64-Bit XP vs 64-Bit Vista

DeathFromBelow

DeathFromBelow

Supreme [H]ardness
Joined
Jul 15, 2005
Messages
7,315
xpvsvistadk4.jpg

Once you go black...

I intended to put this together awhile back, but after being sidetracked by finals, illness, travel, and playing with my new Christmas gear I've just now had the time to sit down and do some benchmarks.

There are a couple of people (one in particular) on this forum who swear by 64-bit Windows XP and think its the best thing since sliced bread. Others claim that Vista is slow/bloated/etc, and otherwise attempt to dissuade others from using Vista.

As a Vista user and fan since the consumer launch I decided I should do some benchmarking to see how they really stand up to each other.

A couple of weeks ago I was playing around with 64-bit XP on my main machine (see my sig), but I've since pulled the drive I used and needed to do a clean install. No problem right? Find my MSDN disks, pop the 64-bit XP disk in the drive, wait for the files to load off the CD, Windows is starting...

xpstabilityinactiongz0.jpg

[smartass]XP's stability in action![/smartass]

Well Jesus Christ on a Cracker, this machine has never thrown a BSOD and it didn't even make it to the formatting phase of setup. After running memtest to ensure that I haven't finally managed to kill my memory with my OC the best I can figure is that 64-Bit XP just doesn't like my new HD4870 for some reason, it worked fine when I was still using an X1900. It's the only thing that's changed since I last tried 64-bit XP.

I wasn't going to swap video cards and mess up my water loop just to benchmark an old OS, so I grabbed a spare rig.

Test Rig Specs:
Athlon 64 X2 4200 (2.2 GHz)
2GB DDR 400
ASUS A8N-VM CSM (nVidia nForce 4 Chipset, integrated geForce 6150 graphics)
Creative Sound Blaster Audigy 2 ZS
250 GB HDD
Since I wanted to run some gaming benchmarks I put in a spare All-In-Wonder X1800XL

I did a clean install of 64-Bit XP, installed drivers (I got the latest nForce/Catalyst/Audigy drivers from their respective websites), installed 64-Bit XP SP2, and installed all availible updates from Windows update. Then I ran the benchmarks. After that I wiped the drive, installed 64-Bit Vista Ultimate, installed drivers, SP1, updates, and did the benchmarks in the same order.

There was no tweaking, vlite'ing, etc. I left Vista's UAC and all services like indexing/superfetch/etc on.

67092494wl2.png


Lets get started then :
(In the screenshots XP is on the left, Vista is on the right)

Install time (I paused the timer while inputing info)
XP: 19:37.9
Vista: 25:36.1

WinRAR Benchmark:
winrarbenchof9.png


Compressing 150 MB of Photos with 64-bit 7zip (I uninstalled WinRAR before installing 7zip):
XP: 0:50.2 seconds
Vista: 0:51.8 seconds

TrueCrypt Benchmark:
truecryptcu9.png

The most important figure is the first one (AES), and its dead even. Re-running the test multiple times shows that all the other benches appear to be a tie as well, with Vista and XP alternating the lead by <1 MB/sec.

Create a 10 GB Encrypted Container in TrueCrypt:
XP: 3:00.9
Vista: 2:58.9

Copy 10 GB Encrypted Container from Desktop to C:\
XP: 5:22.3
Vista: 6:47.8

Transfer a 2GB file from my fileserver to the desktop:
XP: 1:37.3
Vista: 1:31.9

3DMark06:
3dmark06rp7.png


PCMark05
pcmarkrq4.png

I'm not sure why Vista did so much better on PCMark, but running it multiple times gave similar results.

World in Conflict Built-In Benchmark:
worldinconflictki6.png

The game defaulted to slightly higher settings on Vista. I went through and made sure all settings were the same (I even took screenshots of the menus to make sure I didn't forget anything).

Crysis CPU Timedemo (64-Bit):
crysiscpuzs4.png

Average FPS from last 3 runs:
XP: 24.02
Vista: 24.03

Crysis GPU Timedemo (64-Bit):
crysisgpuqy8.png

Average FPS from last 3 runs:
XP: 25.56
Vista: 29.06

Both Crysis timedemos were done with all settings on medium.

Boot Time:
XP: 1:31.3
Vista: 1:11.0

Shutdown Time
XP: 0:29.1
Vista: 0:14.3

______________________________________

In summary, Vista outperformed or was even with XP in every test except 3DMark06, transferring files on the local drive, and (just barely) at compressing files with 64-bit 7zip. I would have liked to do some more benchmarks, especially video encoding, but I really don't have time at the moment.

The only real downside to Vista is that it takes up more space, although the price per GB is so low nowadays that I don't think its a big deal for most people. Back when I was able to actually get 64-bit XP working on my main desktop I had a really hard time getting drivers for my TV tuner card, not to mention it also lacks a media center application. As a long time Vista user it was painful to have to use the playschool looking Luna GUI in XP.

Looking at all the data I can't see how anyone could recommend using 64-bit XP on a new machine. It was a great OS in its day, and I wish I'd started using it back when it was released, but its time has passed. Vista is a fast, secure, and stable OS. Windows 7 still looks to be a ways off, so there's no reason not to make the jump.
 
Zepher said:
ok, I am waiting
Click to expand...

First I accidentally hit post thread, then the forum database took a dump, then my Cable/Internet/Telephone all died. :eek: I thought the world was ending or something.

I finally got it posted though, sorry for the delay. Discuss!
 
I like XP x64 & Vista x64.

I run F@H on all my systems so I have a mix of both on my systems.

Both XP x64 & Vista x64 run all my apps with out problems.

I have not had any issues finding drivers for either O/S.

Granted, there will be some hardware that XP x64 driver support will never exist.

Currently, there are some issues with Vista x64 and F@H so that is one of the reasons I still have a vew XP x64 systems around.

I do like the fact that in XP, file sharing is a lot earier--> Vista makes you set way to many permissions. In a production enviroment / business model, this would be great but not for my home setup.

In the end, I can work and play with both, each having ups and downs.

*Note* - Not sure why you had a issue with your ATI card in XP x64.
The 8.12 XP x64 driver from ATI has worked really well with my 4 x ATI 4850s.
 
EvilAlchemist said:
*Note* - Not sure why you had a issue with your ATI card in XP x64.
The 8.12 XP x64 driver from ATI has worked really well with my 4 x ATI 4850s.
Click to expand...

The bluescreen was during Windows Setup after it copies the files from the CD, hence why I thought it might be my memory crapping out. I couldn't even get to the drive formatting menu. Whatever basic driver it uses during Windows setup doesn't seem to work with my HD4870.
 
Cool, my bud is trading me a 64 bit vista he got for xmas hes not gonna use to trade for my lapped typhoon, Guess whos getting ripped off? :D

Nice article/post. :cool:
 
Conker said:
You gotta do the modification to get 8.12 to work on winxp x64 for some cards. 8.12s have been having problems lately for all os's lately.

http://forums.amd.com/game/messageview.cfm?catid=279&threadid=104796&enterthread=y

If you are using an old version it might affect the benchmarks. Use the latest nforce drivers also. http://www.nvidia.com/object/nforce_winxp64_15.23.html
Click to expand...

As I said before, the BSOD was in the install phase before I even got to formatting the drive.

I used all the latest drivers on the test rig I ended up using.
 
Excellent post, and those benchmarks really raise an eyebrow given all the mindless 'XP is fast, Vista is teh slow' crap that gets spewed daily on the net. Of course I already knew as much, but still. But again I have to commend you, instead of just another mindless fanboy post, you did research and provided evidence, thank you very much.
 
An vista woulda been busy indexing during all of this too.

Yup. I've been through my fair share of OS's and I gotta tell ya I like this Vista stuff.
 
I dunno Conker, Vista is virtually within the margin of error in 90% of the benchmarks it 'loses' to XP-64/XP in those benchmarks you listed. More than good enough considering all the other improvements in Vista, and besides there is a lot more to OSes than pure performance. I'll take a 2% performance hit for all the security in Vista alone, Aero/gpu acceleration, mandatory 64-bit support, stability improvements (due to many drivers being put in user space), etc. is just icing on the cake.
 
devil22 said:
Excellent post, and those benchmarks really raise an eyebrow given all the mindless 'XP is fast, Vista is teh slow' crap that gets spewed daily on the net. Of course I already knew as much, but still. But again I have to commend you, instead of just another mindless fanboy post, you did research and provided evidence, thank you very much.
Click to expand...

agreed, excellent post, and thank you a million times for taking out so much time to do your own testing and post your results like this.... i made my initial post last night when nothing was here but the thread title...

Conker said:
Heres some other benchmarks also of xp64 vs vista64. A lot of the benchmarks favor xp64

http://forums.extremeoverclocking.com/showthread.php?t=309637

http://www.neowin.net/forum/index.php?showtopic=614724

http://forum.notebookreview.com/showthread.php?t=242891
Click to expand...

the first link is absolute crap, as it is just a forum fanboy rant that xp64 is faster (no results shown at all) and the other two have the OSes so close that it is well within the margin of error. Also, specifically in relation to the world in conflict results in one of the links that shows vista 64 getting whipped by xp 64....

you linked to forum posts... tho guys have no better credentials than any of us, but since the OP is a fellow [H]er, i trust his ability to properly install/configure and test his rig with the OSes than those other forum posters.....

devil22 said:
I dunno Conker, Vista is virtually within the margin of error in 90% of the benchmarks it 'loses' to XP-64/XP in those benchmarks you listed. More than good enough considering all the other improvements in Vista, and besides there is a lot more to OSes than pure performance. I'll take a 2% performance hit for all the security in Vista alone, Aero/gpu acceleration, mandatory 64-bit support, stability improvements (due to many drivers being put in user space), etc. is just icing on the cake.
Click to expand...

agreed, only the delusional cling to XP 64.... placebo can be a *powerful* thing....
 
The benchmarks are so close not only in those links but also the data the op posted I think it shows the differences are more likely hardware and drivers.

It certainly disproves those claims that vista is way slower than XP. It's as fast as if not faster than XP. And windows 7 will push the performance even further.
 
vista use to be slower. but since all the updates, release of SP1 and good driver improvments its a very solid OS now.

Problem is every one still spouts fud that was true 2 years ago and has no relavance today.
 
I'm not sure why we fight over these things anymore other than to suit the OS evangelists (on both sides)... Let those who want to use XP (or Vista) do so. The market will sort them out. If people keep using XP while the rest of the world goes Vista (or really, Windows 7), that's their risk to take. There's some truth to "if it works, don't fix it" but with Vista (and 7) Microsoft has taken something that, while not broken, could use improvement, and souped it up. I don't have hardware issues with Vista (which is the only real reason I can see to not upgrade to Vista/7) and I'd never go back to XP.
 
MrGuvernment said:
then your testing is flawed, it is part if the install process.....
Click to expand...

I was trying to ensure that my blundering around didn't affect the benchmark (like if I accidentally mistyped the 25-digit serial number in XP setup).

I don't see how it's flawed since I used the same procedure for both installs. :confused: It just gives you the raw time that Windows setup spent installing.
 
MrGuvernment said:
then your testing is flawed, it is part if the install process.....
Click to expand...

sorry but its the only fair way to do it, its a variable thats directly affected by an outside element and can cause unfair differences.

Like he already stated, if he mis-typed one of the number and had to redo it than the time is going to suffer.

Only way you could do it fairly is to redo the info entering thing like 3 times (3 times without mistakes), average out the time it took and add it to the total time which sounds like a pain in the ass to me.
 
DeathFromBelow said:
First I accidentally hit post thread, then the forum database took a dump, then my Cable/Internet/Telephone all died. :eek: I thought the world was ending or something.

I finally got it posted though, sorry for the delay. Discuss!
Click to expand...

You are not using Cox are you? I have been having internet drop out since Jan 1st.
 
