Pages:
Author

Topic: python OpenCL bitcoin miner - page 29. (Read 1239072 times)

full member
Activity: 171
Merit: 127
February 22, 2011, 08:20:41 AM
New version is up. Changes:

- handling httplib exceptions
- proper handling of wrong user name and password when mining with bitcoind
- support for more than one OpenCL platform
- caching kernel binaries for faster startup
- always flush stdout
full member
Activity: 171
Merit: 127
February 22, 2011, 04:12:29 AM
aistto, see http://blog.zorinaq.com/?e=11

Initially, you can start the miner on first device, then connect your monitor to second device to make it available. If this seems too difficult, see above link for a way to keep second device always active. This is a windows only problem.
legendary
Activity: 1001
Merit: 1005
February 22, 2011, 03:32:37 AM
And what about version  20101126 and earlier? may be I could use it? where can I download it?

maybe you could, i wouldn't recommend it though if you don't really have to use it (have a 5970/dualGPU card)
on a 2x 5870 setup you can easily disable crossfire and use the latest miner-version on both cards.

if you want to use 20101126 anyway, here's a mirror.


when I disable crossfire only first device is available. What should I do to use 2-nd?

crossfire disable
  •      Cypress
  • [1]     AMD Athlon(tm) II X2 220 Processor

    crossfire anable
  •      Cypress
  • [1]     Cypress
    [2]     AMD Athlon(tm) II X2 220 Processor
newbie
Activity: 47
Merit: 0
February 21, 2011, 06:38:49 PM
Any advice on autostarting the miner when the computer boots up (without me logging in)? I configured bitcoind for autostart using this script from the forum.

I suppose for the miner I somehow need to configure envrionment variables (ATI Stream SDK) to be available upon boot. Plus it is confusing as I will start two instances of poclbm, one for each GPU I have.

I suppose I could still copy the script above in principle. One more problem, though: I think I want to pipe the verbose output of poclbm into a script that will parse and log the information, so I suppose I would start it like this

Quote
./poclbm.py -u somebody --pass=pwd -d 2 -v -w 128 --verbose | myparser.py

But what PID will be returned if I feed this to the start-stop-daemon command? If I use pipes, is the script I pipe into dependent on the originating script, that is, if I kill poclbm, will myparser.py also be killed?

Anything else I need to consider?

Thx!

Update: I just tried to start poclbm via ssh login from another computer. If the miner only shows the login screen, I get

Quote
No protocol specified
No device specified or device not found, use -d to specify one of the following

  •    Intel(R) Core(TM) i7 CPU         860  @ 2.80GHz
Once I log in to my account on the miner, I can also start poclbm via ssh.

What could be the reason? I don't suppose it is the environment variables, as they are configured in .bashrc. So I assume they should be known if I log in via ssh, too.
hero member
Activity: 532
Merit: 505
February 21, 2011, 06:31:30 PM
And what about version  20101126 and earlier? may be I could use it? where can I download it?

maybe you could, i wouldn't recommend it though if you don't really have to use it (have a 5970/dualGPU card)
on a 2x 5870 setup you can easily disable crossfire and use the latest miner-version on both cards.

if you want to use 20101126 anyway, here's a mirror.


legendary
Activity: 1001
Merit: 1005
February 21, 2011, 05:33:13 PM
@aistto

Switch crossfire off for now. Until I found why versions until and including 20101126 work with crossfire and later ones don't.


And what about version  20101126 and earlier? may be I could use it? where can I download it?
full member
Activity: 171
Merit: 127
February 21, 2011, 04:33:51 PM
@aistto

Switch crossfire off for now. Until I found why versions until and including 20101126 work with crossfire and later ones don't.

Using python 3.2 on windows  Undecided

No, no, no... use python 2.6

poclbm is not compatible with python 3. Actually python 2 and python 3 are quite different and not compatible with each other.
newbie
Activity: 31
Merit: 0
February 21, 2011, 04:01:59 PM
I'm running at ca. 42 Mhps on Mobility Radeon HD 5650, OpenCL SDK 2.3. Best options so far: -d 0 --user=XXX --pass=XXX -w 128 -v -f 5 (I'm not sure what does -f parameter do really, I'm a dummy).
Do you think that individual mining with laptop GPU is worth it? Note: I'm not paying for the electricity.
From what I've read, "-f x" is "framerate", i.e. the program will try to balance the kernel so that the kernel is run x times per second. This means that a lower value makes the kernel run for longer (-f 5 means the kernel takes about 0.2 seconds). If your OS feels sluggish (windows lag when you drag them on the desktop, videos seem choppy etc.) you can try to set -f higher. I use -f 120. At -f 60 I get slight lag when playing back HD videos, and the performance difference is only about 6% so I would much rather have a fully working machine. At -f 120 I notice no slowdown.

