GekkoScience NewPac / Terminus R606 (BM1387) Official Support Thread - page 94.

goldsounds

newbie

Activity: 14

Merit: 1

im having so much trouble with this.
I have sidehacks usb port and 3 miners and i cant get them running well at all. If i just do one it runs fine, the best at 350mhz (about 65gh/s) any other speed resets and no nonce errors,

i have tried --gekko-noboost, setting the freq, taking it away. I have not turned or moved any of the voltage.

I have tried dec 18 build and this new one to test

auto tune takes me to low eff mode

any advice would be great!

Code:

 (5s):22.85G (1m):24.90G (5m):20.77G (15m):18.21G (avg):15.24Gh/s
 A:7680  R:0  HW:30  WU:213.0/m | ST: 1  SS: 0  NB: 8  LW: 24230  GF: 0  RF: 0
 Connected to us-east.stratum.slushpool.com diff 512 with stratum as user xxxxxxx.worker1
 Block: 56a83cb4...  Diff:6.39T  Started: [19:48:11.093]  Best share: 13.9K
----------------------------------------------------------------------------------------------------
 [U]SB management [P]ool management [S]ettings [D]isplay options [Q]uit
 0: GSH 10031593: BM1387:02  125.00MHz [--] |  29.4% WU:^16% | 11.77G / 4.417Gh/s WU:61.6/m A:1536 R:0 HW: 7
 1: GSH 10031670: BM1387:02  68.75MHz [--] |  68.6% WU:^42%  | 11.36G / 6.517Gh/s WU:90.9/m A:4096 R:0 HW:10
 2: GSH 10031678: BM1387:02  106.25MHz [--] |  41.4% WU:^18% | 18.75G / 4.348Gh/s WU:60.5/m A:2048 R:0 HW:13
----------------------------------------------------------------------------------------------------
 [2019-04-10 19:42:08.294] 2: GSH 2 - no nonce: target frequency 143.75MHz -> 137.50MHz
 [2019-04-10 19:42:32.780] 0: GSH 0 - no nonce: target frequency 156.25MHz -> 150.00MHz
 [2019-04-10 19:43:00.407] 2: GSH 2 - no nonce: target frequency 137.50MHz -> 131.25MHz
 [2019-04-10 19:43:19.137] 0: GSH 0 - no nonce: target frequency 150.00MHz -> 143.75MHz
 [2019-04-10 19:43:56.650] 2: GSH 2 - no nonce: target frequency 131.25MHz -> 125.00MHz
 [2019-04-10 19:44:49.247] 2: GSH 2 - no nonce: target frequency 125.00MHz -> 118.75MHz
 [2019-04-10 19:44:56.526] 0: GSH 0 - no nonce: target frequency 143.75MHz -> 137.50MHz
 [2019-04-10 19:45:43.951] 2: GSH 2 - no nonce: target frequency 118.75MHz -> 112.50MHz
 [2019-04-10 19:45:53.836] Accepted 5abf4528 Diff 722/512 GSH 0
 [2019-04-10 19:46:41.328] 1: GSH 1 - low eff: (1m)55.1 (5m)54.1 (15m)46.4 (WU)36.9  - [90.0]
 [2019-04-10 19:46:41.331] 1: GSH 1 - low eff: target frequency 75.00MHz -> 68.75MHz
 [2019-04-10 19:47:20.297] 2: GSH 2 - no nonce: target frequency 112.50MHz -> 106.25MHz
 [2019-04-10 19:47:29.459] Accepted 4709c247 Diff 923/512 GSH 0
 [2019-04-10 19:47:36.435] 0: GSH 0 - no nonce: target frequency 137.50MHz -> 131.25MHz
 [2019-04-10 19:48:02.402] Accepted 04b55895 Diff 13.9K/512 GSH 1
 [2019-04-10 19:48:11.093] Stratum from pool 0 detected new block at height 571106
 [2019-04-10 19:48:22.693] 0: GSH 0 - no nonce: target frequency 131.25MHz -> 125.00MHz

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: vh on April 10, 2019, 07:22:39 AM

@Kano

Over in the r606 branch.

OK thanks found it

Quote from: vh on April 10, 2019, 07:22:39 AM

@all
A running tally of the pools and its asic-boost condition is at the bottom of This initial post.
The top "okay" section are pools that I've confirmed to accept work and nonces from the extra ab midstate(s).

