Tweaking boot time (classpnp.sys?)

Laker · 09 Aug 2010

madtownidiot said:

exactly right... and why I disagree with some who say startup time is no indication of performance.. to the contrary, if you ignore the time it takes for POST and eliminate as many software and hardware variables as possible, you end up with what I consider a reasonably accurate benchmark.

Well said! +1

Digerati · 09 Aug 2010

why I disagree with some who say startup time is no indication of performance..

Okay. So please explain if it takes 3 or 4 minutes to boot, but once booted, runs great, how is that an indication of performance?

I'm not up to speed until I start my 2nd cup of coffee, but after that, I'm running full speed all day long - not trying to be sarcastic, but is that any different?

I load my PDA's hotsync manager, MailwasherPro spam blocker, and all my security software, dual monitor manager, and have a Quicklaunch toolbar fully loaded that starts at boot, CoreTemp, and a local weather gadget - surely they slow boot times - does that mean my performance is bad? Some graphics card BIOS and drivers take longer to install than others, does that mean the longer one will have poorer performance?

Boot times are affected in a large part by loading hardware drivers. Loading files and programs (which requires disk access) always takes longer than opening the same - which requires RAM access (and maybe PF - which is faster than loading from scratch).

Nothing in startup vs everything needed in startup means nothing in terms of performance after the system has booted - assuming there are no errors with the drivers of services that start - and also assuming there is not a bunch of junk loaded that is never needed - like adobe or java update checkers, or the like.

If you can find a link that shows "boot times affects performance", please post it. I would like to read it.

madtownidiot · 09 Aug 2010

digerati.. I don't think you're even trying to understand what I'm saying.. most of the people who post fast boot times do so with a minimum of peripherals attached and a bare OS with minimal startup items. And I never stated that boot times affect performance, but by removing as many variables as possible (including cleaning and defragging the HDD after a few power cycles) and only considering the time it takes to load windows (ignoring the time it takes for BIOS to decide it's time to start loading the OS), you can get a pretty good base indication of how fast a computer really is. I take into consideration what hardware is installed too.. A raid array, a TPM, multiple video cards, etc are of course going to take a bit longer to initialize, but even so, there's no reason a system with an i7 and SSD and 12 GB of RAM for example should take two minutes or more to boot.. there's a word for that... bloat. You are obviously correct in stating that it takes longer to cold start an app than it does to open one that's already running in the background, but as I've said before, increasing the system load in greater proportion than it's capabilities is actually making it slower. No different than with a car... if you put a 426 hemi in a fully loaded dumptruck towing a backhoe, the engine is still just as powerful as when it was in the Cuda, but it's pretty obviously not going accelerate as quickly, and only a moron would compare them in a drag race...and because I don't have 95 other background processes running at the moment, it takes all of 10 seconds for crysis 2 to load. It takes much longer with all startup items checked off.. I don't want java, itunes, my AV or windows itself to decide it's time to start updating or an IM to pop up while I'm in the middle of a game, nor am I going to check my outlook. I only sync my Ipod 1 in about 10 times when I connect it... all the other times I'm just charging it. I have a printer, but I don't use it every day...most software manufacturers set things to run at startup by default so the user won't forget they exist, not because it's necessary make the program work. My firewall and AV still load the core processes without loading the interface, and they work just fine. My monitor works even though I disabled the ATI control panel, and when I connect a 2nd monitor, that works too.. are you beginning to see my point? It's my personal preference, and you have yours..

Digerati · 09 Aug 2010

madtownidiot said:

digerati.. I don't think you're even trying to understand what I'm saying.. most of the people who post fast boot times do so with a minimum of peripherals attached and a bare OS with minimal startup items. And I never stated that boot times affect performance,

With all due respect madtownidiot. I fully understand what you are saying. But understand this, YOU JUST SAID in your previous post

madtownidiot said:

and why I disagree with some who say startup time is no indication of performance..

And that is clearly contrary to your last statement.

madtownidiot · 09 Aug 2010

Digerati said:

madtownidiot said:

digerati.. I don't think you're even trying to understand what I'm saying.. most of the people who post fast boot times do so with a minimum of peripherals attached and a bare OS with minimal startup items. And I never stated that boot times affect performance,

With all due respect madtownidiot. I fully understand what you are saying. But understand this, YOU JUST SAID in your previous post

madtownidiot said:

and why I disagree with some who say startup time is no indication of performance..

And that is clearly contrary to your last statement.

Now you're misquoting me and out of context

[B said:

echo147][/B]Think I just have to accept that many of the dead quick times from that thread must be systems with the minimum of hardware connected and virtually nothing in startup.

to which I replied:

madtownidiot said:

exactly right... and why I disagree with some who say startup time is no indication of performance..

you conveniently left out the rest of the statement

madtownidiot said:

to the contrary, if you ignore the time it takes for POST and eliminate as many software and hardware variables as possible, you end up with what I consider a reasonably accurate benchmark.

where is the contradiction? And more importantly, how am I wrong? If two computers running close to the same set of programs and comparable hardware take a significantly different amount of time to boot the OS, isn't it reasonable to assume one's going to be a little faster than the other in everything else?

Digerati · 09 Aug 2010

Now I am totally confused. I did not "conveniently" leave it out, it did not, and still does not change the meaning - to me anyway. It appears to me you are saying, "I disagree with some who say startup time is no indication of performance.."

Is that not true?

madtownidiot · 09 Aug 2010

We're both right, you just don't see it. and the second part of my statement makes all the difference in the world. Comparing boot times between two computers without eliminating as many software and hardware variables as possible 1st is pretty moronic, and really, to some people, a slow boot time can be a deal breaker no matter what's inside the case.

Laker · 09 Aug 2010

Yes, you mentioned earlier I was confused when the truth seems to be that you are. I think you are being intentionally obtuse so as to not concede the point. The point of: While boot times do not affect performance they can be considered an indicator of system performance.

gregrocker · 09 Aug 2010

Have never had a Win7 install that started up in less than 20-30 seconds, possibly because I only allow AV and gadgets to start with computer, run CCleaner and Auslogics defraggers monthly, google all Event Viewer errors to resolution.

If I had a startup delay I would create a boot log and monitor closely the Diagnostics-Performance log to see what is hanging.

I would try install without the RAID as Win7 is somewhat allergic to RAID.

madtownidiot · 09 Aug 2010

Wonder what would happen if I put an ssd in my old dell laptop.. the bios takes all of 4 seconds when nothing else is attached, and it boots in about 30 seconds with a spinning HDD