Bits Windows Vista SP1 Core Performance

http://www.bit-tech.net/bits/2008/03/26/windows_vista_sp1_core_performance/1

We've seen how Windows Vista SP1 performs in games but what about the core system performance? We test boot times, disk copying, USB, network, photo editing, file compression, audio and video encoding to see what has changed.

 

wot no x64 love?
strangely mixed results though. it's... interesting to say the least.

 

This is just teasing me now... 2 articles in 2 days!
I was pissing about with this all last night and i STILL can't get SP1 installed

 

The only advantage to me is the network transfer speed and less crashing when doing so. Used to do my box in!!

 

I know it’s mentioned that the difference from Vista - XP Transfer speeds is well cataloged elsewhere, but particularly in the Network & HHD file transfer section it would have been nice to see what Vista File Transfer speeds should be achieving.

Other than that, I thought it was an excellent read.

 

I'll use the same files and same system for the SP2 versus SP3 XP results in a few weeks (providing I can find an XP disk again!)

 

Sweet!

 

Are there any comparisons between XP SP2 and Vista SP1?

 

Great read. I appreciate you guys doing this. I've been considering jumping to Vista Home Premium for the past month with this new build I'm working on, so this gives me some more fat to chew on.

 

Nice report - I just wish you had a copy of XP handy to include in your benchmarks

 

About Vista increase boot time, that is because when you upgrade SP!, it clears pre-fetech, super fetch, and clear all previous boot optimization, once you pass that, Vista DOES start faster. However that is under 64-bit environment.

 

SuperFetch / ReadyBoost and all associated services are disabled because with them turned on, benchmarks get faster and faster over time. There is no reliable way to accurately benchmark a system with the prefetchers enabled. We've seen improvements between run #1 and run #3 on the same hardware for example, which gave us enough reason to just disable it to remove the benchmarking headache.

The boot times were measured on a fresh install with all the hotfixes installed and then again once Vista SP1 was installed - that combined with the above means there should be little to no difference in system state.

 

I would like to see some test on Encrypted FileSystem performance.
I'm seeing particularly bad performance when moving a 100MB file from an unencrypted folder to an encrypted one.
Both files are on the same HDD and performance problems started with the SP1 upgrade. Performance seems dropped by a huge amount...
Could you reproduce this on your test environment?

 

@guanellaluigi, I don't know why, but could it be that it simply encrypts your data before placing it on the drive, and that it could potentially be that Windows Vista pre-SP1 did not do a proper job.

@TimS, I am no expert in anyway on the topic, and I am not trying to questioning your expertise nor protect Vista... just trying to learn. I think it would be best to calculate such task without drivers installed, I say this, because when I install my Creative X-Fi drivers, Windows takes 4-5 times more time to startup then without Creative drivers. And it could be that Nvidia drivers could face the same issue however only effecting SP1, or that the drivers has performance issue with SP1, and it should wither wait for new drivers or maybe try on different configurations.

I think all these benchmark should be done with the next released of drivers, or even better products and drivers released after SP1, as I personally think that Nvidia doesn't care about their old products, as long as it "works" it is fine. What do you think?

Related to Superfetch and prefetch, this is what makes Vista better than XP or better than Vista without SP1, you can't just disable that, else you have to disable EVERYTHING but absolutely EVERYTHING except the bare minimum to run the OS. At some point Superfetch and prefetch won't be able to optimize any more, and this is when you calculate, in my opinion.

@boe_d, Vista would benchmark much slower than XP, for the simple fact that XP has a very very close core system than WinNT, and that companies learned since that time how to make proper optimized drivers. Also benchmark could have potential compatibility issue with Vista that could change results, but personally I think it is just picking at minor insignificant details on the end result score. Between 1000 and 1010 in a score... their is no difference in real life.

 

I'm agog! Superfetch, and to a lesser extent Readyboost are key Vista technologies that are an improvement on XP. Regardless of what it does to your benchmarking, surely they must be enabled?? Otherwise you are comparing apples with pears.

 

Apologies if I have missed this somewhere in the articles...

If I have been diligently installing all recommended updates to VISTA over the course of its lifetime (which I have), is there still benefit from installing SP1 (i.e. is SP1 just a collection of all the critical updates combined) ?

 

The issues regarding file copying are well documented. Pre-SP1 caching (which is enabled in XP) was disabled, and other changes were made in an attempt to improve copying integrity and performance. They succeeded in improving in one or two scenarios but the perceived performance across the board was awful.

In SP1 they have taken a step back and re-enabled caching; the natural effect of this is that the perceived copying speed has improved for smaller/medium files, but little perceived improvement in larger files. When I mention 'perceived' performance it is because when Vista SP1 (or XP for that matter) claims to have finished copying a file, it hasn't really. The file has not truly been copied until it's been removed from cache and written to the actual disk.

As for boot times, we were led to believe that Vista would boot faster than XP, but IMHO this wasn't born out. And I haven't noticed any difference post-SP1.

But in general, I've noticed that SP1 is slightly smoother and slicker, and Vista is just as stable as before - more stable than I ever had with XP.

However, as I said in the gaming performance thread, though SP1 does exactly what MS claimed it would do, it wasn't really ambitious enough - the article clearly states that SP1 is not really going to catch and 'floating voters'. I would have like to see much more visible changes mdae.

Fingers crossed for SP2.

 

SP1 includes all the necessary updates issue previously, plus includes additional changes in core functionality. So installing it would be a good idea...

 

I must agree with cjmUK, however I think Vista fine as it is. If you are begging for 3-5 heck even 10 fps more in your games, than go back to the past with Win95, I think your games will sky rocket in fps. The more features you add, the slower things gets... Take a car! for example, I am most positive that your car will perform way better, if you remove just about everything except the 4 wheels, but you change with bicycle wheels. Anyway...

Also, at Bit-tech did you enable "Enable advance performance", 'cause that does a nice increase in performance, also did you set the computer (in power mode) to "high Performance" instead of "Balance"

 

Depends how often you'd like an article from us.

Once every three weeks while we reinstall everything in between each benchmark run to ensure the system state is the same every time? And even then, there is no guarantee that the system state will be the same, because SuperFetch has its own way of doing things.