bitHopper: Python Pool Hopper Proxy

[creativex]

Penalty should be set to 1.0 not 4.0. You can change it on stats page temporarily but the longterm solution is to change it in user.cfg.

[joulesbeef]

Using Altslicer, Slush currently on 14%, still requesting work...

EDIT:  Stopped it manually at 18% as I'm heading out and don't trust it to stop on its own.  Regardless, it's not quite working as intended yet.

Try mining slush with role:mine instead of mine_slush and penalty set to 4.0

That should make it act like a normal mine only it will hop off at about 11%

I do this cause I was having issues as well and it seems to work out better for me

slush just had a nice 3 in a row

[GenTarkin]

Penalty should be set to 1.0 not 4.0. You can change it on stats page temporarily but the longterm solution is to change it in user.cfg.

for slush?? cuz the user.cfg has it set to 4.0 defaulted.

[creativex]

Correct. Change penalty to 1.0 for mine slush...or as the cow up dere says you can leave it set to 4.0 and change role to mine.

Three quick blocks in a row just now. Gotta love that mine_slush role. Hope you guys got some shares in there.

[joulesbeef]

yeah we are both right
my way gets off at 11% though
that is mine with a penalty of 4
his way gets off at 14%
that is mine_slush with penalty of 1

all of this is in theory as c00w is often up late creating new bugs

[EskimoBob]

in pools.cfg, [btcpool24] is missing a

Code:

url:http://www.btcpool24.com

I changed slush penalty to 2.5, lets see how that works out.

[simonk83]

Slush didn't change until 21% this time.  Definitely needs some tweaking.   Latest c00w's (penalty at 1, mine_slush, altslicer).

[flower1024]

Data point on biclockers:
With Cherry: .513% stales.
2.3Gh/s on one worker from 10 GPUs
Same account and worker since before I started hopping.
FWIW.

with my poclbm mod i am running at 0.28% and 0.33% stales (only twominers running).

and still using the nicer stats grabbing and pool decisions from bitthopper

Cool I'll have to check it out. Dies it work with ryouiki fork or just c00w's?

I think it reinforces that bithopper is still identifiable by bitclockers when it's in between.

i think it should work with ryo too.. never checked; but he forked after i integrated (my lame) first website - so i guess the json data could still be the same.

just try

[bitcoindaddy]

with my poclbm mod i am running at 0.28% and 0.33% stales (only twominers running).

and still using the nicer stats grabbing and pool decisions from bitthopper

so you arent using cherry picker.. just modded poclbm and bithopper?

yes,

my poclbm-mod connects to bithopper and loads a pool list from there.

it checks the page every three seconds and switches (with no stale or rejects) to the new bithopper-current

you'll lose the worker management and submitted shares display in bithopper (and it sometimes shows an exception on a flaky connection - which could safely be ignored [just had no time to fix it right now])

btw. i am working on this atm: http://miner.k1024.de
its pre pre pre alpha. only thing that works:
a nice stats page where two bithopper instances pushes their round share information.
they are both running on different networks - just to make sure the pools dont start playing "whats your ip" with us

(suggestions are welcome)

I'm getting lots of error's like this (on Ubuntu 10.10 64 bit):

/tmp/OCLgKNyWT.cl(197): error: bad argument type to opencl builtin function:
          expected type "uint2", actual type "int"
     W[27] = P1(27) + P3(27);
             ^

Error limit reached.
100 errors detected in the compilation of "/tmp/OCLgKNyWT.cl".
Compilation terminated.

[burp]

Regarding the bitclocker issues: I don't think they slow down or disturb hoppers connections. Their server is just totally overloaded or in any other way can't handle things. There is the utility httping, which is like a ping over HTTP. It measures the time to get a response from the webserver. Using this utility I get from multiple! locations pings like ~0.6s to their main webserver and pings like 3-5 seconds to their pools. I guess they are just unable to set up proper working services, and it affects hoppers as well as their normal miners.

