Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 12, 2014, 11:18:07 PM |
|
Is that it? Mem: 16400K used, 12956K free, 0K shrd, 1724K buff, 4620K cached CPU: 34% usr 26% sys 4% nic 33% idle 0% io 0% irq 0% sirq Load average: 2.65 2.22 2.09 8/65 2079 PID PPID USER STAT VSZ %VSZ %CPU COMMAND 760 752 root S 75220 256% 65% /usr/bin/cgminer --hexmineru-frequenc
Indeed
|
|
|
|
Lucko
|
|
March 13, 2014, 12:49:43 AM |
|
Do boards stop mining from time to time? And then start again... First I thought it is because of too many boards on one router but it happens with only 2 on it too. Is there a setting I can change or something?
|
|
|
|
Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 13, 2014, 01:17:06 AM |
|
Do boards stop mining from time to time? And then start again... First I thought it is because of too many boards on one router but it happens with only 2 on it too. Is there a setting I can change or something?
No, that should not happen. I only had one board so i am not sure what is the cause.
|
|
|
|
Lucko
|
|
March 13, 2014, 02:50:44 AM |
|
This is what happens. On some boards more offend then others:
001: HEX8 1: 260 1000/ 998mV | 10.62G/282.0Gh/s | A:158720 R:1034/0.65% HW:0/0.00% WU: 3.8/m
Sometimes it goes down to less then 1Gh...
|
|
|
|
Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 13, 2014, 03:03:45 AM |
|
This is what happens. On some boards more offend then others:
001: HEX8 1: 260 1000/ 998mV | 10.62G/282.0Gh/s | A:158720 R:1034/0.65% HW:0/0.00% WU: 3.8/m
Sometimes it goes down to less then 1Gh...
There is a variation on hashrate but i never see 10 GH/s.
|
|
|
|
Lucko
|
|
March 13, 2014, 03:25:34 AM |
|
I don't think it is just that
000: HEX8 0: 260 1000/1003mV | 5.871G/252.0Gh/s | A:274432 R:2064/0.75% HW:0/0.00% WU: 3.6/m
Same setting but it happens more offer and I get 30Gh less out...
EDIT: Less then 1GH 000: HEX8 0: 260 1000/1004mV | 553.4M/249.1Gh/s | A:279552 R:2064/0.73% HW:0/0.00% WU: 3.6/m
|
|
|
|
iFA88
|
|
March 13, 2014, 09:09:14 AM |
|
Thank you. When i plug 5 hex8 miner to the RPI, then the CPU usage goes up to 99%. I dont have any solution for this, i use now 2 rpi's. Is that it? Mem: 16400K used, 12956K free, 0K shrd, 1724K buff, 4620K cached CPU: 34% usr 26% sys 4% nic 33% idle 0% io 0% irq 0% sirq Load average: 2.65 2.22 2.09 8/65 2079 PID PPID USER STAT VSZ %VSZ %CPU COMMAND 760 752 root S 75220 256% 65% /usr/bin/cgminer --hexmineru-frequenc
|
|
|
|
Lucko
|
|
March 13, 2014, 09:15:51 AM Last edit: March 13, 2014, 09:39:42 AM by Lucko |
|
Is it running batter with them or is there another reason for using it. I think this drops might be router related...
EDIT: what is that something set to 0 or something like that I remember reading about but can't remember right now and do search.
|
|
|
|
Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 13, 2014, 09:43:57 AM |
|
--hexminer8-set-diff-to-one 1 --hexminer8-chip-mask 255 --hexminer8-voltage 955 --hexminer8-options 8:240 One by one --hexminer8-set-diff-to-one 1/0 the diff when set to 1 like mine - am testing it now brings a lot pressure between UB-HOST and PIC/ASIC communication Coin crafts can filter their response based on difficulty configured. When you set to one you got all nonces match diff 1 which has to be filtered and not send to the pool amusing you mine at diff greater than one. The only real advantage setting it to one is the cgminer 5 secs stats not moving up/down like crazy and average stats are settling down more quickly. Assume that you mine at diff 128. Cgminer sets it individually for each work passed to asic so it stays in sync with your pool always. Even if it is 1! When you set it to zero - ASIC send to the pic only nonces matching the diff 128 and back to host so you have about 128 times less data transfer, and cgminer shorts the time for submitting the nonce not checking against pool diff if it is a HW and all nasty stuff inside. the only disadvantage is that each nonce is counted not as one like before but as work difficulty itself which causes the stats variation. I suggest to use --hexminer8-set-diff-to-one 0 and configure your worker to mine at some constant higher difficulty. for instance when btcguild changes difficulty some old work may become stale oldddiff even though it is valid. In general if you happen to be in between two pools difficulties and your pool changing it up and down frequently you can have higher pool reject ratio - not good --hexminer8-chip-mask 255 - this was for development you can enable/disable chips based on binary mask bin 11111111 - 255 dec all chips enabled. Obviously zero is not a good choice here --hexminer8-voltage 955 the core voltage in milivolts. Do not go above one volt 1000 or do it on your risk --hexminer8-options 8:240 8 is chip count - used for some internal usb timing calculations , 240 is the frequency. The actual chip frequency is multiplied by four so 240 means 880 actually. the maximum value as far as know is 275. But try to stay below 260 in general And finally use --quiet when you have seen enough of course
|
|
|
|
askhimthatsoon
Newbie
Offline
Activity: 15
Merit: 0
|
|
March 13, 2014, 10:12:59 AM |
|
Put some load on the 5v line (old cd-rom/hdd) and your PSU should run without shutting down. I also use very cheap PSUs without any problem.
