Hello,
I'm a masters student at Aarhus university.
We built a provably correct compiler (using Coq) targeting Wasm: Link
<https://womeier.de/files/certicoqwasm-coqpl24-abstract.pdf>
This includes a proof that the modules our compiler generates instantiate
according to the spec, i.e. they are well-typed.
We are evaluating the performance with Node.js, is there a way to improve
our performance
by e.g. disabling some runtime-checks, given that we have the stronger
correctness guarantees?
Node.js doesn't seem to have (publicly documented) flags to allow that.
I included some of our numbers below.
Best,
Wolfgang
All times in ms, avarage of 10 runs:
startup=load binary+instantiate
main=main function
pp=pretty printing the result as an S-expression, we call an imported
function
for every character, thus somewhat slow, but not that relevant for my
question
For a description of the benchmarks, see: Chapter 8.2,
https://zoep.github.io/thesis_final.pdf
Node.js: v18.19.0:
demo1-opt_coalesce-locals : startup: 6, main: 0, pp: 28, sum:
34
demo2-opt_coalesce-locals : startup: 3, main: 0, pp: 8,
sum: 11
list_sum-opt_coalesce-locals : startup: 3, main: 0, pp: 2,
sum: 5
vs_easy-opt_coalesce-locals : startup: 10, main: 33, pp: 1, sum:
44
vs_hard-opt_coalesce-locals : startup: 10, main: 101, pp: 1, sum:
112
binom-opt_coalesce-locals : startup: 18, main: 10, pp: 24, sum:
52
sha_fast-opt_coalesce-locals : startup: 65, main: 70, pp: 7, sum:
142
color-opt_coalesce-locals : startup: 132, main: 44, pp: 2,
sum: 178
Node.js: v20.11.1
demo1-opt_coalesce-locals : startup: 1, main: 3, pp: 24, sum:
28
demo2-opt_coalesce-locals : startup: 2, main: 0, pp: 12, sum:
14
list_sum-opt_coalesce-locals : startup: 2, main: 0, pp: 2, sum:
4
vs_easy-opt_coalesce-locals : startup: 4, main: 38, pp: 4, sum:
46
vs_hard-opt_coalesce-locals : startup: 3, main: 110, pp: 3, sum:
116
binom-opt_coalesce-locals : startup: 3, main: 26, pp: 23, sum:
52
sha_fast-opt_coalesce-locals : startup: 4, main: 228, pp: 10, sum:
242
color-opt_coalesce-locals : startup: 12, main: 332, pp: 2, sum:
346
--
--
v8-dev mailing list
v8-dev@googlegroups.com
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups
"v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to v8-dev+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/v8-dev/e7e5b6c8-0f7b-4787-b66f-592f02873950n%40googlegroups.com.
