On 01/02/2014 01:06, Daniel Hackney wrote:
I have created a library[1] for speculative parallel computing and
used lexing CSS as my test case. I think Servo could benefit from the
speedups enabled by handling large, sequential tasks in parallel and
wanted to make you all aware of my work. I modified rust-cssparser to
do only tokenization, so the results are not perfectly comparable, but
should be similar.
So far, I only get about a 1.5x speedup when running on a 4-core
machine, but I'm sure there are opportunities for improving things
further. Let me know if this is something the Servo project might be
interested in!
[1] https://github.com/haxney/speculate
P.S. I'm not subscribed to the list, so please CC me on replies.
Hi,
This looks very interesting!
I assume the speedups are compared to single-threaded non-speculative
lexing?
If you have speculative lexing (str -> [Token]) followed by "block and
function grouping" ([Token] -> [ComponentValue]) in order to get
equivalent functionality to rust-cssparser, do you still get speedups
compared to current rust-cssparser? (Which does str -> [ComponentValue])
--
Simon Sapin
_______________________________________________
dev-servo mailing list
dev-servo@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-servo
