Record🤔

uri:

"at://did:plc:hdhoaan3xa3jiuq4fg4mefid/app.bsky.feed.post/3lc3r7i75ls2i"

cid:

"bafyreigi4jrdwayx242terzed7mkh3yhj6raeqjjpc6do2zgykbs7xmfeq"

value:

text:

"ok friend you asked for it

after a failed ordered-ish keys run due to my laptop being old and sad (corrupted sqlite file!!!) i put it on speedrun mode (jetstream 24h backfill)

things did not get better for fjall. trending to keep using 4x disk after i think around 50M updates. 30GB vs 7.5"

$type:

"app.bsky.feed.post"

embed:

$type:

"app.bsky.embed.recordWithMedia"

media:

$type:

"app.bsky.embed.images"

images:

• alt:

  "chart of disk space used vs runtime for fjall and sqlite with the same dataset.
  
  both grow somewhat steadily, but fjall maintains a large margin above sqlite for the entire duration"

  image:

  

  View blob content

  $type:

  "blob"

  ref:

  $link:

  bafkreiftdrnc3yjvqeyzlfrbxzltqzeoy4mft2gohzrhoxq53k57hktbjq

  mimeType:

  "image/jpeg"

  size:

  229033

  aspectRatio:

  width:

  2000

  height:

  869

• alt:

  "scatter plot of the same data with sqlite on the x-axis to make the chart time-independent. the slope of the trend line (r^2=.99) is 3.86: this represents how much more disk fjall is trending to use than sqlite for the same dataset."

  image:

  

  View blob content

  $type:

  "blob"

  ref:

  $link:

  bafkreieb2y5nzgq5rizwuxgahlxdaquuamwtq6wvnygpmasxt4lfykfzga

  mimeType:

  "image/jpeg"

  size:

  424375

  aspectRatio:

  width:

  920

  height:

  1246

record:

$type:

"app.bsky.embed.record"

record:

cid:

"bafyreidfddnlbczigt2kztkezesa6s42kxxcyya3sexqmfbfhjvkofadz4"

uri:

"at://did:plc:xyik4pb36ai666wpyume27sx/app.bsky.feed.post/3lc2s722zes2u"

langs:

• "en"

createdAt:

"2024-11-29T14:12:43.789Z"