Fl4shMiner v1.2.7

• Emergency hardfork for Pearlhash - MANDATORY UPDATE

• Supported cards for now:
  RTX 30 Series
  RTX 40 Series
  RTX 50 Series

• Other cards will be added in the next releases

Fl4shMiner v1.2.6

• Improved PearlHash hashrate across multiple NVIDIA GPU generations again:
  GeForce RTX 30 Series
  GeForce RTX 40 Series
  GeForce RTX 50 Series

• Adjusted the dashboard layout.

• Fixed the Diff display

Fl4shMiner v1.2.5

• Improved PearlHash hashrate across multiple NVIDIA GPU generations:
  GeForce RTX 30 Series
  GeForce RTX 40 Series
  GeForce RTX 50 Series

• Include the dedicated SM75/Turing FusionHash runtime in both the HiveOS/Linux and Windows all-architecture launchers.

• Fixed a bug when mining Fusionhash on mixed GPUs on the same rig.

Fl4shMiner v1.2.4

PearlHash

• Migrated PearlHash to consensus-safe, rank-128 kernels optimized specifically for NVIDIA SM86, SM89, and SM120 architectures. Support for additional GPU architectures is planned for future releases.
• Added conditional Kryptex/F2Pool gzip v2 negotiation. The miner now enables gzip v2 only after explicit pool confirmation and continues using the raw proof protocol for all other pools.
• Linux and HiveOS requirements to NVIDIA driver version 580 or newer.

Supported and Live-Tested PearlHash Pools

• HeroMiners
• Kryptex
• F2Pool
• LuckyPool
• AlphaPool
• 2Miners
• JETSKI
• K1Pool

Fl4shminer v1.2.3

• Reduced the stale share rate on Kryptex and F2Pool
• Fixed an issue with the reconnect function
• Fixed an issue that prevented the miner from starting when an unsupported GPU was detected

Fl4shMiner v1.2.2

• Fixed an issue that prevented Kryptex usernames from being used in the wallet field.
• Improved the efficiency on RTX-3000 Series gpus
• Tested pools: Kryptex, HeroMiners, LuckyPool, AlphaPool, 2Miners, JETSKI/Pearlski, K1Pool

Fl4shMiner v1.2.1

• Added support for the PearlHash algorithm used by Pearl (PRL).
• Added optimized NVIDIA CUDA kernels.
• Added support for NVIDIA GPUs based on the following architectures:
• RTX 30 Series - Ampere
• RTX 40 Series - Ada Lovelace

Support for additional NVIDIA architectures will be added in future releases.

Tested performance:

GPU	Hashrate	Core Clock Offset	Locked Core Clock	Memory Clock	Power Limit
RTX 3060 Ti	~67.40 TH/s	+225 MHz	1825 MHz	Locked at 5001 MHz	168 W
RTX 3070	~74.50 TH/s	+250 MHz	1650 MHz	Locked at 5001 MHz	175 W
RTX 3080	~108 TH/s	+250 MHz	1605 MHz	Locked at 5001 MHz	265 W
RTX 3080 Ti	~128 TH/s	+250 MHz	1605 MHz	Locked at 5001 MHz	299 W
RTX 4090	~300.00 TH/s	+250 MHz	2500 MHz	Locked at 5001 MHz	397 W

Dev fee 1.5%

Hashrate and power consumption may vary depending on the GPU model, silicon quality, driver version, operating system, overclock settings, and temperature.

### CPAY, CSD and FXL users: keep using v1.2.0, because v1.2.1 caused a regression on these algorithms. I'm investigating it.

Update: Fixed . The regression affecting CPAY, CSD, and FXL has been resolved. You can safely upgrade.

Fl4shMiner v1.2.0

• Fixed HiveOS status reporting on mixed rigs.
• Cleaned up dashboard labels: CORE TMP → CORE T and MEM TMP → MEM T.

Fl4shMiner v1.1.9

NVIDIA RTX 50 Series Performance Improvements

CSD

• Improved CSD performance by approximately 25% on NVIDIA RTX 50 Series GPUs.

• The optimized CSD path applies across the supported RTX 50 Series. Actual performance gains may vary by GPU model, clock configuration, power limit, and thermal conditions.

• RTX 5080 benchmark:

  • Previous: 6.26 GH/s
  • New: 7.87 GH/s
  • Improvement: +25.7%

Cryptix

• Improved Cryptix performance by approximately 111% compared with the legacy CUDA kernel.

• Approximately 17% faster than the previous OpenCL implementation.

• The optimized Cryptix path applies across the supported RTX 50 Series. Actual performance gains may vary by GPU model, clock configuration, power limit, and thermal conditions.

• RTX 5080 benchmark:

  • Previous: 224 MH/s
  • New: 472 MH/s
  • Improvement: +110.7%

RTX 5080 Benchmark Settings

• Core clock offset: +400 MHz
• Locked core clock: 3,200 MHz
• Locked memory clock: 7,001 MHz

The benchmark results above were measured on an RTX 5080. The same CSD and Cryptix optimizations are enabled for the rest of the supported NVIDIA RTX 50 Series.

Fl4shMiner v1.1.8

• Added new algo : SHA256d (CSD)
• Optimized Kernel for NVIDIA cards
• Support NVIDIA CUDA GPUs from Maxwell through Blackwell:
• GTX 900 series — Maxwell
• GTX 10 series — Pascal
• GTX 16 and RTX 20 series — Turing
• RTX 30 series — Ampere
• RTX 40 series — Ada
• RTX 50 series — Blackwell
• Compatible NVIDIA workstation, datacenter and Jetson GPUs

Tested performance:

GPU	Hashrate	Core Clock Offset	Locked Core Clock	Memory Clock	Power Limit
GTX 1080 Ti	~1.90 GH/s	+150 MHz	1911 MHz	Locked at 810 MHz	205 W
RTX 3060 Ti	~1.90 GH/s	+200 MHz	1825 MHz	Locked at 810 MHz	100 W
RTX 3070	~2.17 GH/s	+200 MHz	1825 MHz	Locked at 810 MHz	115 W
RTX 3080	~3.43 GH/s	+200 MHz	1800 MHz	Locked at 810 MHz	199 W
RTX 3080 Ti	~4.07 GH/s	+200 MHz	1800 MHz	Locked at 810 MHz	220 W
RTX 4090	~9.60 GH/s	+200 MHz	2800 MHz	Locked at 5001 MHz	389 W

Dev fee 1%

Edit: fixed a little bug when trying to mine solo in some pools. Version still the same, if you are trying to mine solo delete the miner and redownload this updated version.