drgr33n (OP)
|
|
May 20, 2011, 03:28:00 PM |
|
booted up fine the first time, but now I keep getting this error when I try to run phoenix/phatk [19/05/2011 16:41:06] FATAL kernel error: Failed to load OpenCL kernel!
You haven't accepted the license agreement It worked the very first time I booted but never since (not using persistant at the moment) I also tried to do it manually by cd /opt/AMD-APP-SDK-v2.4-lnx64 tar xvfz icd-registration.tgz -C / but that still gives the same opencl kernal error does /etc/opencl exist ? And you say it worked the first time ? If so it should work again as without persistence its a fresh filesystem every reboot. /etc/opencl does exist, and everytime I reboot it asks me to accept the agreement and I do... yet I still receive the same error, I've even re-imaged my usb drive multiple times the only thing that I can think of is that I may have re-downloaded the iso when you fixed the ssh bug, but I don't have the original to compare with since I over-wrote it Ok can you pastebin the output of Also are you getting the same error using pcolbm ? cool. maybe someone frontpage it to distrowatch.com site ? p.s. actually im more expect Knoppix foring or SLAX, than Debian itself, but thats only good, IMO p.p.s. can only thank you, cuz i have not enough BitCoins for such product :/ lol its still to early to start adding this distro to distrowatch ;: Thanks for suggesting though !! i got everything running on 3 rigs its great. I have a few questions is there anyway to flash the bios on the cards my xfx radeon 5850s wont clock past 775 and id like to hit 800-850~? i saw on page 2 someone posted a script to see the GPU temps with bash this is the code. How do i run or execute it? while true; do aticonfig --adapter=0 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; aticonfig --adapter=0 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; echo $(aticonfig --odgc --adapter=0| grep GPU); echo $(aticonfig --odgc --adapter=1| grep GPU); # Next lines are to check your balance if solo mining # BALANCE=$(bitcoind getbalance) # echo -ne "Bitcoin Balance: ${BALANCE}\r"; sleep 35; clear done thanks for the great OS its really helpful. Great good to hear your liking the new version. That's my code just a little skiddie script to view the temps of your GPU's and your balance if solo mining. Remember to change the device numbers to match your hardware. For eg if I had four cards in one machine it would be while true; do aticonfig --adapter=0 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; aticonfig --adapter=1 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; aticonfig --adapter=2 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; aticonfig --adapter=3 --od-gettemperature | tail -n1 | awk '{print "Current temp: " $5}' ; echo $(aticonfig --odgc --adapter=0| grep GPU); echo $(aticonfig --odgc --adapter=1| grep GPU); echo $(aticonfig --odgc --adapter=2| grep GPU); echo $(aticonfig --odgc --adapter=3| grep GPU); # Next lines are to check your balance if solo mining # BALANCE=$(bitcoind getbalance) # echo -ne "Bitcoin Balance: ${BALANCE}\r"; sleep 35; clear done You could also use leafpad rather than nano if you would rather use a GUI
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 05:11:27 PM |
|
Ok can you pastebin the output of Also are you getting the same error using pcolbm ? I do get the same error with poclbm output of lsmod user@linuxcoin:/$ lsmod Module Size Used by ext3 112218 1 jbd 41698 1 ext3 mbcache 12930 1 ext3 sco 17555 2 bnep 17424 2 rfcomm 32744 0 l2cap 40163 6 bnep,rfcomm crc16 12343 1 l2cap bluetooth 55503 6 sco,bnep,rfcomm,l2cap rfkill 19014 2 bluetooth fuse 61520 1 dm_crypt 22256 0 dm_mod 62467 1 dm_crypt joydev 17138 0 snd_hda_codec_hdmi 26321 1 snd_intel8x0 30802 0 snd_hda_intel 25946 0 fglrx 2678570 88 snd_hda_codec 67647 2 snd_hda_codec_hdmi,snd_hda_intel snd_ac97_codec 106189 1 snd_intel8x0 ac97_bus 12510 1 snd_ac97_codec snd_hwdep 13148 1 snd_hda_codec snd_pcm 67327 5 snd_hda_codec_hdmi,snd_intel8x0,snd_hda_intel,snd_hda_codec,snd_ac97_codec snd_seq 44678 0 snd_timer 22658 2 snd_pcm,snd_seq