Thanks by the way. Tried it and it did stabilize things quite a bit. Highly recommended.
|
|
|
|
loshia
Legendary
Offline
Activity: 1610
Merit: 1000
|
|
March 13, 2014, 11:38:26 AM |
|
I don't think it is just that
000: HEX8 0: 260 1000/1003mV | 5.871G/252.0Gh/s | A:274432 R:2064/0.75% HW:0/0.00% WU: 3.6/m
Same setting but it happens more offer and I get 30Gh less out...
EDIT: Less then 1GH 000: HEX8 0: 260 1000/1004mV | 553.4M/249.1Gh/s | A:279552 R:2064/0.73% HW:0/0.00% WU: 3.6/m
Thanks Zish Read Change log dude. To avoid speed variation use set-diff-to-one 1 Lucko Read about variation +-10-15 % is fine Watch your pool graps and do your math Hash rate = 8 chips * 32 cores (Amusing all are good check cgminer api) * clock * 4 MINUS HW err% MINUS pool REJECT %
|
|
|
|
Gator-hex
|
|
March 13, 2014, 11:39:34 AM Last edit: March 13, 2014, 10:49:23 PM by Gator-hex |
|
There is a huge variation but it's in both direction over 5s average, but it evens out, the second number is the true average and it will settle down after a few days running. Set diff to 1 while you're tuning Freq/mV settings. Like other people said, put something on the PSU's 5V line, if you have powering down issues.
|
|
|
|
Lucko
|
|
March 13, 2014, 10:34:23 PM |
|
--hexminer8-set-diff-to-one 1 --hexminer8-chip-mask 255 --hexminer8-voltage 955 --hexminer8-options 8:240 One by one --hexminer8-set-diff-to-one 1/0 the diff when set to 1 like mine - am testing it now brings a lot pressure between UB-HOST and PIC/ASIC communication Coin crafts can filter their response based on difficulty configured. When you set to one you got all nonces match diff 1 which has to be filtered and not send to the pool amusing you mine at diff greater than one. The only real advantage setting it to one is the cgminer 5 secs stats not moving up/down like crazy and average stats are settling down more quickly. Assume that you mine at diff 128. Cgminer sets it individually for each work passed to asic so it stays in sync with your pool always. Even if it is 1! When you set it to zero - ASIC send to the pic only nonces matching the diff 128 and back to host so you have about 128 times less data transfer, and cgminer shorts the time for submitting the nonce not checking against pool diff if it is a HW and all nasty stuff inside. the only disadvantage is that each nonce is counted not as one like before but as work difficulty itself which causes the stats variation. I suggest to use --hexminer8-set-diff-to-one 0 and configure your worker to mine at some constant higher difficulty. for instance when btcguild changes difficulty some old work may become stale oldddiff even though it is valid. In general if you happen to be in between two pools difficulties and your pool changing it up and down frequently you can have higher pool reject ratio - not good --hexminer8-chip-mask 255 - this was for development you can enable/disable chips based on binary mask bin 11111111 - 255 dec all chips enabled. Obviously zero is not a good choice here --hexminer8-voltage 955 the core voltage in milivolts. Do not go above one volt 1000 or do it on your risk --hexminer8-options 8:240 8 is chip count - used for some internal usb timing calculations , 240 is the frequency. The actual chip frequency is multiplied by four so 240 means 880 actually. the maximum value as far as know is 275. But try to stay below 260 in general And finally use --quiet when you have seen enough of course Thanks now I see why I see this big differences and that low hashrate. I have difficulty set to 1K... So it is normal from time to time that board doesn't find anything for a lot of time...
