Luxor provides pre-configured profiles with board voltage and chips frequencies set for nominal miner models. These profiles incorporate board voltage and chip frequencies settings and include options for standard, under-clocking, and over-clocking configurations.
As a rule of thumb, higher voltage and higher frequency translate into higher hashrate and power consumption. Similarly, lower voltage and lower frequency translate into lower hashrate and lower power consumption.
The Default profile is generated by reading the EEPROM data from the machine, the Default profile represents the factory hashrate output and power consumption for that miner.
To update the profile:
- Go to the 'Configure' section and click on 'Pre-set Profiles'.
- From the dropdown menu, select your desired profile.
- Verify that the chip frequencies have changed to the selected setting.
Note: Users can create their own custom profiles through the API. Ability to create your own profiles from the user-interface is under development.
LuxOS provides three user-customizable temperatures that drive different actions in the miner.
Temperatures reported by LuxOS are hashboard temperatures and shall not be confused with chip temperatures. For most S19 series miners, stock and other alternative aftermarket firmware's estimate chip temperatures by adding ~15C to the hashboard temperature sensors readings. Since chip temperatures are estimations, LuxOS shows hashboard temperatures only.
Most miners will have 4 temperature sensors per hashboard, although there are a few exceptions of miners coming with 2 temperature sensors per hashboard. LuxOS makes decisions based on the highest reading across all hashboard temperature sensors.
Target Temperature: Used for the fan algorithm. Once the maximum temperature reading reaches the Target Temperature threshold fan speeds will be set to 100% unless the LuxOS fan algorithm is overridden with Manual fans settings.
Hot Temperature: Used by ATM (Advanced Thermal Management) to drive steps up and steps down in profiles.
Dangerous Temperature: When any of the temperatures in a hashboards reaches this temperature the hashboard will be shutdown to prevent damage.
|Temperature Setting||Default||Fiber Glass Hashboards||Aluminium Hashboards|
|Hot Temperature||65C||For units with ATM enabled and if the user wants to experience less underclocking then increase this setting.||To preserve the lifespan of the miner we don't increasing the hot temperature for sustained periods of time.|
|Dangerous Temperature||70C||Maximum of 75C, ideally a 5C buffer from your Hot Temperature for ATM to have enough room to cool-down a miner before it shuts down due to high temperatures.||Maximum of 72C, ideally a 5C buffer from your Hot Temperature for ATM to have enough room to cool-down a miner before it shuts down due to high temperatures.|
Advanced Thermal Management, is designed to adjust a miner's operating parameters (frequency & voltage) based on environmental conditions and user-defined settings and boundaries.
In high-heat environments, ATM will provide a smooth underclock as temperatures rise to prevent miners from shutting down due to high temperatures. Entering into the evening as heat dissipates ATM will drive the miners back to stock settings (or into overclock if user-configured). ATM does not require integrations, triggers, or human intervention to operate.
Additionally, ATM can be used as a lifespan-extending tool as with the right setting miners can decide to operate their equipment at a lower temperature which translates to lower failures and increased lifetime of the mining equipment.
ATM differentiates from Bitmain's Lower Power Mode as it provides a progressive underclock/overclock of the machine as needed throughout the day. For example, ATM might underclock a miner 3% between 10 AM and 11 AM, then 5% until 12 PM, then 10% until 4 PM then 5% until 6 PM and back to stock settings afterwards. While Bitmain's LPM will result in a 25% underclock from start to end. Additionally, Bitmain's LPM might not be enough to prevent miners from overheating in certain scenarios and LuxOS ATM can underclock as much as needed to keep a miner hashing.
Minimum Profile: The minimum profile sets the minimum boundary profile ATM can decrease to. By default the minimum profile is unbounded, meaning ATM will drop to the lowest profile available if required.
If a relatively high power consumption (voltage & frequency) profile is configured as the minimum boundary ATM might not be capable of preventing the machine from overheating in high heat scenarios.
Maximum Profile: The maximum profile sets the top-end boundary profile ATM can increase to. By default, the EEPROM profile (stock settings) is used. Users can increase this setting but acknowledge that miners will draw more power.
The maximum profile can not be unbounded. A maximum profile needs to be selected for ATM to operate. ATM will increase up to the maximum boundary if the temperature allows it to.
Temperature Buffer: This buffer refers to how many degrees C from Hot Temperature are required before increasing a profile.
For example, if the observed hashboard maximum temperature is 59C, the temperature buffer is 5C and the Hot Temperature is 65C then ATM will increase a profile within its boundaries. However, if the observed maximum temperature is 63C then ATM will not increase a profile as it risks overshooting temperature by reaching the Hot Temperature and having to come back down again.
Default value is 5C. If you want to recover nameplate hashrate quicker during hot days recommend setting this value closer to 2 or 3C.
Evaluate Up Minutes: This parameter determines the frequency in which ATM evaluates to increase a profile. Because heat can take some time to build up we recommend a 5-minute evaluation period. Although more conservative also prevents constant profile changes due to overshooting.
Evaluate Down Minutes: This parameter determines the frequency in which ATM evaluates to decrease a profile. We recommend setting it to 1-minute as when the temperature raises fast we want ATM to be able to react quickly to prevent a shutdown of a machine.
During the initial ramp, the slow process of reaching the default profile might cause the miner to overheat before ATM is able to cool it down, since ATM evaluates after the miner ramps to a profile.
In that case, you might need to disable ATM, set a lower profile, and then re-enable ATM. The miner will ramp to the selected lower profile and then ATM will ramp back up when the temperature allows.
Additionally, might be prudent to put the miner on manual fans at 100% during the initial boot sequence. A miner reboot causes fans to stop working which can cause to loss of air pressure between cold and hot aisles which causes a lot of hot air to come back through the overheating miner. Settings fans to 100% will help the miner ramp without overheating.
Once the miner reaches the target fan temperature (60C by default) and its target profile settings fans can be set back to Automatic mode.
You can view your ASIC miner's logs from the dashboard. Click on 'Logs' to see the current log, which is also available for export. Past logs can be accessed in the history section.
Note: For better support. we do recommend NOT REBOOT and send the "Current logs export" if the issue is recent, if the issue pressist or is not often, on the "History" tab you will fine the logs by date, download those and send it on a zip file over our support channel in telegram (opens in a new tab).
A feature provided by LuxOS is hash rate splitting. It provides users the ability to allocate fractions of total hash rate generated by a machine to multiple pools. This feature may be useful for a wide variety of mining operators.
The math is simple enough to support payout configurations as simple or complex as needed. To set up hash rate splitting:
- Configure all of the pools and workers that will receive hash rate.
- Determine the allocations of hash rate per pool (for simplicity, use whole numbers that add to 100).
- Enter these numbers into the quota box of each group.
Luxor Firmware's hashrate fractions are calculated dynamically. If a mining pool becomes unresponsive, the splitting feature automatically adjusts to distribute hash rate to the remaining mining pools accordingly.
Example In Figure 1, the quotas all add up to 100, so the percent of hashrate contributed to all 4 mining pools matches the quotas. Figure 2 shows the updated percentages to account for the removal of Pool 2 as a recipient of hashrate.
Figure 1: All Pools Active
Figure 2: Pool 2 Unresponsive