If you spot something different, using a different combination or set of settings + pools + local environment.
Post as much detail to replicate as you can and I'll continue to recheck/investigate.

Well the easiest way to see if it is responding in the standard AB manner or not is to catch a share, on the wire, being sent to the pool.
AB shares include the extra block version mask.

A non-AB looks like:

Code:

{"params": ["Kano_ftw", "5cae4ce10000000b", "0200000000000000", "5cae4dfd", "71b3a80d"], "id": 5, "method": "mining.submit"}

AB includes the mask in the parameters:

Code:

{"params": ["Kano_ftw", "5cae4ce10000000b", "0200000000000000", "5cae4dfd", "71b3a80d", "00c00000"], "id": 5, "method": "mining.submit"}

i.e. the extra 6th field of 8 hex "00c00000" inside the brackets [] before the "id"

Shares with the default mask are "00000000"
If it generates AB shares, some will have other values like above "00c00000"

gt_addict

hero member

Activity: 2534

Merit: 623

Quote from: vh on April 10, 2019, 12:19:41 PM

You're good. No need to adjust the miner itself.

I have enough from your description to make improvements.
I'll move the software towards a balance between the working build and the R606 line.

Use "--gekko-noboost" for the time being, or the older build.

Would be awesome if you can circle back on the next build and confirm the issue is no more.

No worries they are happily hashing away at 525m now (which was an impossibility on the older build) and still getting 97%. As soon as you have a revised build I will test it out and report back. Sorry my little setup is causing issues Grin

WilcoWi

member

Activity: 88

Merit: 85

Mining on a pool that is not AB compatible, but having no problems at all. Win10 here
This build is working very good and stable.

btw am not using the "--gekko-noboost" option

vh

hero member

Activity: 699

Merit: 666

You're good. No need to adjust the miner itself.

I have enough from your description to make improvements.
I'll move the software towards a balance between the working build and the R606 line.

Use "--gekko-noboost" for the time being, or the older build.

Would be awesome if you can circle back on the next build and confirm the issue is no more.

gt_addict

hero member

Activity: 2534

Merit: 623

VH, its odd that its only happened the once for you. I will give it another go when I’m home later with AB enabled and see what happens. Not like I’m losing out on loads of btc Cheesy

EDIT:

Well I tried again and still the same. Even tried the US server but to no avail. Im happy to not use AB seeing as its working without it. I may well try the "old school" method of turning the pot down to reduce the consumption and up the freq Grin

vh

hero member

Activity: 699

Merit: 666

@Kano

Over in the r606 branch.

@gt_addict

Got one of my test system set up to mimic your test.


Device : Dell Optiplex
OS     : Microsoft Windows 10 Pro - Version 1809 (OS Build 17763.404)
Miners : NewPac x 4
Hub    : GekkoScience
Code   : cgminer-4.11.1-windows-gekko-363d879.7z

I was able to get hashing to get stuck at 60%, on one occasion while connected to eu1, but i didn't have enough logging to see anything useful.

Now I can't seem to trigger it again. Working as expected every run since then.
The pool handles AB correctly from all angles I can tell on Windows & Linux.

10 hour run:

Code:

cgminer version 4.11.1 - Started: [2019-04-09 20:06:36.028]
--------------------------------------------------------------------------------
(5s):400.3G (1m):392.0G (5m):385.0G (15m):383.1G (avg):385.5Gh/s
A:3019264 R:2560 HW:279 WU:5385.1/m
Connected to eu1.bitminter.com diff 512 with stratum as user gtaddict.5xNewpacs
Block: d2ac7486... Diff:6.39T Started: [05:14:49.130] Best share: 1.47M
--------------------------------------------------------------------------------
[U]SB management [P]ool management [S]ettings [D]isplay options [Q]uit
0: GSH 10030040: BM1387:02+ 500.00MHz | 99.6% WU:100% | 112.2G / 113.0Gh/s WU:1578.7/m
1: GSH 10030043: BM1387:02+ 500.00MHz | 100.0% WU:100% | 125.2G / 113.1Gh/s WU:1579.4/m
2: GSH 10031860: BM1387:02+ 337.50MHz | 98.0% WU:^95% | 70.66G / 72.91Gh/s WU:1018.4/m
3: GSH 10031863: BM1387:02+ 387.50MHz | 98.4% WU: 98% | 92.53G / 86.52Gh/s WU:1208.6/m
--------------------------------------------------------------------------------

I have a todo item to loosen up the no nonce check.
It was intended to be overly sensitive, but may be backfiring on windows.

edit:

@all
A running tally of the pools and its asic-boost condition is at the bottom of This initial post.
The top "okay" section are pools that I've confirmed to accept work and nonces from the extra ab midstate(s).

