chore: add transfer benchmark #90
Conversation
achingbrain
force-pushed
the
main
branch
6 times, most recently
from
fbeed88
to
9f9bc60
Compare
| const subjectA = await implA.create() | ||
| const subjectB = await implB.create() | ||
|
|
||
| const addr = await subjectB.addr() | ||
| await subjectA.dial(addr) | ||
|
|
||
| const cid = await subjectA.add(bufferStream(file.size), file.options) | ||
|
|
There was a problem hiding this comment.
Are the CIDs for each file in files unique? If so, we could move the creation of subjectA and subjectB outside of this inner-most for loop to right after the implA->implB log.
There was a problem hiding this comment.
if not, it might be useful to utilize tinybench's beforeEach so it can handle timings appropriately
There was a problem hiding this comment.
The CIDs are randomly generated but the nodes are set up/torn down on each iteration to ensure we're starting from fresh each time.
We're only measuring the transfer speed here so this doesn't affect the outcome.
Adds a benchmark suite for doing various size data transfers between helia, kubo and js-ipfs. Closes #88
achingbrain
force-pushed
the
chore/add-transfer-benchmark
branch
from
5e4b67d
to
0e95ef8
Compare
|
Some stats from the latest release: Using bitswap Helia is consistently faster transferring to and from other Helia nodes, and it's consistently faster reading data from Kubo. Kubo is consistently slower reading data from Helia and transferring to and from other Kubo nodes. 
|
Adds a benchmark suite for doing various size data transfers between helia and kubo.
Closes #88