snd_seq_device 13155 1 snd_seq tpm_tis 13125 0 psmouse 55199 0 snd 52280 10 snd_hda_codec_hdmi,snd_intel8x0,snd_hda_intel,snd_hda_codec,snd_ac97_codec,snd_hwdep,snd_pcm,snd_seq,snd_timer,snd_seq_device edac_core 35344 0 tpm 17756 1 tpm_tis evdev 17475 7 tpm_bios 12903 1 tpm i2c_nforce2 12584 0 k8temp 12531 0 serio_raw 12878 0 pcspkr 12579 0 soundcore 13014 1 snd edac_mce_amd 17103 0 snd_page_alloc 12969 3 snd_intel8x0,snd_hda_intel,snd_pcm i2c_core 23725 1 i2c_nforce2 button 12994 1 fglrx processor 27431 0 squashfs 31002 1 loop 22515 2 aufs 140440 1463 nls_utf8 12456 1 nls_cp437 16553 1 vfat 17165 1 fat 45206 1 vfat sd_mod 35501 3 crc_t10dif 12348 1 sd_mod ide_generic 8217 0 [permanent] ide_core 75593 1 ide_generic usb_storage 43639 2 uas 13151 0 ata_generic 12479 0 ohci_hcd 26400 0 pata_amd 13276 0 sata_nv 26519 0 libata 147240 3 ata_generic,pata_amd,sata_nv ehci_hcd 39529 0 fan 12674 0 scsi_mod 161457 4 sd_mod,usb_storage,uas,libata usbcore 122908 5 usb_storage,uas,ohci_hcd,ehci_hcd thermal 17330 0 thermal_sys 17939 3 processor,fan,thermal skge 40243 0 forcedeth 48684 0 nls_base 12753 5 nls_utf8,nls_cp437,vfat,fat,usbcore
output of lspci | grep VGA user@linuxcoin:/$ lspci | grep VGA 07:00.0 VGA compatible controller: ATI Technologies Inc Hemlock [ATI Radeon HD 5900 Series] user@linuxcoin:/$
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 05:18:51 PM |
|
Ok can you pastebin the output of Also are you getting the same error using pcolbm ? I do get the same error with poclbm output of lsmod user@linuxcoin:/$ lsmod Module Size Used by ext3 112218 1 jbd 41698 1 ext3 mbcache 12930 1 ext3 sco 17555 2 bnep 17424 2 rfcomm 32744 0 l2cap 40163 6 bnep,rfcomm crc16 12343 1 l2cap bluetooth 55503 6 sco,bnep,rfcomm,l2cap rfkill 19014 2 bluetooth fuse 61520 1 dm_crypt 22256 0 dm_mod 62467 1 dm_crypt joydev 17138 0 snd_hda_codec_hdmi 26321 1 snd_intel8x0 30802 0 snd_hda_intel 25946 0 fglrx 2678570 88 snd_hda_codec 67647 2 snd_hda_codec_hdmi,snd_hda_intel snd_ac97_codec 106189 1 snd_intel8x0 ac97_bus 12510 1 snd_ac97_codec snd_hwdep 13148 1 snd_hda_codec snd_pcm 67327 5 snd_hda_codec_hdmi,snd_intel8x0,snd_hda_intel,snd_hda_codec,snd_ac97_codec snd_seq 44678 0 snd_timer 22658 2 snd_pcm,snd_seq snd_seq_device 13155 1 snd_seq tpm_tis 13125 0 psmouse 55199 0 snd 52280 10 snd_hda_codec_hdmi,snd_intel8x0,snd_hda_intel,snd_hda_codec,snd_ac97_codec,snd_hwdep,snd_pcm,snd_seq,snd_timer,snd_seq_device edac_core 35344 0 tpm 17756 1 tpm_tis evdev 17475 7 tpm_bios 12903 1 tpm i2c_nforce2 12584 0 k8temp 12531 0 serio_raw 12878 0 pcspkr 12579 0 soundcore 13014 1 snd edac_mce_amd 17103 0 snd_page_alloc 12969 3 snd_intel8x0,snd_hda_intel,snd_pcm i2c_core 23725 1 i2c_nforce2 button 12994 1 fglrx processor 27431 0 squashfs 31002 1 loop 22515 2 aufs 140440 1463 nls_utf8 12456 1 nls_cp437 16553 1 vfat 17165 1 fat 45206 1 vfat sd_mod 35501 3 crc_t10dif 12348 1 sd_mod ide_generic 8217 0 [permanent] ide_core 75593 1 ide_generic usb_storage 43639 2 uas 13151 0 ata_generic 12479 0 ohci_hcd 26400 0 pata_amd 13276 0 sata_nv 26519 0 libata 147240 3 ata_generic,pata_amd,sata_nv ehci_hcd 39529 0 fan 12674 0 scsi_mod 161457 4 sd_mod,usb_storage,uas,libata usbcore 122908 5 usb_storage,uas,ohci_hcd,ehci_hcd thermal 17330 0 thermal_sys 17939 3 processor,fan,thermal skge 40243 0 forcedeth 48684 0 nls_base 12753 5 nls_utf8,nls_cp437,vfat,fat,usbcore
output of lspci | grep VGA user@linuxcoin:/$ lspci | grep VGA 07:00.0 VGA compatible controller: ATI Technologies Inc Hemlock [ATI Radeon HD 5900 Series] user@linuxcoin:/$
This is very strange your hardware is detected and the kernel modules are loaded ok. Do you have access to your GPU via amdoverdrivectl ? Also when reinstalling to your flash media are you also reformatting your persistent space ? Everything you are saying suggests that there's an error with registration. I've tested on a 5970 and everything seems fine so don't look like a hardware compatibility issue. The only thing I could find about the error you have said is that the registration files do not exist in the right place. Are you using the machine direct or via SSH ?
