V1.4.21 Release Notes

Node Running
1

Happy Release Friday! v1.4.21 has been released for signing. This is the final release before 2.0 and all systems go!

Quick changelog notes:

  • Customizable connection management settings for high end servers and bootstrap peers
  • Deep performance drilldown for BLS48-581, details below.

Before (Golang)

Prior (golang) commits:

poly_size time
16 43ms
128 336ms
1024 2808ms
65536 165.651s

Proofs:

poly_size time
16 43ms
128 359ms
1024 2.826s
65536 178.355s

Before (Rust)

Prior (rust) commits in ms:

poly_size time
16 12ms
128 93ms
1024 736ms
65536 47.521s

Proofs:

poly_size time
16 16ms
128 133ms
1024 1.102s
65536 73.836s

After

Commits:

poly_size time
16 166µs
128 166µs
1024 200ms
65536 9.592s

Proofs:

poly_size time
16 164µs
128 166µs
1024 331ms
65536 17.402s

The final amount of juice to squeeze on this would be via parallelization of the DFT – should be able to bring down the time down on the 65536 case by a strong degree.

14 Likes
2

Quilibrium Guide updates (Last 2 weeks before 2.0 Mainnet !!)

:zap: Upgrading to Quilibrium v1.4.21 from .20-p1 – Quilibrium Guide
:zap: Upgrading to Quilibrium v1.4.21 from versions .19 and older – Quilibrium Guide
:zap: Setting up your Node – Quilibrium Guide

4 Likes
3

Is that parallelization on the roadmap already? And if so, do you have an estimated release for it?

4

I have ten servers, r7252, amd 7h12, 64 cores*2=256 threads, 512g+ memory, 1t+ storage. The reward is only 22 quils per day, while the server with the same network and configuration rewards 100 quils per day. Why is there such a difference? After upgrading to version 21, the problem still exists.

WechatIMG1349
WechatIMG13491920×1440 325 KB

5

latepass here, but so excited for mainnet to launch. huge congratulations on the progress here.