This running-wild behaviour is a well known problem with the somewhat broken OSX OpenCL implementation and the phatk-style miner kernels. No proper solution yet. However, you can use an earlier version of Diablo miner with a different kernel. If you check out the sources, do a "git log" and look for the time the switch to the phatk occurred (iirc Jun 24), and use that one. You can also find an older version packaged in a Mac-friendly app by someone else here on the forums: http://bitcointalk.org/index.php?topic=8994.0This really should be an FAQ entry somewhere... Yeah, but you also lose a ton of important fixes. |
|
|
|
The ghash meter just measures raw ghash produced. If it stops moving, then yes, you have a problem. FPS doesn't sound right unless you're a large -f value in the same ballpark, and the second mhash number doesn't make sense unless you own 2 or more video cards.
I assume its just more OSX insanity. I am not surprised.
It is the original Mac Mini without any hardware changes. (1 video card) How can I help to get rid of the OSX insanity? Or should I get rid of OSX?  Honestly? Don't mine on OSX. Even when it works, it sucks up 40% of your performance. |
|
|
|
I'm running DiabloMiner on a MacMini 10.7 with a ATI Radeon 6630M with the following start parameter: ./DiabloMiner-OSX.sh -u name.macmini -p password -o mining.bitcoin.cz -r 8332 -w 64 -dd After about 15 minutes the system crashes. Are the mhash and ghash values realistic? Last line, before the system crashes: mhash: 235.4/591.2 | a/r/hwe: 46/1/0 | ghash: 2456.8 | fps: 272.2
How can I help to debug the issue? (DiabloMiner works perfectly fine an my old MacMini 10.6.8 that I still have.) Here is the complete output of the DiaboMiner debug log: (Please notice the increasing numbers for mhash and ghash) http://codepaste.net/qhr2esThe ghash meter just measures raw ghash produced. If it stops moving, then yes, you have a problem. FPS doesn't sound right unless you're a large -f value in the same ballpark, and the second mhash number doesn't make sense unless you own 2 or more video cards. I assume its just more OSX insanity. I am not surprised. |
|
|
|
Seems to be a launch4j bug. Tell them to rebuild the binaries for universal. This will happen on any OSX that doesn't have the PPC emulator anymore (such as any 10.7 and I think some 10.6s)
Hi DiabloD3, I will let them know about the bug. Let's see how they react. I tried something else: I used the compilation from my MacBookPro and copied it to the Mac mini and let it run with following result: (name and password replaced) localhost:DiabloMinerComp thomas$ ./DiabloMiner-OSX.sh -u placeholder.placeholder -p placeholder -o mining.bitcoin.cz -r 8332 [8/12/11 8:07:43 AM] Started [8/12/11 8:07:43 AM] Connecting to: http://mining.bitcoin.cz:8332/ [8/12/11 8:07:43 AM] Using Apple OpenCL 1.1 (Jun 25 2011 01:23:15) [8/12/11 8:07:44 AM] Added ATI Radeon HD 6630M (#1) (5 CU, local work size of 256) [8/12/11 8:07:44 AM] ERROR: [CL_INVALID_WORK_GROUP_SIZE] : OpenCL Error : clEnqueueNDRangeKernel failed: local_size[0] (256) != required local_size[0] (1024) [8/12/11 8:07:44 AM] ERROR: Failed to queue kernel, error -54 Known 10.7 bug. Until they fix it, use -w 64. |
|
|
|
I am having errors compiling DiabloMiner on Lion (Mac Mini 2011 i5 CPU) AMD 6630M. I am getting the following errer when running maven: mvn clean install in the DiabloMiner directory. (It works on my MacBook Pro 10.6.8 ) Full debug-log: http://codepaste.net/ziimsi[INFO] launch4j: Launch of "windres" failed: the PowerPC architecture is no longer supported. [ERROR] net.sf.launch4j.BuilderException: net.sf.launch4j.ExecException: Exec failed(2): [Ljava.lang.String;@3a7f1228 at net.sf.launch4j.Builder.build(Builder.java:145) at com.akathist.maven.plugins.launch4j.Launch4jMojo.execute(Launch4jMojo.java:326) at org.apache.maven.plugin.DefaultBuildPluginManager.executeMojo(DefaultBuildPluginManager.java:101) at org.apache.maven.lifecycle.internal.MojoExecutor.execute(MojoExecutor.java:209) at org.apache.maven.lifecycle.internal.MojoExecutor.execute(MojoExecutor.java:153) at org.apache.maven.lifecycle.internal.MojoExecutor.execute(MojoExecutor.java:145) at