If you spot something different, using a different combination or set of settings + pools + local environment.
Post as much detail to replicate as you can and I'll continue to recheck/investigate.

cjsummers82

newbie

Activity: 22

Merit: 30

Quote from: JohnDon87 on April 06, 2019, 04:30:29 PM

Hey guys,

I am having some difficulty with the newpacs and hope someone can shed some light. Currently, I am running 6 newpacs off of a BrovSS USB hub with active cooling. All of the sticks are connected to this hub. The frequency they are running off of is 100 Mhz each. The version of cgminer is the gekkoscience cgminer 4.1 and using on Windows 10. I have been able to mine a couple blocks but eventually cgminer displays a “GSH 3 missing nonces” and “GSH 4 missing nonces” and then displays “Toggling reset on GSH 3” amd “Toggling reset on GSH” and my worker goes offline and have to reboot everything each time. Again, I appreciate any help someone can offer.

I am running 5 sticks at 500 mhz+ with sidehacks GekkoScience USB hub -- hardly get any nonce errors nor do devices frequently reset. With active cooling, this setup has been rock solid. I was able to keep it up and running without issues for 2+ weeks without a reboot when I was on vacation last month.

Link to GekkoScience USB hub: https://www.amazon.com/dp/B07PP9SJ1R/

gt_addict

hero member

Activity: 2534

Merit: 623

I have asked DrHaribo about it but havent seen a reply yet. But VH is going to look into it. Im happy they are working at the moment and it is in Beta stage so bound to be a few bugs, but its performing better than the last implementation for me and is hands down a better way to do it for us hobby miners.

NotFuzzyWarm

legendary

Activity: 3822

Merit: 2703

Evil beware: We have waffles!

Quote from: gt_addict on April 10, 2019, 05:34:27 AM

VH has added the no boost option as it is on by default:-

Quote from: vh on April 07, 2019, 08:35:05 PM

+ cleared issue sending AB stratum message on non AB pools.
+ if pool indicates AB is supported, AB will be enabled.
+ removed option "--gekko-newpac-boost", use "--gekko-noboost" to force non AB mode.
+ added [plus] flag after chip count to indicate asic-boost is functioning.

Thats the issue I was having. With it ON i was getting no nonce and restart of the sticks constantly. With it OFF it runs as expected. VH is going to look at bitminter to see whats going on as the pool is an asicbokst pool.

Then I stand corrected. I have not updated the software running mine in ages and its default is NO AB: AB has to be specifically turned on. In my case just a matter of "it works so leave it alone!" I take it sometime after the version I run VH changed that to be ON by default.

Dumb move (making AB active by default) in my book because the smaller pools that accept stick miners probably did not do their homework and just went with -ck's implementation of enabling the pool end of AB expecting it to work without any issues... Kano delayed activating AB support for 2 months because he not surprisingly found several bugs in the example code -ck provided. As I recall, those bugs were specifically related to the new Stratum calls. Only after Kano fixed those and PROPERLY TESTED THE CODE - unlike a certain other programmer/pool operator/mod who does not do that - did he enable AB on the pool. The large pools of course no doubt had the resources to figure it out fast as well.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: gt_addict on April 10, 2019, 05:34:27 AM

VH has added the no boost option as it is on by default:-

Quote from: vh on April 07, 2019, 08:35:05 PM

+ cleared issue sending AB stratum message on non AB pools.
+ if pool indicates AB is supported, AB will be enabled.
+ removed option "--gekko-newpac-boost", use "--gekko-noboost" to force non AB mode.
+ added [plus] flag after chip count to indicate asic-boost is functioning.

Thats the issue I was having. With it ON i was getting no nonce and restart of the sticks constantly. With it OFF it runs as expected. VH is going to look at bitminter to see whats going on as the pool is an asicbokst pool.

Hmmm, I couldn't find that in the code - where is it?

gt_addict

hero member

Activity: 2534

Merit: 623

Quote from: NotFuzzyWarm on April 10, 2019, 03:23:02 AM

Correct. The only AB switch is to turn it on. The command to add to turn AB on is --gekko-newpac-boost

If that switch is not added to the command line then the miner runs in normal mode without AsicBoost.