At least that is how I've understood it, I'm sure someone will correct me if I am wrong.

Edit: And if you're not paying anything for the electricity and it does not bother you to have it running in the background (presumably when the laptop is plugged in!) then it's free money, so why not? If you think bitcoin is interesting I see no reason why you wouldn't mine Smiley
jr. member
Activity: 56
Merit: 22
February 21, 2011, 03:48:14 PM
Had some errors too.

poclbm should had used a catch elseif for this error, to make more clear that something went wrong.
Or at least make it less scary, just make it print 'Could not fetch new work' Tongue
~


Code:
File "poclbm.py", line 25  print 'No device specified or device not found, use -d to specify one of the following\n'
            ^
SyntaxError: invalid syntax

Using python 3.2 on windows  Undecided
legendary
Activity: 1001
Merit: 1005
February 21, 2011, 03:21:32 PM
MB MSI Socket-AM3 870A-G54 (pci-e x16 in dual cross work as 8/8 (but someone says that as 16/4))
catalyst 11.1
stream 2.2
1. hd5870 gigabyte
2. hd5870 his
I used 2 crossfire bridge
not overclocked

In my crossfire system doesn't work second card. it's look like below
Quote
F:\MySpace\btc\poclbm_py2exe_20110204>poclbm.exe
No device specified or device not found, use -d to specify one of the following

  •     Cypress
  • [1]     Cypress
    [2]     AMD Athlon(tm) II X2 220 Processor

    F:\MySpace\btc\poclbm_py2exe_20110204>poclbm.exe -u usr --pass=pswrd -o 178.79.147.99 -p 8332 -d 0 -v -w 128 -f 30
    21/02/2011 23:07, b2ca3069, accepted
    21/02/2011 23:07, 818e219a, accepted
    312424 khash/s
    Compilation terminated.
    (work as usual)

    F:\MySpace\btc\poclbm_py2exe_20110204>poclbm.exe -u usr --pass=pswrd-o 178.79.147.99 -p 8332 -d 1 -v -w 128 -f 30
    21/02/2011 23:08, 5f141da2, accepted (Here always accepting first block)
    verification failed, check hardware!
    verification failed, check hardware!
    verification failed, check hardware!
Why? Huh

sorry for my english:)
legendary
Activity: 1386
Merit: 1097
February 21, 2011, 03:04:01 PM
I too am getting intermittent errors while using slush's pool with win7-64:

If you hit that before hour or two, I was upgrading the pool servers and many connections were cut.

Code:
BadStatusLine

This just means that data from connection were corrupted (probably because connection was forcibly closed) and poclbm don't catch it. It is just cosmetic.
member
Activity: 73
Merit: 10
February 21, 2011, 02:59:40 PM
I too am getting intermittent errors while using slush's pool with win7-64:


Code:
21/02/2011 13:34:44, Unexpected error:
Traceback (most recent call last):
  File "BitcoinMiner.pyc", line 210, in mine
  File "BitcoinMiner.pyc", line 187, in getwork
  File "httplib.pyc", line 974, in getresponse
  File "httplib.pyc", line 391, in begin
  File "httplib.pyc", line 355, in _read_status
BadStatusLine

Code:
21/02/2011 13:34:45, Unexpected error:
Traceback (most recent call last):
  File "BitcoinMiner.pyc", line 210, in mine
  File "BitcoinMiner.pyc", line 187, in getwork
  File "httplib.pyc", line 974, in getresponse
  File "httplib.pyc", line 391, in begin
  File "httplib.pyc", line 355, in _read_status
BadStatusLine

Those are both from approximately around the same time, but each one occurred on its own graphics card (I am running 2 nvidia gtx 275 cards). It seems that they occur very chronologically close to one another. Further, it either occurs on only one card, or it happens on both of them, one within a few seconds of the other.

