nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 07:26:06 PM |
|
Your aggression seems like it's too high try this in cgminer or reaper worksize 64 aggression 13 threads_per_gpu 4 (or 2 or 1) sharethreads 18 lookup_gap 2 gpu_thread_concurrency 8192
105~130kH/s - going low ...... already thinking to drop mining LTC ......
|
|
|
|
tacotime
Legendary
Offline
Activity: 1484
Merit: 1005
|
|
July 28, 2012, 07:32:47 PM |
|
i try use reaper v13 beta 4 x64 with that settings : worksize 64 aggression 20 threads_per_gpu 1 lookup_gap 5 gpu_thread_concurrency 8192
its work , but only 240~220 kH/s for HD7950OC i think is too little. my old HD6950 reached ~360 kH/s
but i stil lito to use cgminer i try same worksize , thread_concurrency and lookup_gap but again fail.
Use cgminer and these settings --scrypt -I 19 --shaders 1792 --thread-concurrency 8960 -g 1 Should get 500+ kh/s also try --thread-concurrency of 17920, may be a little faster for the 7xxx series you must always use -g 1 to achieve high intensities
|
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 07:44:44 PM |
|
Use cgminer and these settings --scrypt -I 19 --shaders 1792 --thread-concurrency 8960 -g 1 Should get 500+ kh/s also try --thread-concurrency of 17920, may be a little faster for the 7xxx series you must always use -g 1 to achieve high intensities
---------------------------------------------------------------------------------------- [2012-07-28 22:42:05] Started cgminer 2.5.0 [2012-07-28 22:42:06] Probing for an alive pool [2012-07-28 22:42:06] Long-polling activated for http://ltc.xurious.com:9332/LP [2012-07-28 22:42:28] Maximum buffer memory device 0 supports says 536870912, your scrypt settings come to 0 [2012-07-28 22:42:28] Error -61: clCreateBuffer (padbuffer8), decrease CT or increase LG [2012-07-28 22:42:28] Failed to init GPU thread 0, disabling device 0 [2012-07-28 22:42:28] Restarting the GPU from the menu will not fix this. [2012-07-28 22:42:28] Try restarting cgminer. Press enter to continue: ---------------------------------------------------------------------------------------- same for --thread-concurrency 8960 and 17920
|
|
|
|
tacotime
Legendary
Offline
Activity: 1484
Merit: 1005
|
|
July 28, 2012, 07:50:28 PM |
|
Use cgminer and these settings --scrypt -I 19 --shaders 1792 --thread-concurrency 8960 -g 1 Should get 500+ kh/s also try --thread-concurrency of 17920, may be a little faster for the 7xxx series you must always use -g 1 to achieve high intensities
---------------------------------------------------------------------------------------- [2012-07-28 22:42:05] Started cgminer 2.5.0 [2012-07-28 22:42:06] Probing for an alive pool [2012-07-28 22:42:06] Long-polling activated for http://ltc.xurious.com:9332/LP [2012-07-28 22:42:28] Maximum buffer memory device 0 supports says 536870912, your scrypt settings come to 0 [2012-07-28 22:42:28] Error -61: clCreateBuffer (padbuffer8), decrease CT or increase LG [2012-07-28 22:42:28] Failed to init GPU thread 0, disabling device 0 [2012-07-28 22:42:28] Restarting the GPU from the menu will not fix this. [2012-07-28 22:42:28] Try restarting cgminer. Press enter to continue: ---------------------------------------------------------------------------------------- same for --thread-concurrency 8960 and 17920 Hum, that's strange, that's 5x the number of SPUs which should work. You can try the 7970 setting of --thread-concurrency 10240, I don't know why that wouldn't work. That is the normal error for incorrect buffer sizes though. I wonder why it doesn't work? Also --lookup-gap should always be 2, but that's the default
|
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 07:53:59 PM |
|
same whit --thread-concurrency 10240 ....
|
|
|
|
tacotime
Legendary
Offline
Activity: 1484
Merit: 1005
|
|
July 28, 2012, 07:55:50 PM |
|
same whit --thread-concurrency 10240 .... try --shaders 2048 and --thread-concurrency 10240 I'm starting to think something is wrong with your cgminer build though, these should not give 0 memory values
|
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 07:58:59 PM |
|
try --shaders 2048 and --thread-concurrency 10240 I'm starting to think something is wrong with your cgminer build though, these should not give 0 memory values
-------------------------------------------------------------------------------- [2012-07-28 22:57:09] Started cgminer 2.5.0 [2012-07-28 22:57:09] Probing for an alive pool [2012-07-28 22:57:10] Long-polling activated for http://ltc.xurious.com:9332/LP [2012-07-28 22:57:23] LONGPOLL from pool 0 detected new block [2012-07-28 22:57:33] Pool 0 http://ltc.xurious.com:9332 alive [2012-07-28 22:57:33] Error -5: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel) [2012-07-28 22:57:33] GPU 0 failure, disabling! [2012-07-28 22:57:33] Thread 0 being disabled --------------------------------------------------------------------------------
|
|
|
|
dishwara
Legendary
Offline
Activity: 1855
Merit: 1016
|
|
July 28, 2012, 07:59:22 PM |
|
Ok I found my issue, cgminer.conf was conflicting and it overrides the command line settings! So check your config file and make sure things like work size and intensity etc match what youre attempting on the command line. Currently cranking along at 203kh/s on my 5770 w/intensity 18, working on tuning the 6970 still it overheats way to easy and forces the machine to reboot.
