Patrike, Good day -- Antminer Z11 with the latest firmware from Bitmain and AwesomeMiner 8.1.6 seem to be very creative in the performance calculations. The output below is from a Z11e with 1 hashboard running for testing purposes: https://imgur.com/JEQoFZgIt is actually running around 30KSol but Awesome Miner insists it is running at 30MSol. Even better than that, it will report TH instead of M making that single board a significantly profitable according to the calculation. https://imgur.com/mtVeEzQAs best I can tell the API responses from this latest cgminer and from a Z15 are identical; the exception being that Z11 sends an extra \000 at the end of the "summary" and at the end of the "devs" API calls. We have a mutual customer who is using my firmware and your (licensed) AM running into this issue. I'm unsure how to fix this as it is happening with the stock firmware and I cannot find anything wrong with the API output save for the extra newline+\000. What do you need to help troubleshoot the issue? *edit* added in direct links as apparently I can't post image links embedded here at the moment. Thank you, Jason
|
|
|
I expect the entry points to be other than SSH also. cgminer's API is an entry point. The upload/configuration restore mechanism is an entry point (bitmain has tried to patch this as of late with varying levels of success). bitmain's latest additions to cgminer adds new functionality that isn't on the standard API port and probably needs some work....I'm still reverse engineering it but so far haven't seen any authentication/authorization.
Another entry point is... buying used miners. The last variant of this I found came through a reseller in China towards the end of a product cycle.
|
|
|
I found and reverse engineered a variant of this on Z series miners last year. The use of Tor was unexpected... those who ran the virus had SSH servers running on tor with the authentication keys in the malware. .... I was able to fix that for them. :-)
|
|
|
Please be advised of false advertising - Mining on luxor will have DEV fees. Obviously eFu**k is a small time crook but Luxor is quite sizeble so will be interewsting to see if false advertising charges can be filed against them.
Hi. If you are having an issue, try sending me a PM. I haven’t paid any attention to the forums in a while, but I respond almost immediately to PMs here, on Luxor discord and on the Equihash discord. -j
|
|
|
Sell them as home heaters with a built in lottery: If it mines a block, win! Otherwise, heat!
-j
|
|
|
I just started using the Z11 2.3 Fuddware on 4 x Z11. This is the first time I have used any Fuddware. The firmware works great for around 10 hours. Then, the machines stay working (as shown in control panel), but LSTime (last share time) keeps increasing as no new shares are submitted to the pool. This never ends; hours and hours will pass. Poolside hash rate drops to zero. A reboot fixes everything. But very annoying to reboot every 10 hours. Any idea what the problem is? Happened on Luxor and Suprnova, the only 2 pools I tried. Same deal. Obviously dev fee kicks in once in a while but not for hours and hours.
Luxor had issues this past weekend that may be what you saw. I can't speak to Suprnova. This is not my thread, so send me a PM with your supportID. -j
|
|
|
FYI, I'll be releasing a dev-fee free, $0 cost, no callbacks, no licenses, blahblah, completely unencumbered overclock for all of these SHA-256 "soon".
So.. be patient and don't pay $100/miner or $150/miner, etc. when I'll charge $0/miner.
-j
|
|
|
Changing SPI does impact hashing path and is just a waste. You do know what SPI is, right? And it, like I2C and RS232 both have different frequencies (“baud” in RS232). Those are just for communications between components and not hashing.... might as well overclock the LEDs to blink faster while you are in there.
You change the weirdest crap.
Also, “disparate”, not “desperate”... unless your voltage package options really are hard up, but I would assume you meant “unique”.
-j
|
|
|
Apparently you reported me correcting your grammar. Lol.
A reply of yours, quoted below, was deleted by a Bitcoin Forum moderator. Posts are most frequently deleted because they are off-topic, though they can also be deleted for other reasons. In the future, please avoid posting things that need to be deleted.
Quote .... higher THAN this. Not “then”.
Locking thread.
-j
|
|
|
Depending on board quality, should provide a total solution rate between ~60 KSol/s and ~70 KSol/s. ... I'll quote this from the post above. This has been long proven since March.
Have a nice day.
-j
I am not going to go back and forth with you get over your cry baby crap. You didnt want anything to do with the A9 now you want to be all up in it. So if the board quality is good then you should get 65+ k no problem with my mod. There was a lot more done with it then you think. Digging a hole, filling it back up, digging it again, filling it back up is also more work than you think... but completely unnecessary. -j
|
|
|
Depending on board quality, should provide a total solution rate between ~60 KSol/s and ~70 KSol/s. ... I'll quote this from the post. This has been long proven since March.
Have a nice day.
-j
|
|
|
Depending on board quality, should provide a total solution rate between ~60 KSol/s and ~70 KSol/s. ... I'll quote this from the post above. This has been long proven since March.
Have a nice day.
-j
|
|
|
A9 Overclocking guide, free, no need to go to pay someone for finally finding this themselves.This information has been available for since March of 2019 and results in 60-70kSol/second on an A9. Credit to Sto1cNate @ Discord for putting this together. Bellow you will find detailed information about overclocking the A9 using the October 2018 firmware. [2:19 AM] Sto1cNate: SSH Login details (terminal, putty, mobaXterm) ssh root@ipaddress password: blacksheepwall October 2018 Firmware Download Link http://www.innosilicon.com.cn/download/b29+_20181023_130635.swu[2:19 AM] Sto1cNate: So, this is how this works for the October firmware and why it is superior.... [2:19 AM] Sto1cNate: You do not set the VID at all. Voltages are set through a fixed VID table in relation to the frequency. [2:20 AM] Sto1cNate: Clock frequency of the cores are 100% set by the PLL setting in a 1:1 ratio. Ex: A PLL setting of 1500 = 1500MHz. Maximum PLL setting is 1596.[2:22 AM] Sto1cNate: Here is a canned copy paste you can modify (ONLY WITH THE OCTOBER 2018 FIRMWARE).DO NOT PASTE THE FOLLOWING INTO YOUR CONFIG FILE! Paste it directly into your SSH terminal after logging into the A9 and have the # prompt. #Edit the pools before pasting #Copy and paste this whole block of text after making desired edits echo -n \ '{ "pools": [ { "url": "stratum+tcp://na.luckpool.net:3060", "user": "znhGeka9zXmixvw6ufzGpcaSXcSACrjx5WZ.A9-XXX", "pass": "x" }, { "url": "stratum+tcp://eu.luckpool.net:3060", "user": "znhGeka9zXmixvw6ufzGpcaSXcSACrjx5WZ.A9-XXX", "pass": "x" }, { "url": "stratum+tcp://ap.luckpool.net:3060", "user": "znhGeka9zXmixvw6ufzGpcaSXcSACrjx5WZ.A9-XXX", "pass": "x" } ], "api-listen": true, "api-port": "4028", "api-allow": "R:0/0,W:127.0.0.1", "noauto": true, "T1Pll1": "1500", "T1Pll2": "1500", "T1Pll3": "1500", "A1Fanspd": "50", "A1Fanmode": "0" }' > /config/cgminer.conf echo -n '{"mode":"performance","level":"4"}' > /config/profile.conf #The miner must be rebooted for the new configuration to load
[2:27 AM] Sto1cNate: The above configuration on my test A9 results with using ~18.45 amps at 12.0 volts per card.WARNING!!! At that current level, you need to be using a power supply with 16 gauge cable. https://www.powerstream.com/Wire_Size.htmAmerican Wire Gauge Chart and AWG Electrical Current Load Limits ... AWG Wire size chart and ampacity table for design engineers including skin depth frequencies and tensile strength data [2:28 AM] Sto1cNate: That is a total system power of roughly 664.2 watts without adding in the additional power from efficiency losses.704 watts from the wall if running at peak efficiency on a 94% rated PSU. Solution rate per hash board was between ~21 KSol/s and ~23 KSol/s. Depending on board quality, should provide a total solution rate between ~60 KSol/s and ~70 KSol/s.Sol per watt efficiency would fall between 90.3 and 105.4. For reference, the A9++ has a Sol/w rating of 90.3 and the Z11 has a rating of 90.5 [2:33 AM] Sto1cNate: Here is another handy copy paste to enable full privileged access to the A9's API.#Copy and paste this whole block of text #Make the filesystem writable mount / -o rw,remount #Modify cgminer service unit file to enable API access echo \ '[Unit] Description=sgminer4a9 After=network.target config.mount Requires=config.mount StartLimitIntervalSec=0
[Service] Type=simple PrivateTmp=no ExecStartPre=/bin/sh -c "mkdir -p /tmp/log" ExecStartPre=/bin/sh -c "echo "B29+" > /tmp/type" ExecStartPre=/bin/sh -c "echo "G19" > /tmp/hwver" ExecStartPre=/bin/sh -c "if [ ! -s /config/cgminer.conf ]; then cp -f /etc/cgminer.conf.default /config/cgminer.conf; fi" ExecStartPre=/bin/sh -c "if [ ! -s /config/cg.conf ]; then cp -f /etc/cg.conf.default /config/cg.conf; fi" ExecStart=/bin/cgminer -c /etc/cgminer.conf --api-listen --api-network --api-allow W:0/0 --syslog --text-only
Restart=always
[Install] WantedBy=multi-user.target' > /etc/systemd/system/multi-user.target.wants/cgminer.service #Apply changes made to unit file by reloading systemd systemctl daemon-reload #The miner must be rebooted for the new configuration to load [2:33 AM] Sto1cNate: NOTE: I have deleted a few of my earlier posts as I was experimenting and don't want to add to any more confusion that is scattered all over this channel.