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 05:40:21 PM |
|
This is very strange your hardware is detected and the kernel modules are loaded ok. Do you have access to your GPU via amdoverdrivectl ? Also when reinstalling to your flash media are you also reformatting your persistent space ?
Everything you are saying suggests that there's an error with registration. I've tested on a 5970 and everything seems fine so don't look like a hardware compatibility issue. The only thing I could find about the error you have said is that the registration files do not exist in the right place.
Are you using the machine direct or via SSH ?
Yes I have access to it in amdoverdrivectrl (previously had to reflash to default bios first for it to work, lost my voltage increase ), and I reformatted the whole drive multiple times to be sure, I am accessing the machine over ssh/x11vnc
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 05:45:53 PM |
|
Aaaahhhhh are you forwarding xorg ?
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 05:50:33 PM |
|
Aaaahhhhh are you forwarding xorg ?
I don't know the technical term/aspects of it, but I'm assuming that is what vnc/x11vnc might be doing?
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 06:01:36 PM |
|
lol well sort of. To access ATI GPGPU's you need control of X on the machine they are installed. ssh -X user@ipaddress password: live cd /opt/miners/phoenix python ./phoenix.py ...
Oh and don't use start_mining to start sessions via SSH, you will have to start the miners manually.
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 06:03:30 PM |
|
lol well sort of. To access ATI GPGPU's you need control of X on the machine they are installed. ssh -X user@ipaddress password: live cd /opt/miners/phoenix python ./phoenix.py ...
I just tried doing it live on the machine with mouse/keyboard and I still get the same error
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 06:06:20 PM |
|
lol well sort of. To access ATI GPGPU's you need control of X on the machine they are installed. ssh -X user@ipaddress password: live cd /opt/miners/phoenix python ./phoenix.py ...
I just tried doing it live on the machine with mouse/keyboard and I still get the same error I don't get what you mean ? you tried that direct ? or you are using SSH on another machine to access your machine ?
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 06:08:29 PM |
|
lol well sort of. To access ATI GPGPU's you need control of X on the machine they are installed. ssh -X user@ipaddress password: live cd /opt/miners/phoenix python ./phoenix.py ...
I just tried doing it live on the machine with mouse/keyboard and I still get the same error I don't get what you mean ? you tried that direct ? or you are using SSH on another machine to access your machine ? I was working with ssh/x11vnc but I went and rebooted the machine (non-persistent) and tried running phoenix/everything locally at the machine (no ssh/vnc) and I still get the same failed to load opencl message
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 06:15:53 PM |
|
lol Ok well there's not a lot more I can do ? I've tested this on the hardware you have and it works fine my end. That error is stating that phoenix cannot access your GPGPU for some reason. Could be a hardware error ? But if you can mine using another OS very doubtful. If you have direct access to the machine do the following. Boot linuxcoin accept license agreement check /etc/OpenCL/vendors/*.icd exists cd over to /opt/miners/poclbm run python poclbm.py If poclbm lists your GPGPU's it might be a problem with phoenix. If not I would say its hardware or something. i will test extensively this evening to see if I can reproduce the errors you seem to be having. Sorry I can't help you more ATM PS: if you could post the full output of poclbm that might be helpful.