I don't see cgminer.conf in the unziped files on windows. Can point me where i can see cgminer.conf. Thanks.
|
|
|
|
tacotime
Legendary
Offline
Activity: 1484
Merit: 1005
|
|
July 28, 2012, 08:02:31 PM |
|
try --shaders 2048 and --thread-concurrency 10240 I'm starting to think something is wrong with your cgminer build though, these should not give 0 memory values
-------------------------------------------------------------------------------- [2012-07-28 22:57:09] Started cgminer 2.5.0 [2012-07-28 22:57:09] Probing for an alive pool [2012-07-28 22:57:10] Long-polling activated for http://ltc.xurious.com:9332/LP [2012-07-28 22:57:23] LONGPOLL from pool 0 detected new block [2012-07-28 22:57:33] Pool 0 http://ltc.xurious.com:9332 alive [2012-07-28 22:57:33] Error -5: Enqueueing kernel onto command queue. (clEnqueueNDRangeKernel) [2012-07-28 22:57:33] GPU 0 failure, disabling! [2012-07-28 22:57:33] Thread 0 being disabled -------------------------------------------------------------------------------- Yeah, that's what I expected... with reaper and --thread-concurrency of 7168, 8960, or 16000 and number of threads 1 (intensity 19) or 4/5 (intensity 13) should give you at least 400kh/s with a lookup gap of 2 and a worksize of 128 or 256
|
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 08:04:25 PM |
|
I'm starting to think something is wrong with your cgminer build though, these should not give 0 memory values
i use windows build "ckolivas-cgminer-43752ee" i think is newest at the moment
|
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 08:12:58 PM |
|
Yeah, that's what I expected... with reaper and --thread-concurrency of 7168, 8960, or 16000 and number of threads 1 (intensity 19) or 4/5 (intensity 13) should give you at least 400kh/s with a lookup gap of 2 and a worksize of 128 or 256
reaper chrash, chrash, chrash, chrash, chrash, chrash ...... whit : --------------------------------- worksize 256 aggression 19 threads_per_gpu 1 lookup_gap 5gpu_thread_concurrency 7168 --------------------------------- run normal but < 250kH/s
|
|
|
|
tacotime
Legendary
Offline
Activity: 1484
Merit: 1005
|
|
July 28, 2012, 08:14:50 PM |
|
Yeah, that's what I expected... with reaper and --thread-concurrency of 7168, 8960, or 16000 and number of threads 1 (intensity 19) or 4/5 (intensity 13) should give you at least 400kh/s with a lookup gap of 2 and a worksize of 128 or 256
reaper chrash, chrash, chrash, chrash, chrash, chrash ...... whit : --------------------------------- worksize 256 aggression 19 threads_per_gpu 1 lookup_gap 5gpu_thread_concurrency 7168 --------------------------------- run normal but < 250kH/s something is very wrong, lookup_gap 2 should work fine. might want to try uninstalling 12.6 drivers and reinstalling them, and also uninstalling app sdk 2.6 and installing 2.7. other than that i do not know, something is broken.
|
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 09:00:45 PM |
|
something is very wrong, lookup_gap 2 should work fine. might want to try uninstalling 12.6 drivers and reinstalling them, and also uninstalling app sdk 2.6 and installing 2.7. other than that i do not know, something is broken.
exactly the same result crash, crash, crash, crash, crash, crash, crash ... but BTC digging well ~ 530MH/s
|
|
|
|
rahlquist
Newbie
Offline
Activity: 46
Merit: 0
|
|
July 28, 2012, 09:45:41 PM |
|
Ok I found my issue, cgminer.conf was conflicting and it overrides the command line settings! So check your config file and make sure things like work size and intensity etc match what youre attempting on the command line. Currently cranking along at 203kh/s on my 5770 w/intensity 18, working on tuning the 6970 still it overheats way to easy and forces the machine to reboot.
I don't see cgminer.conf in the unziped files on windows. Can point me where i can see cgminer.conf. Thanks. If you have one it will be in the same directory as the cgminer.exe most likely. If you dont have one then this isnt whats holding you back.
|
|
|
|
rahlquist
Newbie
Offline
Activity: 46
Merit: 0
|
|
July 28, 2012, 09:46:44 PM |
|
something is very wrong, lookup_gap 2 should work fine. might want to try uninstalling 12.6 drivers and reinstalling them, and also uninstalling app sdk 2.6 and installing 2.7. other than that i do not know, something is broken.
exactly the same result crash, crash, crash, crash, crash, crash, crash ... but BTC digging well ~ 530MH/s Ok its a stupid question but I havent seen it asked. You are shutting down BTC when testing for LTC right?
|
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 28, 2012, 09:52:46 PM |
|
Ok its a stupid question but I havent seen it asked. You are shutting down BTC when testing for LTC right?
of course, also restart PC when i change digging currency.
|
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 29, 2012, 08:52:06 AM |
|
Come on guys, if not you no one can't help, give ideas
|
|
|
|
-ck
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
July 29, 2012, 09:03:36 AM |
|
Come on guys, if not you no one can't help, give ideas Start it with just your login details and: --shaders 1792 -T -D --verbose -g 1 and see what it reports.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
nitrox
Member
Offline
Activity: 136
Merit: 10
tester
|
|
July 29, 2012, 09:21:21 AM |
|
Come on guys, if not you no one can't help, give ideas Start it with just your login details and: --shaders 1792 -T -D --verbose -g 1 and see what it reports. here is log file: http://www.uploadmb.com/dw.php?id=1343553631
|
|
|
|
-ck
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
July 29, 2012, 09:27:59 AM |
|
Everything looks in order there, but it's still failing. How much system ram do you have? Try forcing smaller multiples of 1792 for thread concurrency. Eg --thread-concurrency 1792 , then 3584, then 5376.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
|