If you do not have the A9 October firmware, you can download and install it by just changing the clock on your computer back a couple of months which will bypass the limitations and upgrade checks that innosilicon put in place. I can make the A9/October firmware available if needed. PM me for questions. Thanks, Jason
|
|
|
pro-tip: some of the wiring on the A9 cannot handle the current when overclocked in a safe manner.
-j
Still within all factory specs so hopefully it can handle it. There are also 2 efuses that need to be careful not to trigger. But thanks edit.... The voltages are actually in the same range as the miner would set in one of its modes. It also appears that the miner is ignoring the 2nd and 3rd T1pllx from the config file. Yup - the voltage stuff has been documented for many months now.. and one of the places where additional code needs to be added to support it. My comment on current was not made out of the blue -- burnt wiring will happen if you push these. core voltage is expected to be no more than 900mV "in range as the miner would set". re: efuse. Don't mistake the existence of these on the die for "what is in any of the code paths". Check and see if pin16 is high or low -- probably low. And if so, that means you can't trigger them via SPI, which means you can't trigger them via "the firmware".... plus it just isn't clear that they do anything, whether they are single fire, or resettable. in other words, nothing to pay attention to nor be "careful not to trigger". ... and if it is high, then simply don't write to d08000000 via SPI. -j
|
|
|
pro-tip: some of the wiring on the A9 cannot handle the current when overclocked in a safe manner.
-j
|
|
|
Locking thread. Please feel free to contact me via PM if you have any questions.
Services will continue to run for users, as promised, until the last miner is turned off.
I will unlock the thread if there are any major updates.
Thank you.
-j
|
|
|
No not at nothing just correcting the wrong information you posted. It is dishonesty when you say there has not been no issues with paid version when in fact you know there have been. I thought you were Mr Honest when it came to users.
You have no problem hiding behind your discord making comments about me and since you block me there I chose to comment here in the open not behind your back. Just like one of your users asked the difference between your and my versions and all you tell them is mine is bullshit. You wanted a truce but then hide and do the same crap, not much of a truce. Sad thing is I have recommended to some users over mine.
*sigh* This was received yesterday from a farm... [2:44 PM] REDACTED: just a follow-up after a few days. your firmware is rock steady without a blip using a license.... But hey.. you are right... one time in the last year, I fucked up for about 4 hours and impacted licensed users. Otherwise, *EVERY SINGLE ISSUE* has been a *USER* or *ENVIRONMENT* issue, not a firmware issue. I have 0 interest in having any discussion about you to anyone, so 'bullshit' is a dismissive statement of non-interest, not an accusation. I'm going to ask you again, to kindly go to your thread and mind your own business. -j I have stayed away except to call you out on the bs you try to claim. And again your full of it because since at least 7/28 you have had a few comments or conversations about me or my work. Anyone who follows this thread can read there have been more problems then you claim. Most of it is dev mode stuff and a lot less then a yr ago. Which randon switching into dev mode is not a user or environmental problem. Sigh. You are making it very hard to be nice to. -j
|
|
|
|