|
|
|
|
Lucko
|
|
March 14, 2014, 07:21:02 PM Last edit: March 14, 2014, 07:52:20 PM by Lucko |
|
Must say that I like this boards... Is there a way you can set voltage on per board bases from router? Something like you could do on GPUs And another question. Setting 1000 mV and 260 MHz. I guess you can drop voltage if the board will work right. Since it works with 240/900 I guess there is no need for full 1V... This is just safe default right? It just sound strange to me that I didn't see any 260/970 settings or something like that Thanks... I plan to write a positive review later. marto74 really know what he is doing... The only negative I can find is the noise that 4 chip board will solved... EDIT: Green, Orange, Red leds... What do they mean?
|
|
|
|
2GOOD
|
|
March 14, 2014, 08:42:48 PM |
|
|
|
|
|
QuiveringGibbage
|
|
March 15, 2014, 08:07:18 AM |
|
5 x units (end Feb batch) arrived Monday, 10th March 2014 in Sydney, Australia. 2 x units DOA but got the other 3 hashing at: 000: HEX8 0: 220 860/ 873mV | 396.4G/223.0Gh/s | A:8710259 R:6248/0.07% HW:522/1.08% WU: 17.0/m 000: HEX8 0: 250 950/ 961mV | 298.0G/232.3Gh/s | A:9033242 R:7328/0.08% HW:1228/2.43% WU: 17.7/m 000: HEX8 0: 250 960/ 971mV | 195.3G/251.4Gh/s | A:9743945 R:5514/0.06% HW:576/1.07% WU: 19.2/m
[13:46:23] *ozbot* | bunnyhopping | 30401698 shares (134 stales; 0.0%) | avg 719753 MH/s |
Pics can be found here: http://imgur.com/a/mRbMZ#bmfxPCChey marto74, what's the chances of getting something (say.. extra h/w) for the DOA and $500AUD to DHL the damn things back to TechnoBit? did I mention the lost shares from half a THs of hashing power while I go through this RMA process? Cheers, QG
|
Bitcoin is at the tippity top of the mountain...but it's really only half way up..
|
|
|
asdasd123
Newbie
Offline
Activity: 2
Merit: 0
|
|
March 15, 2014, 12:58:05 PM |
|
does anyone have tutorial on how to set up miners with TP link i have .0. IP and i cant access TP link on ip 192.168.1.50
|
|
|
|
Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 15, 2014, 01:10:11 PM |
|
does anyone have tutorial on how to set up miners with TP link i have .0. IP and i cant access TP link on ip 192.168.1.50
Yes. You need to set your computer IP with same prefix with TL-MR3020, example 192.168.1.10. For more information read here : https://bitcointalk.org/index.php?topic=476326.0
|
|
|
|
Gator-hex
|
|
March 15, 2014, 06:21:00 PM |
|
5 x units (end Feb batch) arrived Monday, 10th March 2014 in Sydney, Australia. 2 x units DOA but got the other 3 hashing at: 000: HEX8 0: 220 860/ 873mV | 396.4G/223.0Gh/s | A:8710259 R:6248/0.07% HW:522/1.08% WU: 17.0/m 000: HEX8 0: 250 950/ 961mV | 298.0G/232.3Gh/s | A:9033242 R:7328/0.08% HW:1228/2.43% WU: 17.7/m 000: HEX8 0: 250 960/ 971mV | 195.3G/251.4Gh/s | A:9743945 R:5514/0.06% HW:576/1.07% WU: 19.2/m
[13:46:23] *ozbot* | bunnyhopping | 30401698 shares (134 stales; 0.0%) | avg 719753 MH/s |
Pics can be found here: http://imgur.com/a/mRbMZ#bmfxPCChey marto74, what's the chances of getting something (say.. extra h/w) for the DOA and $500AUD to DHL the damn things back to TechnoBit? did I mention the lost shares from half a THs of hashing power while I go through this RMA process? Cheers, QG Sometimes DOA can just mean the USB leaks power and your power protected USB hub wont let your PC talk to it. It's always worth trying it directly plugged into a TP-Link TL-3020 (running the Technobit firmware), bypassing the USB hub, to check if it works. (risky to do this with a PC/Laptop! but it won't break the TP-Link) If it does work, you can solve it by taping out the ground or +5v connector in the USB cable, then it should work back on the power protected hub. USB1 hubs/PC ports can cause problems too as they are not fast enough and some power supplies have issues if there is nothing plugged into the 5V line.
|
|
|
|
Zich
Legendary
Offline
Activity: 1190
Merit: 1000
|
|
March 16, 2014, 07:32:13 AM |
|
Finally found sweet spot Already run rock solid for almost 4 days.
|
|
|
|
|