Discussions

Man, that's just scary.  I hope it's nothing serious, jamie. *fingers crossed*

androidi said:

Duanerrr said:

*snip*

IDK if this is still the case with RTM. I used to run Vista beta on a slower laptop with less memory and it was easy to note differences. What I noted during the *beta* on the slower computer was that turning off things I wouldn't want to turn off like system restore, windows defender, uac, defrag, indexing all together gave notable perf boost. However on 4 GB c2d at 2.4 ghz or more the overhead of above on RTM is not enough to warrant losing any one of those.

Nevertheless XP feels still clearly faster than Vista on that old hardware even with all the useful services turned off (I ran the lappy with RTM for while before going back to XP). This could be explained by Vista being compiled with newer compiler that produce slower code for older computer and faster for Core 2. More likely explanation is that there's still a ton of new lower level stuff in Vista that has very little overhead but with slow system gets more pronounced. I've noticed that 1.6 ghz core 2 is clearly horribly slow compared to 3 ghz core 2. Really need that 2.4 min to make Vista usable though I think to get snappy Vista you need 4 ghz or new i7/Nehalem. And those new Intel SSD of course.

2008 just seems more responsive to me, probably because it comes with a lot less stuff turned on by default. Less is more. Like Raymond Chen says, "Doing nothing is really fast!"

I don't think Microsoft can make an iPhone competitor without severely damaging its relationships with WinMo adopters.

HumanCompiler said:

JChung2006 said:

*snip*

They have so far.  Have you tried IE 8 Beta 2?  Very fast (like FF 3 and Chrome).  Feels about the same to me.

I've tried IE 8 beta 2.  It's comparable to Firefox 3.0 and Safari 3.1 for JavaScript performance.  V8 (Google Chrome) and TraceMonkey (Firefox 3.1) outperform it for JavaScript code, but the network I/O and DOM interactions (layout) on Web pages render those benefits mostly moot anyway.  I'm not sure how SquirrelFish (WebKit 4.0) compares.

The JScript team is making up lots of ground, which is encouraging! For years, they were pretty much the worst for JavaScript performance. It wasn't that it was especially bad (except for the memory leaks), just horribly neglected. They've even patched IE6 JS enough so that Gmail will continue to support IE6 as a platform - New Gmail Code Base Now for IE6 [Gmail Blog, Google]

var is required for anonymous types.  Otherwise, it's a tradeoff between readability and conciseness.

JoshRoss said:

JChung2006 said:

*snip*

I am going through ISA.  Strange.

I have been getting inconsistent results on the homepage from our ISA Server lately, where rarely the results are compressed end-to-end and most times they're not.  No ideas why it's happening but it used to be consistently compressed traffic between the outside world and the ISA server and uncompressed traffic between the ISA server and internal hosts.

Even with GZip compression, with that much script, we're taking a hit on client-side performance with just JavaScript parsing, execution, and DOM interaction. With Mozilla and Google raising the bar on JavaScript performance, I hope that the IE and JScript teams will step up their game for their vNext's.

Chinposing always brings this to mind for me:  ChessCoaster [xkcd.com].

lewellyn said:

JChung2006 said:

*snip*

Hm. OK. At least I'm not living in some parallel universe from everyone else.

Luckily, the majority of my testing right now is to make sure I didn't break IE 6's rendering. So, when I do my run-through on IE 7, I'll reboot the Mac into XP.

But, here's to hoping that someone who can get a new VHD released notices soon.

Me too.  They didn't update the IE 7 VM when they refreshed the IE 6 and 8 VM's. Likely an oversight.