Here is another example, each paste is from a miner running on its own gpu:

Code:
21/02/2011 13:53:46, Unexpected error:
Traceback (most recent call last):
  File "BitcoinMiner.pyc", line 210, in mine
  File "BitcoinMiner.pyc", line 187, in getwork
  File "httplib.pyc", line 974, in getresponse
  File "httplib.pyc", line 391, in begin
  File "httplib.pyc", line 355, in _read_status
BadStatusLine

Code:
21/02/2011 13:53:44, Unexpected error:
Traceback (most recent call last):
  File "BitcoinMiner.pyc", line 210, in mine
  File "BitcoinMiner.pyc", line 187, in getwork
  File "httplib.pyc", line 974, in getresponse
  File "httplib.pyc", line 391, in begin
  File "httplib.pyc", line 355, in _read_status
BadStatusLine
newbie
Activity: 1
Merit: 0
February 21, 2011, 10:04:12 AM
Ok, usually I start the miner on slush's server with regular settings, just not specifying the -w value.
I read about it somewhere else, so I decided to mess around with it.
I'm running a Nvidia GeForce 8600M GT, and generally get about 5000 khash/s without specifying the -w value.
Then, when I set -w to 512, I got a hash rate of 1880356953 khash/s! I'm pretty sure that's messed up. And it ran like that for about 30 seconds before it crashed.
Glitch....or what?
On win7, btw.
newbie
Activity: 5
Merit: 0
February 21, 2011, 09:36:45 AM
I'm running at ca. 42 Mhps on Mobility Radeon HD 5650, OpenCL SDK 2.3. Best options so far: -d 0 --user=XXX --pass=XXX -w 128 -v -f 5 (I'm not sure what does -f parameter do really, I'm a dummy).
Do you think that individual mining with laptop GPU is worth it? Note: I'm not paying for the electricity.
newbie
Activity: 47
Merit: 0
February 21, 2011, 08:03:11 AM
Yeah, I'll probably fork poclbm on GitHub.
full member
Activity: 238
Merit: 100
February 21, 2011, 07:16:48 AM
I saw that, but I think it involves changing the code of poclbm? While doable, it would be a nuisance if new versions of poclbm appear and the modifications had to be repeated.

Do you know how to use git? You can make a modification in your copy and each time m0mchill changes poclbm, you simply pull git changes and if it does not conflict with your version, the changes apply without any fuss. This is a very convenient and standard way of keeping a fork of a program up to date.
newbie
Activity: 47
Merit: 0
February 21, 2011, 06:53:57 AM
Quote
the answer to my question about the miner sending an email every time it finds a block

I saw that, but I think it involves changing the code of poclbm? While doable, it would be a nuisance if new versions of poclbm appear and the modifications had to be repeated.

I think m0mchil mentioned something about hooks into poclbm?

@Raulo thanks for the info about the DISPLAY variable. I urgently need to get an understanding of that display stuff.
full member
Activity: 238
Merit: 100
February 21, 2011, 02:14:18 AM
The problem with that is when you login via ssh not all features of aticonfig work (they only work properly if you're using it directly from the machine which has the screen connected). I am sure Raulo will be able to shed some light here. Wink

Everything works exactly the same with ssh as it works via console if you have DISPLAY=:0 set and you are the user that "owns" the display. If the latter is not fulfilled, being root and having DISPLAY=:0 is all you need.
uck
member
Activity: 94
Merit: 10
February 20, 2011, 09:05:44 PM
I am seeing this exact same error. Let me know if you solve it...
sr. member
Activity: 434
Merit: 251
Every saint has a past. Every sinner has a future.
February 20, 2011, 08:51:11 PM
Is there a way to monitor this miner? Like I have it running on a PC in the cellar that I ssh into. Can I check the current hash rates and success rates somehow?

If you scroll up a few posts you will find the answer to my question about the miner sending an email every time it finds a block (and it's accepted). At least this way I know when a block was found. I think the best thing to do would be having some kind of an interface similar to the one bitcoind has (getbalance, getdifficulty etc.) - just having a few of such features on the miner would be cool (gethashrate, listfoundblocks etc.). As for the temperature (you were asking about it in the other thread) it's easy to check it via the aticonfig tool (unless you're using nvidia, I don't know how to do it for nvidia cards). You could have a script monitoring the output from: sudo aticonfig --odgc

The problem with that is when you login via ssh not all features of aticonfig work (they only work properly if you're using it directly from the machine which has the screen connected). I am sure Raulo will be able to shed some light here. Wink
Pages:
Jump to: