lolMiner 1.60

Note: Windows version re-uploaded Oct 7th 19:40 UTC

Changes

• Significantly improved performance on Flux for Nvidia Turing & Ampere and AMD RX 5000 based GPUs. The expected performance on Ampere and Navi GPUs is about 15-20% above common implementations. On Turing based GPUs the performance matches common codes, but is more energy efficient. Fee is 1.5% for the new codes.
• Equihash 125/4 (Flux), 144/5 and 192/7 stratum can now distinguish between stale shares and other rejected.

Fixes

• Smaller fixes in Kaspa stratum, removing the extra string lines when a share gets rejected and fixing a potential infinite loop that might consume CPU time.
• Windows re-upload: Fixed ETHW / ETC + Kas dual mining on Nvidia GPUs

lolMiner 1.59a

Changes

• Significantly improved Kaspa performance and efficiency on Nvidia Turing and Ampere GPUs. Also the new codes are now working much better when GPUs are power limited.
• Slightly improved Kaspa performance and efficiency on AMD Navi and Big Navi GPUs
• Kaspa performance statistics should now be a bit more smooth

Fixes

• (1.59a) fixed a bug with 1.59 causing defect shares on acc-pool

lolMiner 1.58

Changes

• Added Kaspa only mining mode for Nvidia Pascal and new and AMD Polaris, Navi and Big Navi. Use --algo KASPA to mine it. Fee is 0.75%.

Tuning advice: On Nvidia Turing and Ampere GPUs this algorithm works best when locking the core clock to desired value (higher gives more performance, lower reduces power draw). We recommend to not set the power limit. Also to save energy, the memory clock can be chosen rather low. On Linux you can use --mclk 810 (but when you change back to an other algorithm you might need to reset this or reboot the rig).

Note: If your OS does not list lolMiner supporting Kaspa only mining yet, use the configuration for TON mining (cause temporarily we will accept --algo TON to mine Kaspa.

lolMiner 1.57

Changes

• An other performance improvement of Flux mining on Ampere based GPUs (1.5 - 3.5% depending on model and clocks)

Fixes

• Fixed a bug causing some Ergo pools to disconnect the miner frequently (Introduced with 1.56 :/ )

lolMiner 1.56

Changes

• Major performance increase for Flux mining on Nvidia Ampere GPUs
• Major performance increase for Beam mining on Nvidia Ampere GPUs
• Mining Flux and Beam on all Nvidia GPUs now uses Cuda instead of OpenCL
• Slight performance improvements for Flux mining on AMD RX 500, RX 5000 and RX 6000 (about 0.2 it/s)
• Mining any Equihash based coin (Flux, BTG, Aion, ...) now supports extra nonce subscription via stratum (e.g. for better profit switching pool experience)
• Added option --no-cl to disable detection of OpenCL powered GPUs (fixes a crash on some rigs where the Nvidia OpenCL driver crashes - happens mostly when having a rig with many GPUs)

Fixes

• Disabled DAG checking for ETC mining on 4G AMD GPUs, because this was always failing (investigating to bring it back asap)
• Nvidia LHR cards will no longer run the LHR unlock when not mining Et(c)hash or Ergo - this will prevent them from crashing on some other algorithms.

lolMiner 1.55a

Changes

• Significantly improved Flux mining performance on AMD RX 5000 (+10% on 5700) and AMD RX 6000 based GPUs (+15-22%)
• Mining Flux, Beam, Equihash 144/5 and Equihash 192/7 will now display the iteration/s (it / s) for easier OC tuning and the pool sol/s calculated from submitted and accepted shares.
• Flux mining can now be selected via --algo FLUX without needing to specify --pers.
• Enabling Equihash 144/5 and Equihash 192/7 will now also support --pers auto instead of capital --pers AUTO for enabling pool automatic selection.
• Added DAG check and repair function for Ethereum Classic mining up to epoch 300
• Added ETH / ETC + Kaspa code for Nvidia Pascal based GPUs.
• Added option to mix different Ethash style algorithms when using the fail-over pool function. To use this you now can use eth:, etc: and ubiq: as a prefix for your fail-over pool address to indicate this pool uses a different algorithm then the one configured initially. This will allow to automatically switch from Ethash to an Etchash pool when "the merge" happens as one can configure for normal Ethash mining and configure one of the fail-over pools as Etchash, so when the primary pool gets disconnected or rejects the authorization (these two are expected behavior for Eth pools after the merge), then the miner can switch to a fail-over e.g. configured to mine Etchash. See example files "mine_eth_backup_etc.sh" or "mine_eth_backup_etc.bat" on how to use it.
  Note: when using this feature with prefixes like tls:// or ssl://, then put the new prefix after the prefix for the communication layer. So tls://etc: is valid, while etc:tls:// is not.

