Changelog
Xelis
performance improvement on Nvidia CMP 170hx card (70+ kh/s)
minor performance improvement for Vega cards (depends on oc settings)
Dynex
drop dev fee to 2%
minor bug fixes
Changelog
Xelis:
Performance improvement for RDNA gpus
Minor improvement for Vega 10 and Vega 20
Add Polaris support
Fix RDNA3 initialization
Fix no cuda device found error
Remove the libdrm dependency
Minor bug fixes
Changelog
Xelis:
Minor performance improvement on some cards
Zil:
Reduce the switching delay with applying the OC settings in parallel
Changelog
Xelis:
Efficiency improvement
Add support for solo mining (--xelis-solo)
Add algo alias xelishashv2
Changelog
Xelis:
Add support for xelishashv2(2% dev fee)
Remove xelishashv1
Dynex:
Disable snapshot by default
Add --enable-snapshot flag
Drop the support for GPUs with compute capability lower than 6.0
Changelog
Dynex: Add support for the new enhanced core algorithm
Xelis:
Performance Improvement
This version only supports pool mining.
Changelog
Xelis:
Add support for Stratum protocol (Vipor proposal)
Add --xelis-solo option. Default is Stratum pool mining unless this flag is set or the server port is 8080
Changelog
Xelis:
Performance improvement
Fix an important bug related to the watchdog
Multiple fixes with +Zil
Drop the default max-memory to 70 (use --max-memory to adjust if you need to)
Minor bug fixes
Changelog
Add support for Xelis(xel) coin(Solo mining with 3% dev fee)
Dual mining with Xelis+Zil(refer to the sh/bat files in the packages for sample commands)
Multiple nodes support
Add --max_memory(alias to --chips-memory) for controlling the amount of allocated GPU memory
Changelog
Add ability to take a snapshot of the best found state
This feature will save the lowest local minima found every 5 minutes and is capable of loading it in case the miner exits/reconnects for any reason. It can improve the chance of solving SAT jobs and getting the reward.
Add options --disable-snapshot to disable the feature and --purge-snapshot to entirely remove the cache folder
Fix invalid shares that could happen on jobs other than the main SAT one
Fix incorrect allocation of chips when the number of available chips is lower than the rig's capacity
Fix --nr issue on Windows that could cause lower hashrate on some cards
Minor bug fixes
Changelog
Add support for dynexsolve 2.3.5f protocol
Hashrate improvement on some cards
Fix a fallback pool issue when the primary server is ssl
Multiple fixes related to the reconnection for changing the jobs
Minor bug fixes
Changelog
Minor hashrate/efficiency improvement on Dynex
Fix an issue with Zilliqa that could result in duplicate shares
Overall improvements on the stability of Dynex
Minor bug fixes
Changelog
This version is mainly to support dual mining Dynex and Zil.
Add support for Zilliqa(0% fee)
Changelog
Hashrate/Efficiency Improvement on some gpus
Added new tuning parameter --nr
This is a per gpu option and is related to the number of rounds done in one iteration.
Default is 1 and the values are 1,2,4 and 8. The best value for almost all gpus is either 1 or 2.
Make sure you try both --nr 1 and --nr 2 and let the miner find the best --nt value to get a baseline.
Note that higher --nr values might lead to more power fluctuations and higher stale shares.
Changelog
Hashrate/Efficiency Improvement
Added new tuning parameter --nt
This is a kernel launch parameter and the value can be 1,2,4,8,16 or 32. This can be set per gpu just like OC options.
Added an auto tuner for parameter nt
Revised atomic update failures to allow 5 minutes of running with no atomic update
Minor bug fixes
Changelog
Added support for SSL/TLS
pool urls now support ssl:// or stratum+ssl:// prefix to enable ssl
Added new option --no-cert-validation to skip SSL/TLS certificate verification.
Added rejected shares monitoring to watchdog
Revised watchdog's issue handling
The miner does a full restart of mallob and pool connections on watchdog issues.
Removed the OpenSSL 1.1 dependency and updated internal OpenSSL library to the latest version.
OpenSSL is now statically linked
Deprecated --max-no-job option since it is redundant and max-no-acc does the same job.
This option is disabled in this version and will be removed in the next release.
Added more dev pools and randomization of dev pool selection to reduce the load on pools servers
Fixed a bug that could freeze the miner on multiple mallob reconnections.
Minor bug fixes
Changelog
Minor hashrate improvement(~5%) on some gpus
Fixed an issue that could result in 414 errors on mallob re-connection
Improved overall stability on mallob re-connection
Minor bug fixes
Changelog
Hashrate improvement mostly on 20xx and 30xx gpu series
Added a more robust file downloader with progress bar
Added ability to remove corrupted or unneeded job files
Changelog
This release is mainly to support dynexsolve 2.3.0.
Support for ML, SAT and MAXSAT jobs
Adaptive stepsize
Randomized parameters
Double precision accuracy
Note:
The hashrates of this version is significantly lower than v1.1.3 because of all the changes made.
Changelog
Changelog
* Significant hashrate increase on all gpus.
* More than 100% hashrate on 3000 series. (eg: 3090 does 14 kh/s with stock settings)
Note: The hashrate increase of this release is not because of any actual improvement. It is simply because we are using single precision (doing math calculations up to 7 decimals) for simulating chips compared to double precision(up to 15 decimals) that was used in the previous versions. Double precision is important for dynex jobs especially the machine learning job and we will switch back to it in the future releases. Meanwhile, enjoy the hashrate boost!
Changelog
Fixed a bug with solution submission
Fixed a hashrate reporting issue with small jobs
Increased mallob connection timeout
Changelog
* Hashrate improvement on most GPUs
* Fixed Windows build issue
* Added Minerstat API to HiveOS package
* Added an average period for the reported power consumption.
Power is now calculated by averaging over the past five seconds.
* Minor bug fixes
Changelog
Hashrate improvement on some GPUs. Added overclocking support: --coff --moff --cclk. Only supported on 20xx series and newer models: --mclk. Only supported on 30xx series and newer models: --pl --fan. All of these options are comma separated lists. Use * to skip a gpu. Added full stats of shares, clocks, efficiency, fan and power consumption. Added watchdog options: --max-no-job --max-no-acc. Added support for failover pools. Added support for multiple mallob servers. Fixed poolside hashrate loss on stratum reconnection. Fixed an issue related to timestamps. Revised the handling of atomic update failures. The miner will try for a successful update in 15 minutes instead of shutting down. Added multiple dev pool servers. Minor bug fixes
Changelog
Hashrate improvement: 4090: 10 KH/s. 3090: 5.2 KH/s. 2080ti: 3.7 KH/s. 1080ti: 3 KH/s. Were measured with stock settings and no miner parameters. Fixed hashrate loss issue on 40xx cards. Add support for compute 7.0 (Titan V). Fixed a CPU verification issue. Added miner uptime stat. Minor bug fixes
Changelog
Fixed 8 GPUs limit
Fixed an issue that could cause local minima 0
Added a more robust job downloader
Added job3 file to packages
Added POUW stats to logs
Added decimal support for chips-memory option
Removed intensity option as it's confusing and unnecessary
Some minor bug fixes
Added support for mmpOS.
The HiveOS package should work on mmpOS too