-click to enlarge -
Topic: GekkoScience NewPac / Terminus R606 (BM1387) Official Support Thread - page 107. (Read 61906 times)
December 11, 2018, 07:54:22 PM
I just found this screenshot of cgminer v4.10, running x2 compac sticks on 2017 JUNE 01, and don't see a hardware error percentage.
-click to enlarge -
-click to enlarge -
December 11, 2018, 06:55:09 PM
I just got my Newpac today. Seeing a lot of HW errors even at 100MHz. Could this be due to a dodgy hub?
Like sidehack said, if your getting a good hashrate, the HW errors are unimportant. If your running at 100MHz, you should be seeing around 22.8 GH/s.
If thats not the case, it could be an issue with power from the hub, try plugging directly into the USB port on the back of the PC for direct power from the motherboard. The other thing that tends to cause excessive HW errors is heat, you can try pointing a simple fan at it to help dissipate the heat a little. If neither one of those two things fixes your issues, let me know, there are a few other things but 99% of the time its those first two.
December 11, 2018, 03:44:35 PM
I don't recall explicitly adjusting anything related to hardware error percentages.
Try attaching a screen shot of what you notice was omitted in this build.
Not much has changed recently on the display side.
Jan 2017 (about two years ago)
http://23.108.83.14/images/2Pac-0004.png
Dec 2018 (this build)
http://23.108.83.14/images/2Pac-0012.png
If you are comparing pre 4.10, I wouldn't be able to answer.
There would probably be some notes in GitHub.
Try attaching a screen shot of what you notice was omitted in this build.
Not much has changed recently on the display side.
Jan 2017 (about two years ago)
http://23.108.83.14/images/2Pac-0004.png
Dec 2018 (this build)
http://23.108.83.14/images/2Pac-0012.png
If you are comparing pre 4.10, I wouldn't be able to answer.
There would probably be some notes in GitHub.
Now that I've looked at further, I see that cgminer hasn't included HW % rate; I was confusing it with bfgminer I suppose.
My apologies.
December 11, 2018, 03:43:40 PM
Return the stick to it's factory vcore to ease power draw at high frequency.
http://23.108.83.14/images/NewPac-0006.png
Test them one at a time to see if you can push past 55GHs.
Start at 300MHz, unless you get 68GHs (0.228 x 300) don't go higher .
Try with and without the hub - single stick.
I find my dell's usb port does pretty well with a single stick, fan, no powered hub.
Once you get 68GHs, bump the freq up by no more than 25-50MHz at a time, test for 24hrs...
One at a time first and you should be able to find where the bottleneck is.
http://23.108.83.14/images/NewPac-0006.png
Test them one at a time to see if you can push past 55GHs.
Start at 300MHz, unless you get 68GHs (0.228 x 300) don't go higher .
Try with and without the hub - single stick.
I find my dell's usb port does pretty well with a single stick, fan, no powered hub.
Once you get 68GHs, bump the freq up by no more than 25-50MHz at a time, test for 24hrs...
One at a time first and you should be able to find where the bottleneck is.
Thanks vh, i will give that a go this evening and report back tomorrow.
December 11, 2018, 10:16:59 AM
The NewPac doesn't implement a ticket mask so there's no minimum threshold below which shares are discarded at the chip. All shares, even diff-1 shares, are reported. This means the total volume of data coming back is significantly higher than previous sticks even for the same hashrate. In order to help with the volume, we're also operating the serial bus at about 13 times the bitrate of previous sticks.
All this makes the data bus more susceptible to noise and tight timing and signal-quality constraints, as well as leaving no filter for potentially erroneous shares from the chips.
The HW error rate could be lousy for a near-threshold-voltage stick but as long as the overall hashrate is preserved (within 98% or so of theoretical) I wouldn't worry about it. If your hashrate is reading low, then there's a problem.
All this makes the data bus more susceptible to noise and tight timing and signal-quality constraints, as well as leaving no filter for potentially erroneous shares from the chips.
The HW error rate could be lousy for a near-threshold-voltage stick but as long as the overall hashrate is preserved (within 98% or so of theoretical) I wouldn't worry about it. If your hashrate is reading low, then there's a problem.