Fixes

• Fixed a bug that might cause the miner to crash when an Ergo pool connection needs to reconnect.
• Fixed a bug that can cause random invalid shared when mining Kaspa on AMD GPUs.

Fixes of 1.55a

• Fixed a bug causing Ergo mining not to startup on some Nvidia Rigs
• Fixed a bug causing Flux mining not to startup on some Rigs with RX 400 / 500 GPUs

lolMiner 1.54

Fixes

• Fixed a bug sometimes causing duplicate send shares in Kaspa dual mining.
• Fixed a bug "Received a defect stratum message: conversion of data to type "b" failed" on new Kaspa pools and the solo mining adapters.
• Fixed a bug causing --dualworker not to work right with Kaspa dual mining.

lolMiner 1.53

Changes (1.52 to 1.53)

• Added Eth/Etc/Ubiq + Kaspa dual mining solver for AMD RX 400, 500, 5000 and 6000 series and Nvidia RTX 2000 / 3000 series. Use --dualmode KASPADUAL to select it. Fee is 1% + 0%.
• Slightly improved dual mining performance of Ethash + Alephium on AMD RX 5000 and RX 6000 series
• Slightly improved energy efficiency of Ethash on AMD RX 5000 and RX 6000 series
• Added new parameter --max-latency. If the pool share latency is above this value and failover pools are configured, the miner will terminate the connection and connect to the next failover pool (at the earliest after 10 shares on the active connection). This is repeated if necessary until a pool remains below the latency limit. The default value is 0, which disables the feature. In dual-mining, multiple values can be specified separated by a comma, where the first value is for the first algorithm and the second is for the second algorithm. If only one value is specified, it applies to all connections.

Changes (1.53 compared to 1.53 beta 3)

• Added the --max-latency parameter (see above)
• Improved efficiency of the Nvidia Eth+Kaspa dual mining solvers.

Known issues:

The miner seems to crash on some Windows machines running ETC+Kas on 4G cards. We are not yet sure why, but we will investigate.

lolMiner 1.52a

Fixes

• Fixed a bug that can sometimes cause the pool hash-rate to half after multiple reconnect attempts in a row.
• Adjusted behavior of LHR unlock function so solve issues with some configurations occurred in 1.52.

lolMiner 1.53 public beta 3

Changes (beta 2 to beta 3)

• Added Eth/Etc/Ubiq + Kaspa solver for Nvidia Turing (RTX 2000, CMH) and Nvidia Ampere (RTX 3000, RTX Ax000) based GPUs.
  Note: on Turing the optimization curve for automatic tuning is very flat - therefore on same cards quite different --dualfactor values can be selected. We recommend to run it once and then select your preferred value manually.

Changes (beta 1 to beta 2)

• Significantly improved Kaspa performance in the Eth+Kas solver on RX 400/500. Also improved the possible Eth / Kas ratio on RX 5000 and RX 6000 GPUs or - when running at same ratios - reduced the power consumption.
• Improved Alephium performance in the Eth + Aleph solver on RX 5000 and RX 6000 GPUs
• Slightly reduced clock demand on Ethash only for RX 5000 GPUs
• Added experimental ETC+Kaspa solver for R9 285 / R9 380 (Tonga), RX Fury (Fiji) and RX 460 / 550 (Baffin) GPUs

Fixes

• Fixed "conversion to type 'b'" error message in Kaspa stratum module
• Fixed miner displaying wrong version number and release month at startup
• Fixed miner setting up a dev fee connection for Kaspa - that was just unused, but occupied network resources

Changes of beta 1 to 1.52a

• Added Eth/Etc/Ubiq + Kaspa dual mining for AMD RX400, RX500, RX5000 and RX 6000 series. Use --dualmode KASPADUAL to request this algorithm. Requires --dualpool to be set to a Kaspa stratum pool. GPUs that do not support this dual mode will automatically mine Eth/Etc/Ubiq only. This algorithms prefers the newer architectures. Fee is 1%+0% as for all dual mining modes.