[flower1024]

I'm getting lots of error's like this (on Ubuntu 10.10 64 bit):

/tmp/OCLgKNyWT.cl(197): error: bad argument type to opencl builtin function:
          expected type "uint2", actual type "int"
     W[27] = P1(27) + P3(27);
             ^

Error limit reached.
100 errors detected in the compilation of "/tmp/OCLgKNyWT.cl".
Compilation terminated.

i have answered here: https://bitcointalk.org/index.php?topic=34514.msg437154#msg437154

[ewibit]

Regarding the bitclocker issues: I don't think they slow down or disturb hoppers connections. Their server is just totally overloaded or in any other way can't handle things.

is this the reason too that I have <= 3% rejected and bithopper always disconnects from bitclocker?

[burp]

Regarding the bitclocker issues: I don't think they slow down or disturb hoppers connections. Their server is just totally overloaded or in any other way can't handle things.

is this the reason too that I have <= 3% rejected and bithopper always disconnects from bitclocker?

Yep, that is the reason it always disconnects from bitclockers. Right now they appear to be completely down to me.

[bitcoindaddy]

i have answered here: https://bitcointalk.org/index.php?topic=34514.msg437154#msg437154

I got it to work by removing the phatk2 keyword. I'm answering here and not in the other thread because this highlights that there is a problem with bithopper and bitclockers.  I have 7 miners and bithopper didn't stop showing bitclockers as red and always switching away from it until I got the very last miner off of it and was mining through flower1024/poclbm.

At first I thought it might be due to the fact that with flower1024/poclbm, there are (in my case) 7 different source ports and 7 different TCP/IP connections whereas with bithopper there is one. But why was I still getting problems with only one miner connected to bithopper?

I've been running 14 minutes now and have zero rejects on bitclockers. I do miss the information bithopper gave me though (rounds submitted for example).

[EskimoBob]

Regarding the bitclocker issues: I don't think they slow down or disturb hoppers connections. Their server is just totally overloaded or in any other way can't handle things.

is this the reason too that I have <= 3% rejected and bithopper always disconnects from bitclocker?

Yep, that is the reason it always disconnects from bitclockers. Right now they appear to be completely down to me.

cgminer has a really cool way to change pools while mining so I switched manually to bitclocker. It gives me few connection errors here an there but shares are submitted and I get new work.

Code:

... 14:59:56] Accepted 479df52a GPU 0 thread 0 pool 8
[2011-08-07 15:00:02] Accepted 259ced85 GPU 0 thread 1 pool 8
[2011-08-07 15:00:10] Pool 8 communication failure, caching submissions
[2011-08-07 15:00:10] Stale share detected, discarding

BitHopper script is too hysterical about minor connection issues and switches pools way too often for no serious reason.

[burp]

cgminer has a really cool way to change pools while mining so I switched manually to bitclocker. It gives me few connection errors here an there but shares are submitted and I get new work.

Code:

... 14:59:56] Accepted 479df52a GPU 0 thread 0 pool 8
[2011-08-07 15:00:02] Accepted 259ced85 GPU 0 thread 1 pool 8
[2011-08-07 15:00:10] Pool 8 communication failure, caching submissions
[2011-08-07 15:00:10] Stale share detected, discarding

BitHopper script is too hysterical about minor connection issues and switches pools way too often for no serious reason.

Yeah, that is the issue. As soon as bitHopper gets an empty/invalid getwork reply or hits a timeout it switches the pool. Not sure if this is the correct spot, but something like this might work:

Code:

iff --git a/work.py b/work.py
index 34330b2..9b23a6b 100644
--- a/work.py
+++ b/work.py
@@ -126,10 +126,10 @@ def jsonrpc_getwork(agent, server, data, j_id, request, bitHopper):
     work = None
     while work == None:
         i += 1
-        if data == [] and i > 1:
+        if data == [] and i > 4:
             server = bitHopper.get_new_server(server)
         try:
-            if i > 4:
+            if i > 8:
                 time.sleep(0.1)
             work = yield jsonrpc_call(agent, server,data,bitHopper)
         except Exception, e:

[cirz8]

Not sure if it's a bug or what it is, but if you start bitHopper from some other directory using a command like this

Code: (How to trigger)

cd / && /home/user/bithopper/bin/bitHopper/bitHopper.py --scheduler=AltSliceScheduler

