v1.6.22

Performance improvement 10%

v1.6.20

Continuous performance improvement, enjoy!

v1.6.12

Continuous performance improvement, big surprise, enjoy!

v1.5.18

Bugfix and performance optimization

v1.5.12

We are still working hard for the performance improvement, about 10%～20% we make it again, so check it by yourself.

N.B. For some gpu out of memory issue, pls. change the NONCE_TIMES value according to your GPU memory, default value is 12, you can reduce the value to 10, 8, 6, 4..., bigger is better.

v1.4.5

Performance improvement more than 20%

v1.4.0

Performance improvement more than 10%, GPU working mode improvement is much better, enjoy!

v1.3.10

Add on cluster and device parameters for f3-prover, which can be checked from the log of pool endpoint.

v1.3.6

Performance improvement

1. Optimize the GPU usage
2. Optimize the binary file size
3. Improve performance during the hard puzzle computation
4. Output the epoch difficulty during the puzzle computation

v1.2.4

1. bugfix for -g option
2. add on merkle tree leaves size output
3. a new release with adx instruction support

Description:
--- epoch leaves = 23245 with adx enable
2024-07-01 17:20:38.346 INFO 1.2.4 2001010 perf: 39968 (1m: 39.90 P/s, 5m: 39.87 P/s, 30m: --- P/s, 60m: --- P/s)
2024-07-01 17:21:38.346 INFO 1.2.4 2001010 perf: 42360 (1m: 39.87 P/s, 5m: 39.87 P/s, 30m: --- P/s, 60m: --- P/s)
2024-07-01 17:22:38.346 INFO 1.2.4 2001010 perf: 44752 (1m: 39.87 P/s, 5m: 39.87 P/s, 30m: --- P/s, 60m: --- P/s)

--- epoch leaves = 23245 without adx
2024-07-01 17:21:09.731 INFO 1.2.4 2000995 perf: 43283 (1m: 38.77 P/s, 5m: 38.75 P/s, 30m: --- P/s, 60m: --- P/s)
2024-07-01 17:22:09.731 INFO 1.2.4 2000995 perf: 45614 (1m: 38.85 P/s, 5m: 38.78 P/s, 30m: --- P/s, 60m: --- P/s)
2024-07-01 17:23:09.731 INFO 1.2.4 2000995 perf: 47940 (1m: 38.77 P/s, 5m: 38.79 P/s, 30m: --- P/s, 60m: --- P/s)

To check the cpu adx feature support: lscpu | grep adx, the "adx" will be in the Flags string.