|
|
|
|
de4l
Member
Offline
Activity: 82
Merit: 10
|
|
May 20, 2011, 07:40:15 PM |
|
lol Ok well there's not a lot more I can do ? I've tested this on the hardware you have and it works fine my end. That error is stating that phoenix cannot access your GPGPU for some reason. Could be a hardware error ? But if you can mine using another OS very doubtful. If you have direct access to the machine do the following. Boot linuxcoin accept license agreement check /etc/OpenCL/vendors/*.icd exists cd over to /opt/miners/poclbm run python poclbm.py If poclbm lists your GPGPU's it might be a problem with phoenix. If not I would say its hardware or something. i will test extensively this evening to see if I can reproduce the errors you seem to be having. Sorry I can't help you more ATM PS: if you could post the full output of poclbm that might be helpful. the .icd file does exist, poclbm and phoenix both list my card, but still give opencl failure (that's all it gives is that error, none others), I went back to try the linuxcoin 0.1b and I can run the phatk kernal after downloading it just fine, my issue is that I need amdoverdrivecntrl (only in .2, I couldn't get it to work in .1 on my own) to get my memory down to 300, aticonfig won't do it, thanks for the help, still can't figure out why it worked the very first time but no more I startup linuxcoin, click close/ok on agreement, then cd to /opt/miners/phoenix and try running my usual commands, I've tried both using amdoverdrivectrl and not using it, that doesn't seem to be it. Here is my favorite string, but I've tried a bunch with all the same opencl error too python phoenix.py -u http://uname:pass@deepbit.net:8332/ -k phatk VECTORS BFI_INT AGGRESSION=13 WORKSIZE=256 DEVICE=1
|
|
|
|
greenlander
Newbie
Offline
Activity: 28
Merit: 0
|
|
May 20, 2011, 08:05:44 PM Last edit: May 21, 2011, 12:40:57 AM by greenlander |
|
you need 2 partitions : 1st one is vfat and 2nd one is ext3 and in debian based, name must be "live-rw", not "casper-rw" +1 gusti I've been trying to get persistence to work for days and this was the missing data point. The procedure I did that worked was: 1. Put keyfob into a linux box. Use "sudo fdisk /dev/sd-" to nuke all partitions. Set up two partitions. Mark the first partition as bootable. Write the results using "w" in fdisk to write the partition table. (replace "-" with your actual drive) 2. Run "sudo mkfs.ext3 -b 4096 -L live-rw /dev/sd-2" to format the SECOND partition. The name must be "live-rw", not "casper-rw" (replace "-" with your actual drive) 3. Put keyfob into your windows box. A prompt comes up asking if you want to format. Press "format disk" and then "start" then "ok" to allow windows to format the disk. It will format the FIRST partition as fat32. when you hit ok. 4. Use unetbootin-win-549.exe. Click the "diskimage" radio button and point it at the bitcoin iso image. DON'T put anything in the "Space used to preserve files across reboots": leave that at zero. Hit OK to start the transfer. It takes a few minutes. You're all set. Your image should now be persistent. (at least it was for me...)
|
|
|
|
drgr33n (OP)
|
|
May 20, 2011, 10:34:16 PM Last edit: May 21, 2011, 12:49:26 AM by drgr33n |
|
@ de4l you got skype ? PM me your details and I'll screen share with you. Would be nice to figure whats going on. you need 2 partitions : 1st one is vfat and 2nd one is ext3 and in debian based, name must be "live-rw", not "casper-rw" +1 gusti I've been trying to get persistence to work for days and this was the missing data point. The procedure I did that worked was: 1. Put keyfob into a linux box. Use "sudo fdisk /dev/sd*" to nuke all partitions. Set up two partitions. Mark the first partition as bootable. Write the results using "w" in fdisk to write the partition table. 2. Run "sudo mkfs.ext3 -b 4096 -L live-rw /dev/sd*2" to format the SECOND partition. The name must be "live-rw", not "casper-rw" 3. Put keyfob into your windows box. A prompt comes up asking if you want to format. Press "format disk" and then "start" then "ok" to allow windows to format the disk. It will format the FIRST partition as fat32. when you hit ok. 4. Use unetbootin-win-549.exe. Click the "diskimage" radio button and point it at the bitcoin iso image. DON'T put anything in the "Space used to preserve files across reboots": leave that at zero. Hit OK to start the transfer. It takes a few minutes. You're all set. Your image should now be persistent. (at least it was for me...) You don't need a linux machine ! I've said about 5 times lol Partition your drive using whatever OS you like. Just make sure you have two partitions. The first needs to be at least 1G the second can be as large or as small as you like. use unetbootin to install LinuxCoin on your FIRST drive and then reboot into linuxcoin. Open a shell find what your drive has been named. If you only have one HDD on the computer its normally sdb BUT double check with the following command. You should see a list of HDD / Partitions and it should be obvious what one is your USB drive. Now format your second partition with the following command. sudo mkfs.ext3 -b 4096 -L live-rw /dev/sd[yourdrive]2
Reboot and you have persistence I'm looking into why it wont recognize casper-rw won't work using unetbootins method.
|
|
|
|
greenlander
Newbie
Offline
Activity: 28
Merit: 0
|
|
May 21, 2011, 12:44:45 AM |
|
You don't need a linux machine ! I've said about 5 times lol Yeah, I know you said it like five times... but I couldn't figure out how to do it under windows. Windows always wants to create a single partition as soon as you plug in the disk. (I admit that I didn't consider trying to use the built-in disk administrator or any other options.) Linux seems to "like" the idea of multiple partitions on a usb key than windows does. @greenlander could you change / remove these lines from your guide Agreed, done. I am having another problem now. If I boot the system and don't select a boot mode at the first screen and just let it time out on "default" it boots the non-persistent mode. This is true even though the configuration file for the menu indicates that it should boot to persistent mode by default... so I'm trying to figure that out now. drgr33n, I just sent you 3.5 BTC (about $20?) to say thanks for putting this release together.
|
|
|
|
drgr33n (OP)
|
|
May 21, 2011, 12:52:20 AM Last edit: May 21, 2011, 01:07:11 AM by drgr33n |
|
You don't need a linux machine ! I've said about 5 times lol Yeah, I know you said it like five times... but I couldn't figure out how to do it under windows. Windows always wants to create a single partition as soon as you plug in the disk. (I admit that I didn't consider trying to use the built-in disk administrator.) Linux seems to "like" the idea of multiple partitions on a usb key than windows does. @greenlander could you change / remove these lines from your guide Agreed, done. Lol using a windows system can be a little cryptic when you want to setup multiple partitions. I have a wiki setup and will be going live very soon. Also working on documentation for linuxcoin. Thanks bud. I was just worried someone would blindly follow the commands and end up formatting a HDD they didn't intend to EDIT: Yes just received them. thanks greenlander !! if you want to set persistent as the default boot option just add the following line to the beginning of live.cfg. default LinuxCoin_persistent
I think unetbootin uses its own config gimmie two secs and I'll let you know what file you want to edit if you used unetbootin. If you used unetbootin edit syslinux.cfg and change default menu.c32 to default ubnentry2
|
|
|
|
gusti
Legendary
Offline
Activity: 1099
Merit: 1000
|
|
May 21, 2011, 12:58:30 AM |
|
I am having another problem now. If I boot the system and don't select a boot mode at the first screen and just let it time out on "default" it boots the non-persistent mode. This is true even though the configuration file for the menu indicates that it should boot to persistent mode by default... so I'm trying to figure that out now.
you can delete all unused menu options in syslinux.cfg
|
If you don't own the private keys, you don't own the coins.
|
|
|
greenlander
Newbie
Offline
Activity: 28
Merit: 0
|
|
May 21, 2011, 01:12:05 AM |
|
I am having another problem now. If I boot the system and don't select a boot mode at the first screen and just let it time out on "default" it boots the non-persistent mode. This is true even though the configuration file for the menu indicates that it should boot to persistent mode by default... so I'm trying to figure that out now.
you can delete all unused menu options in syslinux.cfg +2 gusti Yeah, that was the ticket. The configuration file I was changing was live.cfg, but that file doesn't seem to do anything. I modified the default option in syslinux.cfg and then it works as intended.
|
|
|
|
greenlander
Newbie
Offline
Activity: 28
Merit: 0
|
|
May 21, 2011, 02:08:42 AM |
|
I'm trying to append to the init sequence to automatically run the miner after the boot is complete.
Do any of you Linux people know in what file I would put a command like that? I imagine that the command would fail if it ran before X was completely up.
|
|
|
|
REF
|
|
May 21, 2011, 02:18:04 AM |
|
is there anyway i get flash the bios of my cards with linuxcoin? Or someway to pass the 775 clock limit on my cards? Iv seen lots of guides for win7. i tried google to but i cant seem to find anything.
|
|
|
|
|