14:26:33 <maybefbi> i saw a a pow function like this: input -> (sha3 -> Salsa20 expansion -> bzip2) -> (sha3 -> Salsa20 expansion -> bzip2) -> sha3. even if someone develops an ASIC/GPU algorithm for it, it will speed up bzip2 which will help us handle archival of all those exabytes of data the internet is generating.
14:27:21 <moneromooo> NSA approves
14:27:43 <maybefbi> :)
14:30:49 <hyc> just like sha2-256 ASIC development has improved SHA2 performance all over the internet
14:37:51 <maybefbi> Intel SHA extensions was introduced in 2013. who knows, may be sha2 pows in many coins led to instructions like SHA256RNDS2, SHA256MSG1, SHA256MSG2
14:38:40 <sech1> I saw that pow function too
14:38:49 <sech1> Even wrote code for it in xmrig :D
14:39:28 <sech1> It's not bzip2 there, it's BWT transform
14:39:54 <maybefbi> ah yes
14:40:09 <maybefbi> i bet BWT has more applications than bzip2
14:40:52 <maybefbi> googling says "bioinformatics and computational biology, full-text compressed indexes, prediction and entropy estimation, and recent approaches in joint-source channel coding"
14:41:10 <maybefbi> and "hape analysis in computer vision, and in machine translation"
14:41:17 <maybefbi> *shape analysus
14:45:54 <gingeropolous> i mean, sure maybe? with randomx, we've put a bounty of increasing efficiency, performance, and speed of general compute.... BWT will benefit
14:46:02 <gingeropolous> *bounty on
14:47:26 <maybefbi> lets assume everyone starts using ARM after Apple moves to ARM. will randomx test vectors still produce the same output?
14:47:35 <hyc> don't be an idiot
14:47:40 <gingeropolous> jesbut
14:47:41 <hyc> of course the ydo
14:47:47 <hyc> just like they already do today
14:49:25 <sech1> I literally checked randomx mining on ARM 2 hours ago. Shares were accepted.
15:08:16 <maybefbi> thanks. i didnt know.
15:15:41 <maybefbi> WASM doesn't implement IEEE 754-2008 standard completely. "WebAssembly uses the round-to-nearest ties-to-even rounding attribute, except where otherwise specified. Non-default directed rounding attributes are not supported."
15:16:01 <hyc> tough luck for WASM
15:23:04 <maybefbi> "PowerPC processors (Macs until about 2006-2007, tons of current IBM servers) use a 128 bit format consisting of two doubles for long double, instead of the IEEE 754 extended format."
15:23:28 <maybefbi> https://www.nextplatform.com/2019/07/08/new-approach-could-sink-floating-point-computation/ "New Approach Could Sink Floating Point Computation"
15:27:22 <hyc> nobody cares
15:27:52 <moneromooo> An IBM engineer is seen crying in a corner
15:28:11 <hyc> randomX doesn't use long doubles