes a screen with a list of processes every second or so. You may see that memory used by ethcmdminer64 process grows every time.

top - 22:23:27 up 9:10, 2 users, load average: 0.26, 0.13, 0.10 Tasks: 159 total, 1 running, 158 sleeping, 0 stopped, 0 zombie %Cpu(s): 0.2 us, 2.9 sy, 0.0 ni, 97.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st KiB Mem: 3529812 total, 2176220 used, 1353592 free, 171264 buffers KiB Swap: 0 total, 0 used, 0 free. 1194856 cached Mem PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 986 root 20 0 467712 160300 117116 S 4.0 4.5 14:43.68 Xorg 29995 work 20 0 7559356 193588 71832 S 1.7 5.5 1:02.57 ethdcrminer64 9472 work 20 0 727680 76300 8864 S 0.3 2.2 0:04.77 node 28301 root 20 0 0 0 0 S 0.3 0.0 0:01.51 kworker/0:1

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Quote from: osnwt on June 10, 2016, 01:15:24 PM

Quote from: BitJunkie223 on June 10, 2016, 07:32:43 AM

Also, I'm guessing in Linux, you're using Wine to run Claymore's miner? I know that Linux seems to increase hashrate, but does this also happen when using Wine for windows mining software, if that makes sense? I'm just thinking that windows is an inherent bottleneck or something, so I assume it would be the same if using any windows software..?

There is a Linux native version of this miner, but:

@Claymore:
- can you confirm my observations that Linux version has lower hashrates comparing to Windows one? With 7950 I see around 1MH/s difference using same PowerTune, cclock and mclock set externally.
- did you notice that Linux version has some memory leaks? I see around 8MB/s of virtual memory leak using top utility. I can restart miner once a week, but still not a good sign.

1. Yes I see that in Linux it works with a bit less speed, by about 1-1.5%. Unfortunately, I don't know why and I don't know Linux much to try to find the reason. Also perhaps it works with the same speed (since it uses same OpenCL binaries) but Linux port of the miner shows hashrate a bit differently, may be the reason is in timers, I'm not sure. I know Windows pretty well, but not Linux, I just made Linux port by requests.
2. Memory leak means that memory grows all the time. How can I see it?

pikachuy

sr. member

Activity: 544

Merit: 250

This might relate to you @osnwt.

Quote from: fredeq on June 07, 2016, 02:41:19 PM

Quote from: melloyellow on June 07, 2016, 01:54:52 PM

Quote from: pikachuy on June 07, 2016, 01:35:48 PM

Quote from: Claymore on June 07, 2016, 01:28:59 PM

Quote from: pikachuy on June 07, 2016, 01:16:26 PM

@Claymore can you explain this?
I have a radeon 380 4gb on a windows 7 x64 running 15.12 drivers and running CDM 4.5 that is undervolted -100mv, core clock at 840, memory clock at 1125, and power limit -20% for days using sapphire trixx. And it hashes at a stable 19~ MH/s constantly and never above 20 MH/s.
I then decided to change the memory clock recently to 1500 just to see if it makes any difference on trixx and pressed apply, and let it run for couple of minutes, but the hash rate barely changed or any change at all. So i changed the memory clock back down to 1125 and press apply, and now the hashrate is hashing at a stable 20~ MH/s.
I am clueless of why setting the memory clock from 1125 to 1500 and then back down to 1125 would increase the reported hashrate in CDM. Any thoughts?

I have no ideas, sorry.

No worries, thought I ask since I can reproduce this result everytime, so now I have to go through this exact pattern everytime I run CDM to get that extra 1 MH/s.

Can someone run the same settings to see if they get the same outcome? Wonder if it's CDM related.

First, let it run with undervolted -100mv, core clock at 840, memory clock at 1125, and power limit -20%. then once you see the reported hashrate, changed the memory clock to 1500 and apply, then after you see the reported hashrate again, changed the memory clock back down to 1125 and press apply. That's when I then see the increased 1 MH/s everytime.

It's not CDM related, read here https://forum.ethereum.org/discussion/comment/39962/#Comment_39962 It's a problem with 380's

How is that related to what pikachuy is reporting? He says that the hashrate goes up AFTER you downclock again to 1125.
Btw I can confirm this as well, pretty weird... got gigabyte 380 on samsung chips

Well, after further testing, it seems like it could be CDM related, because once I did all the steps, and see the increase hashrate, I left the overclocking tool open (saphire trixx) so all the settings don't get changed, and I then exit CDM only and nothing else and reload CDM, but I don't see the increase hashrate. So I have to do the exact steps ONLY when CDM is running again to get the increase 1 MH/s per card.

osnwt

sr. member

Activity: 353

Merit: 251

Quote from: BitJunkie223 on June 10, 2016, 07:32:43 AM

Also, I'm guessing in Linux, you're using Wine to run Claymore's miner? I know that Linux seems to increase hashrate, but does this also happen when using Wine for windows mining software, if that makes sense? I'm just thinking that windows is an inherent bottleneck or something, so I assume it would be the same if using any windows software..?