VH has added the no boost option as it is on by default:-

Quote from: vh on April 07, 2019, 08:35:05 PM

+ cleared issue sending AB stratum message on non AB pools.
+ if pool indicates AB is supported, AB will be enabled.
+ removed option "--gekko-newpac-boost", use "--gekko-noboost" to force non AB mode.
+ added [plus] flag after chip count to indicate asic-boost is functioning.

Thats the issue I was having. With it ON i was getting no nonce and restart of the sticks constantly. With it OFF it runs as expected. VH is going to look at bitminter to see whats going on as the pool is an asicbokst pool.

NotFuzzyWarm

legendary

Activity: 3822

Merit: 2703

Evil beware: We have waffles!

Quote from: kano on April 10, 2019, 03:19:20 AM

I don't see --gekko-noboost anywhere in vh's public git cgminer.c source ... so that will stop it from running.

Correct. The only AB switch is to turn it on. The command to add to turn AB on is --gekko-newpac-boost

If that switch is not added to the command line then the miner runs in normal mode without AsicBoost.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

I don't see --gekko-noboost anywhere in vh's public git cgminer.c source ... so that will stop it from running.

gt_addict

hero member

Activity: 2534

Merit: 623

Quote from: vh on April 09, 2019, 01:59:36 PM

That's some mighty fine sleuthing there...

Thanks VH, I did scan the readme but must have missed it.

My details are as follows:-

Code:

cgminer -o stratum+tcp://eu1.bitminter.com:3333 -u gtaddict.5xNewpacs -p x --widescreen --gekko-tune-down 90 --gekko-newpac-freq 500 --gekko-noboost

It could well be because I'm using the EU server and not the US one. But I dont see how that would make a difference if it all ends up at the same place Huh

sidehack

legendary

Activity: 3374

Merit: 1859

Curmudgeonly hardware guy

Yeah, you're missing about 15GH. 350MHz should see 79.8GH which means your 350MHz is running at 81% and you'd get the same hashrate with probably lower power use out of 285MHz.

goldsounds

newbie

Activity: 14

Merit: 1

Tried out 363d879

used code:

Code:

-o stratum+tcp://us-east.stratum.slushpool.com:3333 -u xxxx.xxxx -p x --gekko-newpac-freq 500 --suggest-diff 512 --balance --gekko-noboost

no nonce on 500 ---> 200mhz
no nonce on 200 ---> 187.5mhz
never could get it stable after running for a while, always under 50% WU^

however, if i use the Dec 18 build, i can get 350mhz going, with an average of 64gh/s

Code:

cgminer version 4.11.1 - Started: [2019-04-09 15:15:20.876]
--------------------------------------------------------------------------------
(5s):53.37G (1m):66.19G (5m):51.19G (15m):25.96G (avg):64.62Gh/s
A:5120 R:8192 HW:0 WU:903.5/m
Connected to us-east.stratum.slushpool.com diff 512 with stratum as user xxxxx.xxxxx
Block: cb080a68... Diff:6.39T Started: [15:15:20.877] Best share: 2.29K
--------------------------------------------------------------------------------
[U]SB management [P]ool management [S]ettings [D]isplay options [Q]uit
0: GSH 10031593: BM1387:2 350.00MHz | 61.66G / 64.61Gh/s WU:903.5/m
--------------------------------------------------------------------------------
[2019-04-09 15:15:20.877] Network diff set to 6.39T
[2019-04-09 15:15:21.584] GSH 0: found 2 chip(s)
[2019-04-09 15:15:21.644] GSH 0: setting frequency to 350.00MHz
[2019-04-09 15:15:21.661] GSH 0: open cores @ 350.00MHz
[2019-04-09 15:15:25.846] GSH 0: start work @ 350.00MHz
[2019-04-09 15:16:06.409] Pool 0 difficulty changed to 1638
[2019-04-09 15:16:06.409] Stratum from pool 0 requested work restart
[2019-04-09 15:16:32.965] Pool 0 difficulty changed to 512
[2019-04-09 15:16:36.777] Stratum from pool 0 requested work restart
[2019-04-09 15:16:44.673] Accepted 54156740 Diff 779/512 GSH 0
[2019-04-09 15:17:08.360] Accepted 723a5cb3 Diff 574/512 GSH 0
[2019-04-09 15:17:08.360] Accepted 60dea91d Diff 677/512 GSH 0
[2019-04-09 15:20:14.278] Accepted 47636631 Diff 918/512 GSH 0
[2019-04-09 15:21:22.356] Accepted 723effc2 Diff 574/512 GSH 0
[2019-04-09 15:21:26.244] Accepted 796f4e0d Diff 540/512 GSH 0
[2019-04-09 15:21:41.957] Accepted 6b2712dc Diff 612/512 GSH 0
[2019-04-09 15:22:00.313] Accepted 62fa6309 Diff 662/512 GSH 0
[2019-04-09 15:22:17.297] Accepted 4b9ca8e7 Diff 867/512 GSH 0
[2019-04-09 15:22:58.517] Accepted 1ca92894 Diff 2.29K/512 GSH 0