Code: (reply from http://localhost:8337/stats)

web.Server Traceback (most recent call last):
<type 'exceptions.IOError'>: [Errno 2] No such file or directory: 'index-altslice.html'

/usr/lib/python2.6/dist-packages/twisted/web/server.py, line 125 in process
123    try:
124      resrc = self.site.getResourceFor(self)
125      self.render(resrc)
126    except:
Self
site	
twisted.web.server.Site instance @ 0x96d45cc
<twisted.web.server.Site instance at 0x96d45cc>
Locals
resrc	
website.dynamicSite instance @ 0x995f74c
<website.dynamicSite instance at 0x995f74c>
self	
twisted.web.server.Request instance @ 0x995f38c
<GET /stats HTTP/1.1>
/usr/lib/python2.6/dist-packages/twisted/web/server.py, line 132 in render
130  def render(self, resrc):
131    try:
132      body = resrc.render(self)
133    except UnsupportedMethod, e:
Locals
resrc	
website.dynamicSite instance @ 0x995f74c
<website.dynamicSite instance at 0x995f74c>
self	
twisted.web.server.Request instance @ 0x995f38c
<GET /stats HTTP/1.1>
Globals
UnsupportedMethod	<class 'twisted.web.error.UnsupportedMethod'>
/usr/lib/python2.6/dist-packages/twisted/web/resource.py, line 210 in render
208      from twisted.web.error import UnsupportedMethod
209      raise UnsupportedMethod(getattr(self, 'allowedMethods', ()))
210    return m(request)
211
Locals
m	<bound method dynamicSite.render_GET of <website.dynamicSite instance at 0x995f74c>>
self	
website.dynamicSite instance @ 0x995f74c
<website.dynamicSite instance at 0x995f74c>
request	
twisted.web.server.Request instance @ 0x995f38c
<GET /stats HTTP/1.1>
/home/user/bithopper/bin/bitHopper/website.py, line 57 in render_GET
55    except:
56     index = index_name
57    file = open(index, 'r')
58    linestring = file.read()
Locals
index	'index-altslice.html'
index_name	'index-altslice.html'
<type 'exceptions.IOError'>: [Errno 2] No such file or directory: 'index-altslice.html'

Code: (what the shell says)

Traceback (most recent call last):
  File "/usr/lib/python2.6/dist-packages/twisted/protocols/basic.py", line 251, in dataReceived
    why = self.lineReceived(line)
  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 1573, in lineReceived
    self.allContentReceived()
  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 1641, in allContentReceived
    req.requestReceived(command, path, version)
  File "/usr/lib/python2.6/dist-packages/twisted/web/http.py", line 807, in requestReceived
    self.process()
--- <exception caught here> ---
  File "/usr/lib/python2.6/dist-packages/twisted/web/server.py", line 125, in process
    self.render(resrc)
  File "/usr/lib/python2.6/dist-packages/twisted/web/server.py", line 132, in render
    body = resrc.render(self)
  File "/usr/lib/python2.6/dist-packages/twisted/web/resource.py", line 210, in render
    return m(request)
  File "/home/user/bithopper/bin/bitHopper/website.py", line 57, in render_GET
    file = open(index, 'r')
exceptions.IOError: [Errno 2] No such file or directory: 'index-altslice.html'

Other than the non-working web-page, the hopper works as intended.

[paraipan]

cirz8 you could try changing to bH dir before launching it, like this:

Code:

cd /home/user/bithopper/bin/bitHopper/ && ./bitHopper.py --scheduler=AltSliceScheduler

absolute path detection it's not enabled for all the bH files like the html pages

[cirz8]

That's what I'm doing since that is the only way to make it work as intended.

Is there something that would prohibit an absolute path detection to be implemented on all files?

[paraipan]

That's what I'm doing since that is the only way to make it work as intended.

Is there something that would prohibit an absolute path detection to be implemented on all files?

hehe nope, nothing and no one prohibits that but atm is not really a priority and this behavior is considered somehow normal for allot of programs and scripts, you're welcome to contribute on this project if you like