There is a Linux native version of this miner, but:

@Claymore:
- can you confirm my observations that Linux version has lower hashrates comparing to Windows one? With 7950 I see around 1MH/s difference using same PowerTune, cclock and mclock set externally.
- did you notice that Linux version has some memory leaks? I see around 8MB/s of virtual memory leak using top utility. I can restart miner once a week, but still not a good sign.

JuanHungLo

hero member

Activity: 935

Merit: 1001

I don't always drink...

Quote from: merc84 on June 10, 2016, 10:35:41 AM

Quote from: JuanHungLo on June 10, 2016, 09:48:17 AM

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

I guess u have hynix mem? My cards have elipdia and barely break 19Mh/s @ 1100/1500

Yeah, they're Hynix. I had to do a lot of shopping and reselling to find those.

Claymore

donator

Activity: 1610

Merit: 1325

Miners developer

Confirmed, today after reboot Windows 10 closed v4.5 (on one rig it even removed .exe). I disabled "Windows Update" service in past, but it didn't help, Windows downloaded some update anyway. Damn, how to stop it? It seems I can disable all services and it still downloads all that it wants.

DragonSlayer

full member

Activity: 181

Merit: 100

Yep, last Defender update is jacking everyone. I went thru my boxes and added exceptions to folder.

calculatorduck

member

Activity: 93

Merit: 10

Quote from: Pustul on June 10, 2016, 10:38:26 AM

Windows 10 now detects v4.5 as a virus (Trojan), it deleted the file from my computer and when I try to download it again, it says: Virus detected... Is it safe to restore it?

Yep, my rig stopped mining and I realized it was because Windows Defender updated and then quarantined the exe.

elzium

member

Activity: 96

Merit: 10

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

Thats minor difference, not enough to cool down overheated card. Unless it doesnt change some other voltages automatically. I tried it right now, underclocked 2x290x 40MHz both and consumption went down 15W from 600, 1 degree drop... Not much point. But cards are different of course.

-dcri gives much better effect.

phzi

hero member

Activity: 700

Merit: 500

Quote from: Pustul on June 10, 2016, 10:38:26 AM

Windows 10 now detects v4.5 as a virus (Trojan), it deleted the file from my computer and when I try to download it again, it says: Virus detected... Is it safe to restore it?

Another typical false positive. Just exclude the directory.

I wonder how many people this impacted - looks like network difficulty took a bit of a dive.

Pustul

sr. member

Activity: 350

Merit: 250

Windows 10 now detects v4.5 as a virus (Trojan), it deleted the file from my computer and when I try to download it again, it says: Virus detected... Is it safe to restore it?

merc84

hero member

Activity: 799

Merit: 1000

Quote from: JuanHungLo on June 10, 2016, 09:48:17 AM

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

I guess u have hynix mem? My cards have elipdia and barely break 19Mh/s @ 1100/1500

CoRpO

full member

Activity: 185

Merit: 100

Quote from: JuanHungLo on June 10, 2016, 10:16:18 AM

Quote from: CoRpO on June 10, 2016, 10:07:51 AM

Quote from: JuanHungLo on June 10, 2016, 09:48:17 AM

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

Tried 875/1500 ? should hash better with very little more power

No difference with mine whether 1250 or 1500.

You're lucky. @875/1250 i got 17.7 Mhs

JuanHungLo

hero member

Activity: 935

Merit: 1001

I don't always drink...

Quote from: CoRpO on June 10, 2016, 10:07:51 AM

Quote from: JuanHungLo on June 10, 2016, 09:48:17 AM

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

Tried 875/1500 ? should hash better with very little more power

No difference with mine whether 1250 or 1500.

CoRpO

full member

Activity: 185

Merit: 100

Quote from: JuanHungLo on June 10, 2016, 09:48:17 AM

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

Tried 875/1500 ? should hash better with very little more power

JuanHungLo

hero member

Activity: 935

Merit: 1001

I don't always drink...

Quote from: merc84 on June 10, 2016, 09:28:37 AM

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

I undervolt those same cards to 950 then mine ETH with 875/1250 for maximum savings. Still gets over 19 M/s that way.

merc84

hero member

Activity: 799

Merit: 1000

Clocks most certainly do change temps. I have several gigabyte wf3 280x with fixed voltage and the temp increases with clock, higher clock = more power draw = more heat.

elzium

member

Activity: 96

Merit: 10

Quote from: merc84 on June 10, 2016, 08:45:01 AM

Quote from: Claymore on June 10, 2016, 05:08:22 AM

Quote from: merc84 on June 10, 2016, 04:48:32 AM

Quote from: Claymore on June 10, 2016, 01:43:22 AM

Quote from: baykan on June 09, 2016, 04:40:06 PM

@Claymore
I need a parameter saying to miner, if the temperature of a card is greater than a threshold of let's say 80 degrees than switch to ether only mining rather then dual mining.
samething for cold environment, if gpu temp is colder than 60 degrees than dual mine.
Summer is beginning here in İstanbul so this would be very helpfull.
I Appreciate your good work.

Ok, I will add it in next update, it will be available in 3-4 days.

How would this work, would all cards switch to eth only or only cards that are below threshold?

Every card will be managed separately.

How about adding dynamic OC also for example adjust clocks within specified range depending on temp threshold.

Clocks dont change temps, voltage does. Or u mean oc potential?

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 1276. (Read 6590757 times)