just seeing if i was missing something? I like the peak hash to get a little more oomph out of my stick.

thanks!!

vh

hero member

Activity: 699

Merit: 666

That's some mighty fine sleuthing there.

If you have a user.worker you can PM me, I'll take a closer look at bitminter.

Most pool post a hashrate and luck value. WU would be the luck component.
Hash% WU:Luck%

the "o" and "-" was just a quick way to tag which chip the no nonce is associated with.
it's there temporarily to help with the R606.

edit:
My explanation on WU was a little too generically put.
From the README:

Code:

Q: What should my Work Utility (WU) be?
A: Work utility is the product of hashrate * luck and only stabilises over a
very long period of time. Assuming all your work is valid work, bitcoin mining
should produce a work utility of approximately 1 per 71.6MH. This means at
5GH you should have a WU of 5000 / 71.6 or ~ 69. You cannot make your machine
do "better WU" than this - it is luck related. However you can make it much
worse if your machine produces a lot of hardware errors producing invalid work.

gt_addict

hero member

Activity: 2534

Merit: 623

@VH

So finally managed to play around with swapping the USB port and using just one stick with my "no nonce" "0 chips" quandary. After adding the "--gekko-noboost" option when running a single stick it now climbs and after 5mins is currently upto 356.25m (and still climbing) when set to 450m.

So it looks like the AB function and Bitminter pool isn't compatible.

I will try the original port and multiple sticks and report back with further findings

EDIT:
So hooked up All 5 sticks and plugged back into the original usb port. Set the target freq to 500m and within 15mins all the sticks are running at 500m (set tune down to 90).

I can't remember what the top freq is for the TWO sticks per regulator. Anyone have that info to hand? As I dont want to push the freq any higher and burn anything out. The whole hub is using 83w (at the wall).

EDIT 2:
Whats the "WU" with the percentage after it? Average performance? Maybe a dumb question but I gotta ask Cheesy

Also what do the "- or o" in the brackets represent?

rockmoney

sr. member

Activity: 439

Merit: 297

www.amazon.com/shops/MinersSupply

First, I must say this is a GREAT combined effort between sidehack and vh. I don't believe most people truly realize the complexities behind a project like this. My hat off to BOTH of you for your hard work!

[if there is such a thing] I am fairly well versed with Windows. I also use the latest Windows 10 Home version to run all of my GekkoScience gear, which at times, is 3 GekkoScience hubs each with 6 NewPac sticks, all running together simultaneously from the same PC. And, they do it with ease and perfection, using the latest stable December 13 vh build of cgminer!

Please let me know if there are any NewPac stick tests that need run on this latest build, and I can get the proper logging info over to vh. If that would help, I'm here and more than willing.

NOTHING about this is easy, so again, great work @sidehack and @vh for pulling this miracle off!

Thanks to your work, it's looking like these pod miners would be super powerful miners for such little power usage. Not to mention, it's almost silent operation, making them perfect in every way for home use! It would be a definite game changer to say the least! Wink

As a side note, I have recently brought my cousin on board to help with GekkoScience product related sales and support, as needed.

If for any reason you are unable to get in contact with me, my cousin CJ will be able to assist you with any sales inquiries and/or support related matters. For security, any direct forum sales payments MUST go DIRECTLY through ME (solely to protect customers from potential imposters). Please NEVER send payments to ANYONE else, or anyone claiming to be us.

CJ's official bitcointalk account is: cjsummers82

Feel free to contact me or CJ if needed. Thanks!

Topic: GekkoScience NewPac / Terminus R606 (BM1387) Official Support Thread - page 94. (Read 61906 times)