Phoenix - Efficient, fast, modular miner

[jedi95]

I'm getting "Failed to connect, retrying..." errors. I have python-twisted installed and my network connection is fine. Any ideas?

Turns out this was only happening with slushs pool, I switched to deepbit and it is working fine.

I'm actually having the same problem on one of my rigs.  For some reason it will not connect to slush's pool (tried multiple login credentials), but will connect to bitcoinpool and deepbit just fine.

I just tested with:

phoenix.py -u http://jedi95.worker:password@mining.bitcoin.cz:8332 -k poclbm FASTLOOP VECTORS BFI_INT PLATFORM=1 DEVICE=0 AGGRESSION=8

It works fine for me. What command line are you using?

[ghost]

I'm getting "Failed to connect, retrying..." errors. I have python-twisted installed and my network connection is fine. Any ideas?

Turns out this was only happening with slushs pool, I switched to deepbit and it is working fine.

I'm actually having the same problem on one of my rigs.  For some reason it will not connect to slush's pool (tried multiple login credentials), but will connect to bitcoinpool and deepbit just fine.

What's odd is I have two identical computers. One will connect to slushes just fine, the other won't. I've tried using different workers and even creating a new worker on slushes website.

Here is the command I am running:
phoenix.py -u http://username.worker:password@mining.bitcoin.cz:8332 -k poclbm DEVICE=0 VECTORS=on AGGRESSION=10 WORKSIZE=64 FASTLOOP BFI_INT

[CFSworks]

That's a nice ~8% increase.  The only bummer is that I still get 100% CPU usage.

ATI's implementation of OpenCL (but not nVidia's if I recall...) on Windows likes to use the CPU a lot,
for some reason. This is not a problem with any miner. You'll probably see 100% CPU use on any OpenCL app
you try to run on that card.

The next pair would use 'platform=1' and then the next pair would use 'platform=2'.  Hopefully somebody can confirm this or correct it.

Speaking of ATI's implementation vs. nVidia's implementation, that's how you switch between them.
If you have multiple "platforms" installed, (which usually only happens when you have both nVidia and ATI cards in the same system),
Phoenix will require you to tell it which platform you want to use.
If it's not yelling at you to select a platform, you can get away with not setting it. To Phoenix (and the rest of the OS in general), a 5970
is no different from putting two 5870s in your system, so if you were to run a dual-5970 rig, you would address all 4 GPUs with DEVICE=0...3

Confusing, does OS and/or server location affect the hashrate?

I'm not sure what you mean about "server location" - where on the mining rig you extracted Phoenix? No, there's no reason that would affect hashrate.
Or did you mean the distance between your mining rig and the pool server? That shouldn't affect hashrate either, since Phoenix queues work.
OS does affect hashrate. It seems ATI Stream runs faster on Linux than elsewhere.

I just tested with:

phoenix.py -u http://jedi95.worker:password@mining.bitcoin.cz:8332 -k poclbm FASTLOOP VECTORS BFI_INT PLATFORM=1 DEVICE=0 AGGRESSION=8

It works fine for me. What command line are you using?

Don't forget to include an askrate if you're on any [RPC] server! (You are fine excluding it on [MMP] and [RPC (+LP)] servers)

Something to note about how Phoenix parses URLs: You must include the trailing / if you want to use an askrate.
Why?
Phoenix sees http://mining.bitcoin.cz:8332;askrate=10 as this:
Hostname: mining.bitcoin.cz
Port: 8332;askrate=10
Path: NONE
Parameters: NONE
Will have problems loading the page.

Phoenix sees http://mining.bitcoin.cz:8332/;askrate=10 as this:
Hostname: mining.bitcoin.cz
Port: 8332
Path: /
Parameters: ;askrate=10
Will work correctly.

This is a problem with how urlparse parses the URL internally and I'm thinking of a way to better handle this now.

[Cryptoman]

I'm having some problems getting phoenix to run under OpenSUSE 11.3.  First I got this error:

Code:

ImportError: cannot import name IBodyProducer

It turns out that the version of twisted in the OpenSUSE repository is only 8.2.    Installing twisted 11.0 fixes that problem.  However, now I get:

Code:

[26/04/2011 12:36:06] FATAL kernel error: Failed to load OpenCL kernel!

I've been running poclbm for months, so I obviously have OpenCL.  Do I need a symlink?

Here's my command line:

Code:

./phoenix.py -u http://user:pass@127.0.0.1:8332 -k poclbm DEVICE=0 VECTORS AGGRESSION=7 FASTLOOP BFI_INT WORKSIZE=64

[EDIT] Solved.  I was using the wrong device number.  DEVICE=0 refers to the CPU.

[mjsbuddha]

has anyone figure out how to connect this to deepbit? -u http://address@site.com.minername:password@deepbit.net:8332 doesnt work because it thinks my login is address@site and my miner is com.minername.

[proudhon]

has anyone figure out how to connect this to deepbit? -u http://address@site.com.minername:password@deepbit.net:8332 doesnt work because it thinks my login is address@site and my miner is com.minername.

You need to use this:

http://address@site.com:password@deepbit.net:8332

[mjsbuddha]

has anyone figure out how to connect this to deepbit? -u http://address@site.com.minername:password@deepbit.net:8332 doesnt work because it thinks my login is address@site and my miner is com.minername.

You need to use this:

http://address@site.com:password@deepbit.net:8332

yeah i know that works but you cant specify a miner that way. there's no way to connect and specify a miner? I have 8 GPU's and I'd like to know which is doing what.

[eleuthria]

For deepbit, use the email address with the underscore for the worker (so:  email_address@blah.com_1, _2, _3, etc).  There is no '.miner' required on deepbit.

[mjsbuddha]

For deepbit, use the email address with the underscore for the worker (so:  email_address@blah.com_1, _2, _3, etc).  There is no '.miner' required on deepbit.

ah. perfect. thanks

[sniper_sniperson]

What's exactly doing PLATFORM=ID option?

[CFSworks]

What's exactly doing PLATFORM=ID option?

Short answer:
The PLATFORMS=ID switch is to specify which platform ID on your system you want to use.
It's only important if you have multiple OpenCL devices in your system, from different vendors (mixing AMD/ATI and
nVidia cards in one system, for example.)
If you only have one platform on your system, you don't have to specify it. If you have multiple platforms,
Phoenix will tell you that it is needed. So, don't worry about it.

Long answer:
OpenCL is a standard developed by Khronos.
The "core" of OpenCL was not created by AMD/ATI nor nVidia, although they probably contribute to its development.
The problem is that, even though OpenCL is common, AMD/ATI and nVidia have different ideas about how to go about
bridging the gap between OpenCL and GPU. AMD/ATI gets from OpenCL to raw GPU code in a different way from how
nVidia does it, which makes nVidia's version of OpenCL totally incompatible with AMD/ATI cards and vice versa.

This can cause a serious problem when there is a system with some nVidia and some AMD/ATI cards in it: One vendor's
version of OpenCL will not work with another vendor's card, so Khronos needed a way of making sure independent
versions could exist on the same system. Their solution was to use an "installable client driver" (ICD) for each vendor.
So, when you install both AMD/ATI's OpenCL and nVidia's OpenCL, they exist as separate "platforms." One accesses
the AMD/ATI cards, the other accesses the nVidia cards.

[foxmulder]

Great miner, donated   .

[JWU42]

Seeing many more rejected shares than with poclbm - averaging around 2% over the past 24 hours.  This is on a 5970, 6950 and 5770 (a good mix of hardware).

One thing I found is 2 (and up to 4) consecutive rejected shares.  It seems this happens within about a 10-15 second time frame.

[travex]

I've just realized that with phoenix miner, I dont even have to turn my bitcoin client on ! Lol ?

[jedi95]

Seeing many more rejected shares than with poclbm - averaging around 2% over the past 24 hours.  This is on a 5970, 6950 and 5770 (a good mix of hardware).

One thing I found is 2 (and up to 4) consecutive rejected shares.  It seems this happens within about a 10-15 second time frame.

Poclbm clears its result queue when work is received from a new block. Phoenix uses Twisted to send results asynchronously, which makes this difficult. No computation time is wasted, it's just that Phoenix doesn't cancel sending already found nonces like poclbm.

This is something we will fix if we can find a good solution to it.

[marcus_of_augustus]

Good work on the miner guys ... donation coming your way.

PS: I don't know to love you or hate you for single-handedly making the network hashrate 10-15% bigger.

[shivansps]

using  -v -u http://pooluser@miner:pass@btcmine.com:8332 -k poclbm device=0 WORKSIZE=128 VECTORS AGGRESSION=8 FASTLOOP BFI_INT

HD6870 from 238 to 271!
HD6850 from 211 to 233!

Excelent job! interface lags like hell... but i can hadle it.

[zzzd]

are rejected the same as stale shares?

[Raulo]

are rejected the same as stale shares?

Rejected shares are either stale or invalid. But unless you have broken hardware (or overclocked too much) or buggy software which would result in invalid solutions, rejected ones are stale.

[BitLex]

i get this message everytime i start phoenix on ubuntu:

Code:

/home/noodles/phoenix-1.2/KernelInterface.py:139: DeprecationWarning: struct integer overflow masking is deprecated
  hashInput = pack('>76sI', staticData, nonce)
/home/noodles/phoenix-1.2/KernelInterface.py:148: DeprecationWarning: struct integer overflow masking is deprecated
  formattedResult = pack('<76sI', range.unit.data[:76], nonce)

it just spits out that warning and starts to work anyway,

but from time to time, a miner just stops after work queue is empty, like it did about 1hour ago:

Code:

[27/04/2011 00:20:39] Result: 83228c5b accepted             
[27/04/2011 00:20:39] Warning: work queue empty, miner is idle

and i have to restart it (and again get the warning shown above)