December 11, 2018, 10:03:49 AM
I just got my Newpac today. Seeing a lot of HW errors even at 100MHz. Could this be due to a dodgy hub?
December 11, 2018, 09:24:51 AM
Return the stick to it's factory vcore to ease power draw at high frequency.
Test them one at a time to see if you can push past 55GHs.
Start at 300MHz, unless you get 68GHs (0.228 x 300) don't go higher .
Try with and without the hub - single stick.
I find my dell's usb port does pretty well with a single stick, fan, no powered hub.
Once you get 68GHs, bump the freq up by no more than 25-50MHz at a time, test for 24hrs...
One at a time first and you should be able to find where the bottleneck is.
Test them one at a time to see if you can push past 55GHs.
Start at 300MHz, unless you get 68GHs (0.228 x 300) don't go higher .
Try with and without the hub - single stick.
I find my dell's usb port does pretty well with a single stick, fan, no powered hub.
Once you get 68GHs, bump the freq up by no more than 25-50MHz at a time, test for 24hrs...
One at a time first and you should be able to find where the bottleneck is.
December 11, 2018, 05:26:09 AM
Well after turning down the pot to the 6oclock position and setting it to 400freq im still only getting 55ghs each. Can anyone help out? I cant figure ot why they wont break the 55ghs mark.
December 11, 2018, 12:12:37 AM
vh:
Thanks for the update! I noticed that the hardware error percentage rate has been removed/omitted from this version...was that intentional?
Thanks for the update! I noticed that the hardware error percentage rate has been removed/omitted from this version...was that intentional?
I don't recall explicitly adjusting anything related to hardware error percentages.
Try attaching a screen shot of what you notice was omitted in this build.
Not much has changed recently on the display side.
Jan 2017 (about two years ago)
Dec 2018 (this build)
If you are comparing pre 4.10, I wouldn't be able to answer.
There would probably be some notes in GitHub.
December 10, 2018, 09:12:01 PM
A new build is available (9d66fd3):
Windows binary: cgminer-4.11.1-windows-gekko-9d66fd3.7z
- Fixed datetime display.
- Closed gaps in data transmission to reduce line power fluxations.
- Spin off dedicated work tx thread.
- Switch from blocking get_work to get_queued.
- Send null work to keep asic busy during network glitches and when pool is reacting to block events.
Reminder best practice update process: keep old working (or favorite) build readily available.
Windows binary: cgminer-4.11.1-windows-gekko-9d66fd3.7z
- Fixed datetime display.
- Closed gaps in data transmission to reduce line power fluxations.
- Spin off dedicated work tx thread.
- Switch from blocking get_work to get_queued.
- Send null work to keep asic busy during network glitches and when pool is reacting to block events.
Reminder best practice update process: keep old working (or favorite) build readily available.
- move the existing "cgminer" folder to "cgminer.yyyymmdd"
- clone repository to create a new cgminer folder as noted from first post instructions.
vh:
Thanks for the update! I noticed that the hardware error percentage rate has been removed/omitted from this version...was that intentional?
December 10, 2018, 10:43:34 AM
Makerbots are good, just not cheap. I wouldnt buy one, but if you have access to one through a school, take advantage of that.
If enough people need it, I can print them (for a small fee, maybe $3 + shipping?) and mail them out.
If enough people need it, I can print them (for a small fee, maybe $3 + shipping?) and mail them out.
Definitely another venture for the site. If you do 3D printing. I was on talks with a guy a few months ago to design a duct for the HP server psus so you could have a stronger fan blowing through from the back but he never got back to me. A fan mount for usb sticks on the ther hand is always a winner.
December 10, 2018, 10:20:20 AM
3D printing I wonder what we have access to? I think makerbot? They any good 419?
Makerbots are good, just not cheap. I wouldnt buy one, but if you have access to one through a school, take advantage of that.
If enough people need it, I can print them (for a small fee, maybe $3 + shipping?) and mail them out.
December 10, 2018, 09:55:48 AM
And theres me thinking it wasnt high enough. Ill be turning them down when I get home then lol. Where abouts should it be sidehack? Its been a while since I dabbled with the compacs.
December 10, 2018, 09:33:04 AM
That's pretty bad. 350MHz should get you 80GH each. If it takes that much fan to cool them and the hashrate is still that bad at speed, your voltage is probably way the heck too high. If the flat of the pot is anywhere on the left side it's too high.
December 10, 2018, 08:38:44 AM
Nice, I'm sitting at 350 myself. Waiting to figure out better cooling. Think that's your issue? they get to hot and throw nonce? I'll might try to hit 375 an 400 this week if I get the chance to sit around tinker.
3D printing I wonder what we have access to? I think makerbot? They any good 419?
3D printing I wonder what we have access to? I think makerbot? They any good 419?
I think that may well have been my issue. Put a 120x38 delta fan on them through a duct and they seem to be homding at 55ghs each at 350m now.
December 10, 2018, 12:47:31 AM
Thanks for the help. Got two instances running one stick will do no more than 300mhz and the other no more than 350mhz. Any high and they drop out. Can’t figure out how you guys are running them so high.
Nice, I'm sitting at 350 myself. Waiting to figure out better cooling. Think that's your issue? they get to hot and throw nonce? I'll might try to hit 375 an 400 this week if I get the chance to sit around tinker.
3D printing I wonder what we have access to? I think makerbot? They any good 419?
December 09, 2018, 08:06:06 PM
check out the Creality3D Ender 3, its what I run, its got a good size build platform and the price was very reasonable.
will do, thank you
December 09, 2018, 05:55:37 PM
looking good, now i just need to get a 3d printer.
check out the Creality3D Ender 3, its what I run, its got a good size build platform and the price was very reasonable.
December 09, 2018, 04:10:19 PM
Have to run multiple batch files with "--usb :1" in the line along with the change in freq. I think there's a way to target specific ports but I don't recall how to do that.
Thanks for the help. Got two instances running one stick will do no more than 300mhz and the other no more than 350mhz. Any high and they drop out. Can’t figure out how you guys are running them so high.
December 09, 2018, 03:54:16 PM
Summmary from a Saturday of rearranging the system...
After much mucking about, the end result is pretty good:
The 9 NewPacs are rock solid - no `missing nonces' and such. 6 @ 550 MHz and 3 @ 475 MHz in a `3 by 3' set of cgminer instances... I pulled the 2Pacs out of the Eyeboot USB3.0 Hub... they can have a rest until the new 2.0 hubs arrive... well, all but one 2Pac which is quietly running at 200 MHz in its own cgminer instance... the Terminus is also running happy at 250 MHz, also in its own cgminer instance...each type of GS `flavor' is on its own controller and hub... Lots of fans everywhere...All miners, including the one suspect of the six new arrivals from yesterday are running at or within 1/2 percent of expected hashrate...solid, solid and solid... I'm not the least bit panicked about heat and the 9 NewPacs have been running for 14 hours in this configuration... still show as GSH 0, 1 and 2 on each cgminer - with no restarts. If I haven't said so before, these NewPacs are fantastic...
There's no point in calling them impossible or unbelievable. They clearly have happened!
After much mucking about, the end result is pretty good:
The 9 NewPacs are rock solid - no `missing nonces' and such. 6 @ 550 MHz and 3 @ 475 MHz in a `3 by 3' set of cgminer instances... I pulled the 2Pacs out of the Eyeboot USB3.0 Hub... they can have a rest until the new 2.0 hubs arrive... well, all but one 2Pac which is quietly running at 200 MHz in its own cgminer instance... the Terminus is also running happy at 250 MHz, also in its own cgminer instance...each type of GS `flavor' is on its own controller and hub... Lots of fans everywhere...All miners, including the one suspect of the six new arrivals from yesterday are running at or within 1/2 percent of expected hashrate...solid, solid and solid... I'm not the least bit panicked about heat and the 9 NewPacs have been running for 14 hours in this configuration... still show as GSH 0, 1 and 2 on each cgminer - with no restarts. If I haven't said so before, these NewPacs are fantastic...
There's no point in calling them impossible or unbelievable. They clearly have happened!
Nice, glad everything is humming a long. Any shots of the set up?
Jump to: