libera #monero-pow

21 Oct 2023

Monero's Proof of Work algorithm (RandomX) and related topics. Use #p2pool-log and #p2pool-mini for P2Pool discussions. Anti-spam active: PM hyc for voice.

• 00:33
  hyc
  hi sorry I'm out all this weekend. sech1 were you able to kill the redundant cache flushes for this release?
• 07:17
  sech1
  hyc no, it's still there. I don't think it hurts performance much, the benchmark showed more than 2200 h/s on M1
• 10:53
  selsta
  riscv gitian fails here github.com/monero-project/monero/bl…/common/compat/glibc_compat.cpp#L93
• 10:53
  selsta
  glibc_compat.cpp:(.text+0x30): undefined reference to `glob_old'
• 10:54
  selsta
  I thought about adding a similar line with `#elif defined(__riscv)` but don't know how to find the GLIBC version
• 11:52
  moneromoooo
  external/db_drivers/liblmdb/mdb.c:#define GLIBC_VER ((__GLIBC__ << 16 )| __GLIBC_MINOR__)
• 11:56
  m-relay
  <t​obtoht:monero.social> selsta: remove -DBACKCOMPAT=ON for riscv64 builds. glibc only added riscv support in 2.27
• 12:03
  selsta
  thank you, trying that now
• 12:54
  selsta
  tevador: can you test the if the gitian binary works on your risc-v system? github.com/selsta/monero/actions/runs/6597189374
• 13:06
  m-relay
  <t​obtoht:monero.social> matrix.monero.social/_matrix/media/…ero.social/INAXBnvIykbDirAJSXyDcwjf
• 13:06
  m-relay
  <t​obtoht:monero.social> :D
• 13:09
  selsta
  nice :) now I just have to confirm that the hash stays consistent
• 13:38
  m-relay
  <t​obtoht:monero.social> sent a test transaction with cli, seems to be working fine
• 13:38
  m-relay
  <t​obtoht:monero.social> getting I/O errors with the daemon but this might be attributable to whatever F-tier sd card I'm running this on
• 13:44
  tevador
  the monerod binary seems to be running fine
• 13:57
  sech1
  Nice. Now we just need some RISC-V Raspberry Pi 5 level board
• 13:59
  selsta
  if anyone else wants to confirm the hash, I tried with two separate Github Action VMs and they matched
• 13:59
  selsta
  ./gitian-build.py --setup --docker --url github.com/selsta/monero github-actions riscv-gitian
• 13:59
  selsta
  ./gitian-build.py --docker --detach-sign --no-commit --build -j 3 -o l --url github.com/selsta/monero github-actions riscv-gitian
• 13:59
  selsta
  688fd44ffa1dd27b8a48e2abc1ead1e0bfb048f6acd9e7ac97b200e2d3697b48 monero-riscv64-linux-gnu-c56932183.tar.bz2
• 14:02
  tevador
  btw, monerod estimates 28 days to sync on my board... no sure how it compares to raspberry pis
• 14:06
  selsta
  monero-project/monero #9028
• 14:07
  m-relay
  <t​obtoht:monero.social> selsta: looks good, just running a build to see if hash matches
• 14:21
  m-relay
  <t​obtoht:monero.social> hash matches
• 14:21
  selsta
  ty, also added the steps to check if hashes match in case someone else wants to try it
• 14:23
  plowsof
  iirc the estimate can not see into the future, so its wildly inaccurate to begin with (it assume all blocks are emptyish in the beginning) but im not certain
• 16:48
  hyc
  hmm, lemme plug in a USB SSD so I have somewhere to store the blockchain
• 20:01
  hyc
  hmm, my commit id was 9ea325777 for your PR. not c56932183
• 20:03
  hyc
  so my build hash doesn't match yours
• 20:03
  hyc
  selsta ^ where did your commit ID come from?
• 20:06
  hyc
  ah I see. from your test branch
• 20:10
  hyc
  ok trying again
• 20:10
  hyc
  OPT="-o l -j8" ./dockrun.sh c56932183
• 20:19
  hyc
  yeah hash matches
• 20:25
  hyc
  started monerod
• 20:32
  hyc
  sad. gigabit network, node traffic rarely hits 1Mbit
• 20:32
  hyc
  this is syncing with --add-exclusive-node against my rockpro64. probably the rockpro is too slow
• 20:36
  moneromoooo
  IIRC the node will download blocks till 100 MB, then wait to refill.
• 20:37
  hyc
  yeah I doubt the 4GB RAM is enough for the rockpro to serve the tip of the chain and also replay from the beginningo f the chain for the new node
• 20:38
  hyc
  so the final blocks should transfer quicker because they'll already be cached in memory