org.apache.maven.lifecycle.internal.LifecycleModuleBuilder.buildProject(LifecycleModuleBuilder.java:84) at org.apache.maven.lifecycle.internal.LifecycleModuleBuilder.buildProject(LifecycleModuleBuilder.java:59) at org.apache.maven.lifecycle.internal.LifecycleStarter.singleThreadedBuild(LifecycleStarter.java:183) at org.apache.maven.lifecycle.internal.LifecycleStarter.execute(LifecycleStarter.java:161) at org.apache.maven.DefaultMaven.doExecute(DefaultMaven.java:319) at org.apache.maven.DefaultMaven.execute(DefaultMaven.java:156) at org.apache.maven.cli.MavenCli.execute(MavenCli.java:537) at org.apache.maven.cli.MavenCli.doMain(MavenCli.java:196) at org.apache.maven.cli.MavenCli.main(MavenCli.java:141) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) at java.lang.reflect.Method.invoke(Method.java:597) at org.codehaus.plexus.classworlds.launcher.Launcher.launchEnhanced(Launcher.java:290) at org.codehaus.plexus.classworlds.launcher.Launcher.launch(Launcher.java:230) at org.codehaus.plexus.classworlds.launcher.Launcher.mainWithExitCode(Launcher.java:409) at org.codehaus.plexus.classworlds.launcher.Launcher.main(Launcher.java:352) Caused by: net.sf.launch4j.ExecException: Exec failed(2): [Ljava.lang.String;@3a7f1228 at net.sf.launch4j.Util.exec(Util.java:142) at net.sf.launch4j.Cmd.exec(Builder.java:206) at net.sf.launch4j.Builder.build(Builder.java:98) ... 22 more Seems to be a launch4j bug. Tell them to rebuild the binaries for universal. This will happen on any OSX that doesn't have the PPC emulator anymore (such as any 10.7 and I think some 10.6s) |
|
|
|
|
You seem to be the only one posting in the thread.
|
|
|
|
|
Update: Improved error handling more, remove usage of System.exit() and blanket Exception catching, DiabloMiner now passes FindBugs
|
|
|
|
DiabloD3,
Ok, but if this stable pool becomes unreacheable (for example) does your miner move back to one of the previous less-stable pools or stops there waiting for it to become alive again?
Sorry to bother, just to understand its logic.
TIA
spiccioli
Yeah, it doesn't keep track of things like that. It'll normally do a 15 second timeout, fail, and rotate to the next. It is just by virtue of it not rotating off a good pool that it will gravitate towards the more functional pools. |
|
|
|
Hi DiabloD3,
I'm testing latest version, zip file downloaded yesterday (on a linux coin pc), with a three pools config: mineco.in, slush and btcguild.
I've found that sometimes one or even two pools get 'discarded' and from there on it only mines on the remaining one(s).
Is this by design or something to correct?
TIA
spiccioli.
Its by design. Anytime a pool fails to connect during a getwork, that execution thread rotates to the next pool. Which pools execution threads are set to on startup is randomly selected. The client will generally gravitate towards the most stable pool. Ok, I see the logic, but is there a way to force it to try again (after some time) a pool which gives an error (apart from restarting the program)? spiccioli No. Although I've been considering having it change pools after a semi-random timeout, but you risk being sent back to the unstable pool. It really doesn't matter what pool you mine on as long as you're mining, honestly. You'll eventually get the payouts, it just takes a little longer. |
|
|
|
Hi DiabloD3,
I'm testing latest version, zip file downloaded yesterday (on a linux coin pc), with a three pools config: mineco.in, slush and btcguild.
I've found that sometimes one or even two pools get 'discarded' and from there on it only mines on the remaining one(s).
Is this by design or something to correct?
TIA
spiccioli.
Its by design. Anytime a pool fails to connect during a getwork, that execution thread rotates to the next pool. Which pools execution threads are set to on startup is randomly selected. The client will generally gravitate towards the most stable pool. |
|
|
|
In the Bitminersunion.org topic: Just want to let everyone know we are currently at DEF CON !!! we will be back on Sunday and Monday we are going to have a 2-3 hour scheduled down time to move the pool to a DEDICATED 100 MBS FIBEROPTIC LINE!
