lolMiner 0.8.3

lolMiner 0.8.3

Changes from 0.8.3

• Added 16 GByte solver for Cuckatoo-31 (GRIN). Better performance and less energy use for Radeon VII, Radeon Pro Duo Polaris, Vega Frontier and RX 570 16GByte

Performance

Radeon VII ~1.15 g/s
Radeon Pro Duo Polaris ~0.95 g/s

lolMiner 0.8.2

lolMiner 0.8.2

Changes from 0.8.1

• Added support for ZelHash (EquihashR 125/4/0) for the hard fork (July 2nd)

Notes:

If you want to mine ZelCash pre-fork please use version 0.8.1 or earlier. The new algorithm is available with --coin ZEL and replaced the old.

The new parameters can already be used in the test network

lolMiner 0.8.1

lolMiner 0.8.1

Note: will build the Windows version when back at a Windows PC (later today)

Changes from 0.8

• Added support for NiceHash on Grin
• Fixed a bug with some pool software (e.g. grinmint.com)
• Reduced number of stale shares submitted

Changes from 0.7

• Added support for Grin (Cuckatoo-31) for 4G (Slean) and 8G cards. The parameter is "--coin GRIN-AT31"
• Grin cycle finding is completely done on GPU, the miner has almost 0 CPU load (on AMD, Nvidia has some due to OpenCL back end)
• Stratum bug fixes (NiceHash) for 144/5, 192/7 and 96/5
• Fixed AUTO192_7 configuration
• Improved general stability
• Added a distinct 1G / 3&4G / 6&8G kernel for mining MNX (Windows, Linux had this in 0.7.1)

Approximate Performance for Cuckatoo-31

Note: Stock clocks; fidelity is shown in miner and in API and should converge to 1

Card	8 G Code	4 G Code
RX 560	-	0.14 g/s
RX 570	0.45 g/s	0.26 g/s
RX 580	0.49 g/s	0.28 g/s
RX 580	0.49 g/s	0.28 g/s
Fury Nano	-	0.33 g/s
Vega 56	...	-
Vega 64	0.9 g/s	-
Radeon VII	1.1 g/s	-
GTX 1060 6G	-	0.27 g/s
GTX 1070	~0.6 g/s	...
GTX 1080	0.75 g /s	-
RTX 2060	-	0.29 g/s
RTX 2070	...	-
RTX 2080	...	-

Minus sign = Will not be used or is not compatible
... = Compatible but value not yet known

lolMiner 0.8

lolMiner 0.8

Changes from 0.7

• Added support for Grin (Cuckatoo-31) for 4G (Slean) and 8G cards. The parameter is "--coin GRIN-AT31"
• Grin cycle finding is completely done on GPU, the miner has almost 0 CPU load (on AMD, Nvidia has some due to OpenCL back end)
• Stratum bug fixes (NiceHash) for 144/5, 192/7 and 96/5
• Fixed AUTO192_7 configuration
• Improved general stability
• Added a distinct 1G / 3&4G / 6&8G kernel for mining MNX (Windows, Linux had this in 0.7.1)

Approximate Performance for Cuckatoo-31

Note: Stock clocks; fidelity is shown in miner and in API and should converge to 1

Card	8 G Code	4 G Code
RX 560	-	0.14 g/s
RX 570	0.45 g/s	0.26 g/s
RX 580	0.49 g/s	0.28 g/s
Vega 56	...	-
Vega 64	0.9 g/s	-
Radeon VII	1.1 g/s	-
GTX 1060 6G	-	0.27 g/s
GTX 1070	...	...
GTX 1080	0.75 g /s	-
RTX 2060	-	0.29 g/s
RTX 2070	...	-
RTX 2080	...	-

Minus sign = Will not be used or is not compatible
... = Compatible but value not yet known

lolMiner 0.7.1 (Linux only)

Updated lolMiner to version 0.7.1

Changes from 0.7

• Added a distinct 1G / 3&4G / 6&8G kernel for mining MNX
• Added a new parameter --asm. The parameter activates the use of binary kernels, see below.

Binary kernels in 0.7.1

lolMiner 0.7.1 features binary kernels for mining BEAM that are tuned with AMD assembler commands. This kernels currently require the AMD ROCm OpenCL driver in Linux (thats why 0.7.1 is Linux only).
When you run a system with this OpenCL drivers using --asm 1 or the pendant in config file, the miner will try to load the binary kernels.

The kernels feature a speedup of approx 8% on Radeon RX 580 series and 10% on Vega 56, 64 and Radeon VII. Also the energy consumption should slightly be reduced. There are kernels included for RX 470, 480, 570, 580, 590, Vega 56, 64 and Radeon VII. Fiji GPUs can be compatible but are untested yet.

It is planned for future releases to extend the support for other drivers and Windows, but these other drivers make it complicated to handle the improvements. Stay tuned.

Performance for BEAM

(Tested with ROCm 2.1 driver)

GPU	Non-ASM Kernel	ASM Kernel
RX 580	12.1 sol/s	13.3 sol/s
Vega 56	18.5 sol/s	20.5 sol/s
Vega 64	21 sol/s	23.1 sol/s
Radeon VII	24.4 sol/s	27.2 sol/s

lolMiner 0.7

Updated lolMiner to version 0.7

Changes from 0.6

• Added support for Beam (BEAM, modified Equihash 150/5), only tuned for AMD cards
• Added TLS support for stratum. The default is off for all coins except Beam but on for Beam. This adds a new parameter --tls to control TLS on / off (see manual on usage)
• Some bugfixes and reworks in the stratum code. It is more stable now
• Complete rework of the mining back end. Lower CPU load for AMD graphic cards.
• GPU sorting changed. Its now sorted by PCIE bus address. This addresses are also shown at startup and in API.
• Integrated all kernel files to the executable
• Lowered fee of ALL algorithms / coins to 1%
• Added --help parameter (needs formating)
• API update interval is now fixed to 10 seconds
• API now smooths the performance data over approx one minute

Coin specific changes:

• Removed workbatch parameter for MNX, its obsolete now
• Removed 96/5 Nvidia specific kernel
• Changed Safecoin (SAFE) from Equihash 144/5 to 192/7
• Added Vidulum (VDL), Equihash 192/7

lolMiner 0.6

lolMiner 0.6

Changes over 0.6 alpha 4:

• Fixed a bug with defect shares when running on >4 GPUs
• Slightly improved AION performance on AMD
• Added support for edit ExchangeCoin (shortcut: EXCC) (Equihash 144/5) pow structure change that happens today (!)
• Rules back to the alpha 3 MNX code for AMD. Worked better for most cards
• Reduced Equihash 144/5 fee to 1.5%

Changes over 0.5 stable:

• New codes for all coins on all cards (better speed)
• New API and statistics module (API updating every 10 seconds)
• Better stratum stability
• Improved stability for many GPU systems
• and many more small stuff I can not remember ;)

lolMiner 0.5

Update README.md