Or store the metadata immediately before the code, as GHC does for thunk entry
code.
keegan
- Original Message -
From: "Niko Matsakis"
To: "Patrick Walton"
Cc: dev-servo@lists.mozilla.org
Sent: Monday, March 31, 2014 3:10:39 AM
Subject: Re: [dev-servo] Crazy idea:
On 31/03/2014 16:11, Patrick Walton wrote:
On 3/31/14 3:10 AM, Niko Matsakis wrote:
Interesting thought. You could also use an embedding involving either
coded no-ops at the start of each instruction and/or an unconditional
jump over some meta-data. This would reduce coupling at the cost of
some
On Mon, Mar 31, 2014 at 08:11:32AM -0700, Patrick Walton wrote:
> I also realized yesterday that one could just re-parse the style
> sheet into the AST if the CSSOM or debugging tools are used. I wonder
> whether that would be just as fast as disassembling.
Good point. I imagine perf of parsing is
On 3/31/14 3:10 AM, Niko Matsakis wrote:
Interesting thought. You could also use an embedding involving either
coded no-ops at the start of each instruction and/or an unconditional
jump over some meta-data. This would reduce coupling at the cost of
somewhat more memory. Naturally you'd want to me
Still nice. Perhaps it would make a nice project for research students.
Cheers,
David
On Mon Mar 31 04:37:55 2014, Patrick Walton wrote:
> Of course, I should try on a snapshot of a real Web page (the HTML5
> spec, perhaps), but I don't expect to do much better. 27% is not bad,
> but there are o
Interesting thought. You could also use an embedding involving either
coded no-ops at the start of each instruction and/or an unconditional
jump over some meta-data. This would reduce coupling at the cost of
somewhat more memory. Naturally you'd want to measure the effect on
performance too, though
On 3/29/14 7:22 PM, Patrick Walton wrote:
On a related note, I have been tossing around ideas today for using
SIMD to match multiple selectors that have the same "shape" in
parallel. For example, if we have ".foo #a" and ".bar #a" it may be
possible to use the packed comparison instructions in SS
Yes, it does have pretty tight coupling. That is the biggest risk.
On a related note, I have been tossing around ideas today for using SIMD to
match multiple selectors that have the same "shape" in parallel. For example,
if we have ".foo #a" and ".bar #a" it may be possible to use the packed
co
On 3/29/14 8:23 PM, Patrick Walton wrote:
This is just 29 bytes of code when assembled. This is likely larger than
the equivalent `nsRuleNode`
The right comparison is not nsRuleNode but nsCSSSelector, right?
These are actually pretty bloated in Gecko right now. For example,
".foo #a" is pars
Hi everyone,
I've been discussing this idea with a few people in person over the past
week, and nobody told me it was completely insane. ;) So I thought I'd
send this idea around.
By now many people have heard of WebKit's CSS JIT. It's a surprisingly
small amount of code. One of the issues t
10 matches
Mail list logo