With our powerhouse server super broad band connection and custom coded pushpool front end
we are going to have a huge week this week, and I am sure that we will have our first block this week as well.
"Powerhouse server super broad band connection" <3 Gratz to the upgrades, but what you're talking about is regular low/medium-range hosting.. Still better then nothing tho! @Jine Why do you have to be such a dick sometimes? You do that often in other pool-topics. Afraid for the smaller pools? You started big? Do you think people will hop over to you when you act like that? Dude, don't be a dick. |
|
|
|
|
Update: Remove rollNTimeExpire variable and use refresh/1000, make LP flush all only flushes works for that pool, support X-Reject-Reason in debug output, unify json/rollNTime/fetch time cross-thread signaling in one object
|
|
|
|
No, tonal uses different glyphs for digits (ie, it doesn't recycle letters), and comes complete with units of measure, time, etc. As I write this, it is 7 Kolumbian 16 at about .T. My keyboard's space key is .8 metertons long.
Am I the only one seeing squares here?  We all see it. Unicode fonts typically don't contain the official symbols for meth and cocaine. |
|
|
|
|
Update: Make roll ntime disabling even more paranoid, change LP timeout back to 10 minutes, re-add force getwork on rejected shares to the new async sendwork code
|
|
|
|
|
Update: Improved error handling, improved roll ntime disabling support, changed LP hard timeout to 5 minutes
|
|
|
|
So basically, use multipool mode on DiabloMiner. Or latest poclbm. Sample batch file: cd C:\poclbm
start /affinity 1 /D"C:\poclbm" /MIN poclbm.exe assman:password@api2.bitcoin.cz:8332#serv1 assman:password@api.bitcoin.cz:8332#serv2 assman:password@uscentral.btcguild.com:8332#serv3 -t 10 -b 1000 -d 0 -v -w 128
start /affinity 1 /D"C:\poclbm" /MIN poclbm.exe assman:password@uscentral.btcguild.com:8332#serv1 assman:password@uswest.btcguild.com:8332#serv2 assman:password@api.bitcoin.cz:8332#serv3 -t 10 -b 1000 -d 1 -v -w 128 That is just a silly script. With DiabloMiner, you just use commas with -l or -u -p -o -r, and it automatically handles it. |
|
|
|
Too many people in one pool very bad for Bitcoin. http://www.bitcoinwatch.com/Many people mine on DeepBit due to it's crunching power well here is a great alternative. Put your even video cards on slush's pool @ http://mining.bitcoin.czPut your odd video cards on BTC Guild's pool @ http://www.btcguild.comMaybe put every third card pointed towards @ http://www.bitcoins.lc/You add the hashing power together and get 2473 + 2084 + 506 = 5063 GH/secBetter for Bitcoin and better for you in so many ways. Such as lower fees or more consistent payouts due to crunching so many blocks at one time. Comments.. So basically, use multipool mode on DiabloMiner. |
|
|
|
I just reinstalled win7 (64) today and downloaded the latest diablo miner, unfortunately this error pops up: [31-7-11 3:05:52] Started
Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 0
at com.diablominer.DiabloMiner.DiabloMiner.execute(DiabloMiner.java:427)
at com.diablominer.DiabloMiner.DiabloMiner.main(DiabloMiner.java:133)
Yay, I didn't catch the error condition where the user doesn't type any pools to access. |
|
|
|
Can i still use Phoenix with Eligius now? is there a list of accepted mining software? You can still use all miners with Eligius. This change should only affect those that have broken implementations of the rollntime extension. Phoenix doesn't support this at all. Since it is a helpful feature, I would recommend switching to poclbm if you can get the same hashrate out of it (it's usually better once you find the right settings). Except poclbm is generally slower than DiabloMiner. Right now I'm seeing over 380 shares sent in the last hour on the ~410MH poclbm miner on my desktop machine. I haven't seen anything near that efficiency with any other client/pool combination. Shares generated is an essentially random value. According to statistics, you should get about 344 shares in an hour at 410 mhash. |
|
|
|
Can i still use Phoenix with Eligius now? is there a list of accepted mining software? You can still use all miners with Eligius. This change should only affect those that have broken implementations of the rollntime extension. Phoenix doesn't support this at all. Since it is a helpful feature, I would recommend switching to poclbm if you can get the same hashrate out of it (it's usually better once you find the right settings). Except poclbm is generally slower than DiabloMiner. |
|
|
|
|
Show Posts
