You can find all published reports here:

List of Published Reports

Overview

This test was aimed towards identifying the maximum number of blobs that each CL could handle. Therefore, we only ran single combination of node in the tests

• lighthouse/geth: 40 blobs/block
• prysm/geth: 25 blobs/block
• teku/geth: 30 blobs/block
• lodestar/geth: 25 blobs/block
• grandine/geth: 20 blobs/block
• nimbus doesn’t yet have version of PeerDAS with getBlobs, so we’ll run this test in a later time

From above tests:

• we’re investigating how geth’s blobpool behaves and how good getBlobs hit rate correlates with blob throughput
• with ideal EL blob hitrate, the clients could handle more blobs than previous tests.
  • However, there exists differences in blobs throughput between clients.

The testing methodologies are identical to the previous tests.

Lighthouse

With 50 lighthouse-geth nodes, with 25 fullnode, 25 supernode: