Tom;s Hardware Review: Vista vs. XP

[dfumento]

Tom;s Hardware Review: Vista vs. XP

http://www.tomshardware.com/2007/01/29/xp-vs-vista/

[pianowizard]

Summary: Vista is slower as expected, though, surprisingly, it consumes about the same amount of power as XP.

[wehugheog]

WOW!

I was surprised by the article. I mean really really surprised.

I would have thought that Vista would at least run equally fast in ALL aspects than XP. But a decrease in so many areas makes me think twice about using Vista in the very near future.

I will get the free Vista business upgrade for my new T60p, and I will install it, at least once to take a good look at it, but will revert back to XP for the time being.

Once I will NEED to use Vista, I will ugrade (compatibillity with clients).

just my 2 cents...

[tomh009]

Really, the decreases in performance were generally quite small -- under 5% in most cases. When you think of that spread over five-plus years, and how much faster systems have gotten in that time, it's really a pittance.

I took their power measurements with a grain of salt since they were testing on some sort of extreme desktop system that uses 150W of power when on this board we talk about power consumption in the range of 10-15W. In the latter case there may still be a difference, but probably not as radical as some posters have been predicting.

[brentpresley]

That article is PURE FUD.

Keep this in the back of your mind:
XP - 5 years of driver TWEAKING
Vista - 1 day old (1/2 hardware still running on BETA drivers)


Leaps and bound were made on Vista in performance between Beta 2 and the RCs. As better drivers are release (Intel, Nvidia, ATI - ARE YOU LISTENING TO ME?) the performance will incrementally increase as well.

Sure, Aero does tax the graphics card more than XP, but this was known WELL in advance and is expected.

And honestly, this is the performance difference we saw between XP and 2000 on XP release as well. In fact, benchmarkers STILL use windows 2000 for benching b/c there are far fewer tasks running in the background.

[jdhurst]

<snip> (Intel, Nvidia, ATI - ARE YOU LISTENING TO ME?) <snip>

No.

[brentpresley]

<snip> (Intel, Nvidia, ATI - ARE YOU LISTENING TO ME?) <snip>

No.

Rhetorical.

[pianowizard]

That article is PURE FUD.

But it's still informative in the sense that it tells you what to expect TODAY. I agree with you that Vista performance will improve over time, but right now it's not as good as XP. In other words, it's better to migrate to Vista after better drivers become available.

[dfumento]

5% difference in times are not noticable.

[pianowizard]

5% difference in times are not noticable.

But I seem to recall that that's the average. For certain tasks, Vista was found to be much slower than XP. And 5% is just for the high-end desktop computer that they used. For machines with lesser specs (i.e. including all Thinkpads), the difference would be bigger.

[makaveli559m]

Is there really a difference betweent the RTM and retail? Are the numbers different on each build?
_____________________
Thinkpad A20m Pentium 3 M 850mhz 512MB of RAM 80 GIG hd

[brentpresley]

5% difference in times are not noticable.

But I seem to recall that that's the average. For certain tasks, Vista was found to be much slower than XP. And 5% is just for the high-end desktop computer that they used. For machines with lesser specs (i.e. including all Thinkpads), the difference would be bigger.

The differences for other computers would NOT be more than 5%. The ABSOLUTE times would be larger on a slower computer, but the percentages would remain intact.

[brentpresley]

Is there really a difference betweent the RTM and retail? Are the numbers different on each build?

You bet there are. The build numbers are the same, but the activation and verification for retail vs. OEM is VERY different. MS put about 50 million dollars into the anti-copy protection for Vista. Can't say I blame them as XP was pirated so badly there losses to piracy were estimated in the BILLIONS.

[pianowizard]

The differences for other computers would NOT be more than 5%. The ABSOLUTE times would be larger on a slower computer, but the percentages would remain intact.

Consider the absolute minimum specs that can support Vista, i.e. 800MHz PIII with 512MB RAM, 16MB graphics, and a 4200rpm HDD. You are saying even on such a computer, XP is still only 5% faster? I am talking about what happens in reality, not about benchmarks. On such a computer, I would expect WinXP to run smoothly, whereas Vista would be struggling big time.

[brentpresley]

The differences for other computers would NOT be more than 5%. The ABSOLUTE times would be larger on a slower computer, but the percentages would remain intact.

Consider the absolute minimum specs that can support Vista, i.e. 800MHz PIII with 512MB RAM, 16MB graphics, and a 4200rpm HDD. You are saying even on such a computer, XP is still only 5% faster? I am talking about what happens in reality, not about benchmarks. On such a computer, I would expect WinXP to run smoothly, whereas Vista would be struggling big time.

Assuming NO RAM bottleneck, then YES. Remember we are talking about percentages here, not absolute numbers. You forget that on the above computer, Vista will automatically see the minimum hardware and configure itself accordingly. You won't have a shot in hell of running Aero in that HW, and when Aero is off the differences b/w Vista and XP are much smaller.

I've been using Vista for 6 months now, and can tell you from a variety of hardware (nothing as slow as the minimum you listed, but not all top-end stuff either) that there is absolutely NO perceptible difference in day-to-day usage than with XP. And in my opinion, the advance networking features, GREATLY improved kernel security, and more intuitive interface outweigh the small performance hit that is claimed. Heck, I even ran for TWO MONTHS without antivirus protection installed (and didn't get a virus). I would not recommend that at all, but you can't say you could make it 2 months on XP that way without getting hosed.

[pianowizard]

Remember we are talking about percentages here, not absolute numbers. You forget that on the above computer, Vista will automatically see the minimum hardware and configure itself accordingly.

Yes, I know we've been talking about percentages, and didn't forget that Vista would configure itself to adjust to the hardware. I based my conjecture upon personal experience. I used to have an X40 (1.2Ghz P-M, 1.5GB, 16MB graphics, 4200rpm HDD) and I compared WinXP Pro with Vista Beta 2, with both in classic mode. With 1.5GB RAM, the RAM shouldn't be a bottleneck. I felt a huge difference in speed, for things as simple as opening folders. Had the difference been only 5%, I wouldn't have noticed any difference. It felt more like 100%. But of course, the caveat is that it was an early version of Vista.

[wehugheog]

Well, basically, I rest my case.

I will wait for all proper drivers for my t60p and have them iron out any other issues they might currently have.

After spending 5 years on Vista, I would assume, and expect that at least all the proper drivers would have been ready (by Lenovo, Intel, etc. etc.). Or have they been sleeping?

With respect to the speed, Vista IS slower than XP. However we did see this in all previous Windows versions, so that is not really new. (or should I revert back to Win3.11 ? )

Concerning viri, I never have a virus on my laptop, have not had one for... uhm.... a long time. I use the standard firewall, am behind a NAT and hardware firewall (simple Linksys router) and use Symantec corporate. Also I "do" all updates.

So, as stated before, I will keep using XP-pro for some time to come, but eventually will use Vista.

Call me cautious, call me an old fart, but right now(!) I do not want to "fix" something that is not broken.

[dfumento]

In my view, the only compelling reason to not yet go to Vista is if there is a key program that you need to run that is not available yet on Vista. The new system is fun to use and I really cannot tell any speed differences on a modern laptop.

The one caveat is that Dell recommends running Vista with 2 GB RAM and I would suggest this upgrade of running with 1.5 GB or 2 GB.

[tomh009]

The one caveat is that Dell recommends running Vista with 2 GB RAM and I would suggest this upgrade of running with 1.5 GB or 2 GB.

Ahh, but inject a large hunk of salt here, too. Microsoft recommends a minimum of 1 GB, Dell recommends more -- but which one of the two companies has a financial interest in selling more memory to its customers?

[brentpresley]

Remember we are talking about percentages here, not absolute numbers. You forget that on the above computer, Vista will automatically see the minimum hardware and configure itself accordingly.

Yes, I know we've been talking about percentages, and didn't forget that Vista would configure itself to adjust to the hardware. I based my conjecture upon personal experience. I used to have an X40 (1.2Ghz P-M, 1.5GB, 16MB graphics, 4200rpm HDD) and I compared WinXP Pro with Vista Beta 2, with both in classic mode. With 1.5GB RAM, the RAM shouldn't be a bottleneck. I felt a huge difference in speed, for things as simple as opening folders. Had the difference been only 5%, I wouldn't have noticed any difference. It felt more like 100%. But of course, the caveat is that it was an early version of Vista.

There is your problem. Beta 2 was a DOG in terms of performance. If you want to eval Vista, use RC1 or RC2. They are MUCH faster than Beta 2.

Why? Beta 2 still had LOTS of the debug code included in it. This was properly stripped out in the Release Candidates.