Repository navigation
Releases: swdots/bitaxe-benchmark-webgui
Release list
v0.2.4
v0.2.3
- Adds dashboard option to pause polling when the window is hidden, as well as the option to clear all accumulated coins when the window is hidden
- Improved dashboard memory handling
- On window un-hide, the system should poll and spawn share coins proportionally up to 50% of the total max coins allowed
TabChange_01.mp4
v0.2.2
- Fixed a database issue that prevented new deployments from setting up correctly after the braiins api was added.
- Repaired database issues with older versions, dashboard database should identify the issue and dynamically fix if present.
- Improved system detection and identification in network scan - some non-miner systems that respond similarly to the api queries can be misidentified still as Braiins OS systems - adding them to the dashboard should not cause any issues, and they can be removed.
- Fixed the "add by ip address" button
- Added a debug api endpoint for the dashboard to report the database structures for future bug clarity.
v0.2.1
Improved dashboard system compatibility with Avalon systems - tested on nano3s. The dashboard cards for the nano3s should show the "TAvg" as the asic temp approximation, and the "OTemp" as an output air temp. Fan speed is directly pulled from the api, while the power reading is pulled from the last value in PS[...] when querying estats.
v0.2.0
v0.1.9
- Adds support for the avalon q api in the dashboard. Currently unvalidated, needs to be tested.
- Adds a new share color for the coin animations - when a new highest difficulty is achieved, the share is blue. Toggleable between session highest difficulty and overall highest difficulty.
- Adds a star and gold best difficulty when a system has found at least one block - can be deactivated in settings.
v0.1.8
This version adds a "Block Odds" section in the top dashboard bar - the network difficulty is pulled from a running system if one exposes it in the api - if not, it is fetched online and cached for 12 hours. Odds are recalculated with the current hash rate total every minute, or if there is a 10% or 500GH/s change in the total - this is to stabilize the readings a bit, but keep them reasonably real-time.
This version also adds red coins for rejected shares!
v0.1.7
dashboard_03.mp4
- Added a "Physics" option for the coin animations on the dashboard - user defined coin maxes are respected, but this allows coins to pile up at the bottom of the screen as shares are found.
- Fixed a small memory leak due to previously creating excessive event listeners for menu nav.
v0.1.6
Added a dashboard to monitor the current activity of multiple miners. You can set your own background image and sound effect for shares, and can enable a falling coin animation when shares are found.
- Multi-miner overview: Live status cards for NerdQAxe/Bitaxe-style miners with hashrate, power, temps (ASIC/VRM), fan speed, shares, and best difficulty. This uses the /api/system/info for information, and so far is only tested on the Nerdqaxe++ and Bitaxe Gamma.
- Live polling: Keeps the last known good readings on-screen during brief network hiccups, and marks devices offline only after sustained failures.
- Benchmark visibility: Devices automatically show a pink “Testing” status while a benchmark is running.
- Clean / Cards-only mode: Optional streamlined dashboard view that hides the top controls after you're done adding systems.
- Customizable units & formatting: Toggle hashrate display between GH/s and TH/s, choose hashrate decimal precision, and control when rejected shares turn red via a configurable % threshold (default 1%).
- Theme & transparency controls: Theme toggle plus a card transparency setting for better readability over backgrounds.
- Sound + variation: Optional share sound notifications with adjustable pitch randomness to reduce “same-y” alerts.
dashboard_02.mp4
v0.1.5
#Improved grid testing#
- Added expected hash rate, percentage of expected hash rate achieved, and rejected share percentages to the heatmap options
- Added a system that allows the user to highlight cells in the heatmap that meet stability/headroom criteria, to better select an "optimal" point instead of the highest raw hash rate - the test run still applies the highest hash rate settings when the test completes, but a more balanced approach may achieve lower rejection rates
