Krak
|
|
May 14, 2012, 06:53:48 PM |
|
No, I haven't. What does the flag do?
--gpu-reorder Attempt to reorder GPU devices according to PCI Bus ID
Although in my experience, it was only necessary when I used Windows. So far it's been accurate without that flag on Ubuntu.
|
BTC: 1KrakenLFEFg33A4f6xpwgv3UUoxrLPuGn
|
|
|
MrTeal
Legendary
Offline
Activity: 1274
Merit: 1004
|
|
May 14, 2012, 06:55:48 PM |
|
I tried --gpu-reorder, and it did organize them in a more logical layout but it doesn't fix the problem. [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit GPU 0: 64.0C 3192RPM | 273.0/274.5Mh/s | A:15 R:0 HW:0 U: 4.82/m I: 8 GPU 1: 59.0C 4408RPM | 331.0/330.2Mh/s | A:25 R:0 HW:0 U: 8.04/m I: 8 GPU 2: 60.5C 4403RPM | 88.6/177.9Mh/s | A: 5 R:0 HW:0 U: 1.61/m I: 8 GPU 3: 36.0C | 331.3/330.2Mh/s | A:14 R:0 HW:0 U: 4.50/m I: 8 Here, I changed the core of the single GPU 5970 to 200MHz and the voltage to 0.950V. The hashrate dropped on GPU2, but the corresponding drop in temperature happened on GPU3.
|
|
|
|
MrTeal
Legendary
Offline
Activity: 1274
Merit: 1004
|
|
May 14, 2012, 06:58:42 PM |
|
Check the debug log file. Sometimes ADL does return -1 from fan speed/temp APIs. I've had people reporting it with 6000 series cards (not cgminer, my akbash watchdog) BTW, -1, is "Most likely one or more of the Escape calls to the driver failed".
Not sure why, maybe it is "not always supported" (as their ADL docs says) ?!?. I raised this issue with AMD support. Waiting for their response.
Not sure how re-order would help, ADL APIs use adapter index, not opencl gpu #.
The lack of RPM for the one 5970 has to do with how it died. The GPU that failed was the one closest to the output, so this card isn't capable of outputting an image, or controlling/reporting on fan speed. The fan is just always pegged at 100.
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4242
Merit: 1644
Ruu \o/
|
|
May 14, 2012, 11:52:31 PM |
|
Please take a good look through the readme, and read carefully the extensive documentation on the advanced option --gpu-map.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
mdude77
Legendary
Offline
Activity: 1540
Merit: 1001
|
|
May 15, 2012, 12:13:24 AM |
|
Question: Is there a way to mine on multiple pools at the same time with CGMiner? In other words, if I have 2 BFL miners, and want to point one to one pool, and another to another pool, how would I go about setting that up?
The --load-balance flag will basically do that for you. How accurately would it load balance between the pools? I heard -haven't tried- that it doesn't quite work as expected on balancing. I use it for failover only.
|
I mine at Kano's Pool because it pays the best and is completely transparent! Come join me!
|
|
|
mdude77
Legendary
Offline
Activity: 1540
Merit: 1001
|
|
May 15, 2012, 12:14:34 AM |
|
It's a 480 GB SSD, I don't think Spinrite will help The box is a fairly recent reload, I'm not keen on doing it again. I was hoping someone might have a magic bullet. Actually, I read recently that spinrite level 1 is read only, and can work wonders on SSDs.
|
I mine at Kano's Pool because it pays the best and is completely transparent! Come join me!
|
|
|
Inaba
Legendary
Offline
Activity: 1260
Merit: 1000
|
|
May 15, 2012, 12:18:54 AM |
|
It's a 480 GB SSD, I don't think Spinrite will help The box is a fairly recent reload, I'm not keen on doing it again. I was hoping someone might have a magic bullet. Actually, I read recently that spinrite level 1 is read only, and can work wonders on SSDs. I'm not even sure how that would make sense?
|
If you're searching these lines for a point, you've probably missed it. There was never anything there in the first place.
|
|
|
mdude77
Legendary
Offline
Activity: 1540
Merit: 1001
|
|
May 15, 2012, 12:52:44 AM |
|
It's a 480 GB SSD, I don't think Spinrite will help The box is a fairly recent reload, I'm not keen on doing it again. I was hoping someone might have a magic bullet. Actually, I read recently that spinrite level 1 is read only, and can work wonders on SSDs. I'm not even sure how that would make sense? You can read from SSDs almost indefinitely without causing wear. It's the writing that causes problems. So by doing a level 1 read across the drive, you're forcing the wear logic to realize failing sectors are having problems, which it then swaps out to a spare one. Steve's words: And all of our listeners just got a new tip for running SpinRite, if you have a drive which, like this - the problem is that all of the other levels are writing something. Level 1 is a read-only pass. And that's why it's safe to run on thumb drives, because it doesn't write anything, absolutely nothing. It only reads. But the beauty of that is that, as we were saying before, the act of reading shows the drive it has a problem. And clearly this, whatever was going wacky with this and a couple other drives that Mike found, writing gave the drive fits, but reading was okay. So reading was sort of eased into it more gently and allowed the drive to fix the problems so that then writing to them was writing to different areas because the bad spots had been relocated to good areas on the drive. So that's a great tip. It'll definitely make it into our notes for the future.
from: https://www.grc.com/sn/sn-343.pdf
|
I mine at Kano's Pool because it pays the best and is completely transparent! Come join me!
|
|
|
PolymorphicAssasin
Newbie
Offline
Activity: 46
Merit: 0
|
|
May 15, 2012, 09:26:51 PM |
|
... Updated git tree: I've added dynamic adjustment of intensity for usage on a normal desktop. This flag is enabled by default and tests to ensure the GPU is constantly available for desktop use and will scale intensity down when you are watching a movie, gaming or just about any other intense GPU usage, and scale it up when otherwise idle. This is best disabled on dedicated miners: ... The difference can be quite dramatic in terms of how much smoother the desktop is, and how much higher the throughput is when it's idle.
I just switched over from Diablo on my work PC, and I have to say this is the one of the coolest things about cgminer. I don't have to kill my miner whenever I'm doing graphic intense work (2x or 3x a day). Kudos. I'll send you some bitcents when next I access my wallet (from cold storage) Thanks for some great software! Finally got around to sending those cents. You should see them in the next block. Thanks again for the great contribution to the community. I'd send more if I was richer.
|
|
|
|
P_Shep
Legendary
Offline
Activity: 1795
Merit: 1208
This is not OK.
|
|
May 16, 2012, 12:30:15 AM |
|
Another this I've noticed: When you change the priority of a pool to the top (via the API) and you save the config file, the order of the pools in the config file remains unchanged from is original. Maybe this should change to match the current priority order?
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
May 16, 2012, 02:21:34 AM |
|
Another this I've noticed: When you change the priority of a pool to the top (via the API) and you save the config file, the order of the pools in the config file remains unchanged from is original. Maybe this should change to match the current priority order?
However, if your main pool has a temporary outage, then it may be moved from the top position. So it would have to take that into account also ...
|
|
|
|
P_Shep
Legendary
Offline
Activity: 1795
Merit: 1208
This is not OK.
|
|
May 16, 2012, 03:44:34 AM |
|
Another this I've noticed: When you change the priority of a pool to the top (via the API) and you save the config file, the order of the pools in the config file remains unchanged from is original. Maybe this should change to match the current priority order?
However, if your main pool has a temporary outage, then it may be moved from the top position. So it would have to take that into account also ... Unlucky to hit save at that moment, but yeah, possible. Not so sure that's something to worry about. Just an idea/suggestion anyway.
|
|
|
|
ddd1
|
|
May 16, 2012, 05:54:51 AM |
|
What should temp-cutoff and overheat be to avoid card throttling down in speed?
I have watercooling and want the app to shutdown if core reaches 60c.
I found this in my cgminer.cfg "temp-cutoff" : "95", This I put to 60c "temp-overheat" : "85", This I can keep at 85? "temp-target" : "75", This I can keep at 75? ^^^^^^^ There is no fans controlled by GPU, it's radiator with fans.
So MAX temps on my 7950 is: Core MAX 52c, VRM MAX 65c.
I'm assuming just changing the "temp-cutoff" to 60 and hysteria at 3 and not changing other stuff incase of waterpump failure the GPU once it reaches 63c will turn off the mining then?
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4242
Merit: 1644
Ruu \o/
|
|
May 16, 2012, 02:07:40 PM |
|
What should temp-cutoff and overheat be to avoid card throttling down in speed?
I have watercooling and want the app to shutdown if core reaches 60c.
I found this in my cgminer.cfg "temp-cutoff" : "95", This I put to 60c "temp-overheat" : "85", This I can keep at 85? "temp-target" : "75", This I can keep at 75? ^^^^^^^ There is no fans controlled by GPU, it's radiator with fans.
So MAX temps on my 7950 is: Core MAX 52c, VRM MAX 65c.
I'm assuming just changing the "temp-cutoff" to 60 and hysteria at 3 and not changing other stuff incase of waterpump failure the GPU once it reaches 63c will turn off the mining then?
It will cut off at 60, but it uses the temp target to re-enable the mining, so best to set that lower or it will keep enabling and disabling it.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
May 16, 2012, 05:21:13 PM |
|
It will cut off at 60, but it uses the temp target to re-enable the mining, so best to set that lower or it will keep enabling and disabling it. This. Also remember if you specify nothing for a param then cgminer uses the default. I alway prefer explicit values rather than suprises due to implicit values. I just have it set at target & overheat @ 58, and cutoff @ 60.
|
|
|
|
lazyranma
Newbie
Offline
Activity: 14
Merit: 0
|
|
May 16, 2012, 05:46:49 PM |
|
I have a strange issue with cgminer... After a day or two of work hash rate starts to slowly drop from usual 440 Mhash/s and reaches 340 Mhash/s within 30-40 minutes where it stops. Restarting cgminer brings hashrate back to usual values. Any ideas why this happens? Right now I have a script which checks hash rate periodically and restarts cgminer when needed. Ubuntu Linux 12.04, Catalyst 12.4, HD7870.
|
|
|
|
os2sam
Legendary
Offline
Activity: 3583
Merit: 1094
Think for yourself
|
|
May 16, 2012, 05:49:03 PM |
|
I have a strange issue with cgminer... After a day or two of work hash rate starts to slowly drop from usual 440 Mhash/s and reaches 340 Mhash/s within 30-40 minutes where it stops. Restarting cgminer brings hashrate back to usual values. Any ideas why this happens?
Does the utility drop too? Sam
|
A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
May 16, 2012, 11:58:48 PM |
|
I have a strange issue with cgminer... After a day or two of work hash rate starts to slowly drop from usual 440 Mhash/s and reaches 340 Mhash/s within 30-40 minutes where it stops. Restarting cgminer brings hashrate back to usual values. Any ideas why this happens? Right now I have a script which checks hash rate periodically and restarts cgminer when needed. Ubuntu Linux 12.04, Catalyst 12.4, HD7870.
At a guess ... 12.4 is your problem ... since I heard it would often just pack up and die for no reason with cgminer.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
May 17, 2012, 03:08:28 AM |
|
I've got an Icarus pull request up that should increase performance: https://github.com/ckolivas/cgminer/pull/204The gain would be due to it running the Icarus much closer to when it becomes idle The code change is rather extensive since it now includes a new (hidden) option --icarus-timing If you have an Icarus device (that works the same as Icarus and has 2 FPGA) then this option will allow you to use that device also and at optimum performance (also e.g. if you have an Icarus with a faster bitstream) For more details see the new FPGA-README in the pull request (the code is extensively commented also) The new default MH/s calculation is actually a VERY slight underestimate (at the 5th digit the default time to do a single hash is rounded up) So the increase seen should actually be valid (assuming no bugs ) This calculation only comes into play when a piece of work is aborted at the new 11.2s (or due to an LP)
|
|
|
|
ddd1
|
|
May 17, 2012, 06:47:56 AM |
|
Win32/Coinminer is virus?
My windows defender reported that this was harmful, googling it lead to sites saying it steals the wallet.
Now is this something to worry about? I have removed it when the window popped up.
Now it removed cgminer.exe, wery wierd??
|
|
|
|
|