These results are pretty good, and I find them much more consistent with my own experience with Vista and XP than a lot of the tests that throw XP way ahead of Vista.
 
brucedeluxe169 said:
agreed, excellent post, and thank you a million times for taking out so much time to do your own testing and post your results like this.... i made my initial post last night when nothing was here but the thread title...



the first link is absolute crap, as it is just a forum fanboy rant that xp64 is faster (no results shown at all) and the other two have the OSes so close that it is well within the margin of error. Also, specifically in relation to the world in conflict results in one of the links that shows vista 64 getting whipped by xp 64....

you linked to forum posts... tho guys have no better credentials than any of us, but since the OP is a fellow [H]er, i trust his ability to properly install/configure and test his rig with the OSes than those other forum posters.....



agreed, only the delusional cling to XP 64.... placebo can be a *powerful* thing....
Click to expand...

lol your delusional in reality. This is probably the only test i see that goes faster than xp x64. That first link was done on directx9 on xp64 and vista64 and xp64 was faster. The guy had more games to base off his results and had averages and peaks more. I already know the person trying to sway everyones thinking making this thread probably downclocked his cpu or something. Atleast the guy in the other forum posted up his cpuid snapshot while taking some of these tests. As it is not many games were tested either. Some of these tests have to be taken more than once also to know the actual average. Not even sure if he installed the amd dualcore optimization. He should have some screenshots open of his cpu and gpu settings in windows on the side to show he wasn't cheating.

I bet some of the guys here bashing xp64 haven't ever even installed it. I had both and xp64 is just better performance wise.
 
Conker said:
I already know the person trying to sway everyones thinking making this thread probably downclocked his cpu or something.
Click to expand...

Please take your trolling elsewhere. I'm not going to say anything more to you.
 
Those benchmarks are in no way 'more legit' than these, and they are all within the margin of error except the one WIC benchmark, however Vista in turn kills XP in PCMark, not only in the OP but at http://www.extremetech.com/article2/0,2845,2302499,00.asp as well. The performance is so close otherwise that it's a non-issue. XP should die, it was a good OS in it's day, but Vista is way better.
 
With any luck, Windows 7 64-bit will absolutely destroy both Vista & XP x64.


/current Vista x64 user
 
What's the use of creating yet another thread with the exact same subject? Scared the other thread holds too much information IN FAVOR of XP64 are you?

perfressmall.gif


I currently have 3 different operating systems on one Intel SSD, and have had them on this machine for about 5 months now;
- CentOS 5.2 (x64),
- nlited Windows XP x64, and
- vlited Vista Business x64.

See http://www.clunk.org.uk/forums/reviews/3201-gigabyte-x48t-dq6-living-review-84.html#post39078 for my system specs.

The XP64 install by far outperforms both others in almost everything I throw at it, all I'd like to do with this machine can be done best in XP64. The only REAL reason one might favor Vista is because it has some specific software written for it that you fancy (UAC or the likes), or because you like the ridiculous shiny bullshit on your screen nobody really needs in actual production use of daily life.

If you're the only one using your machine, XP64 is the fastest choice by far, no doubt about it. Vista is only interesting for secondary PC-newbie stuff, like multiple less experienced users on one system might require.

The only thing you should watch out for is making the XP64 install secured, which is not hard to do at all for the computer user with a few years of experience. Use some software from this list and you'll be fine. Vista by default closes things down for you (which can be very annoying by the way, especially if you don't like an OS bossing you around).
 
devil22 said:
Those benchmarks are in no way 'more legit' than these, and they are all within the margin of error except the one WIC benchmark, however Vista in turn kills XP in PCMark, not only in the OP but at http://www.extremetech.com/article2/0,2845,2302499,00.asp as well. The performance is so close otherwise that it's a non-issue. XP should die, it was a good OS in it's day, but Vista is way better.
Click to expand...

Windows Xp sp3 32bit is not the same as xp 64bit since windows xp 32bit was based on windows 2000 and windows xp x64 was based on windows server 2003. Completely different guts with a similar gui though.
 
Vista does pretty well there, especially compared to XP 32-bit. What program made those benchmarks? Is that one program or did you make a graph of many programs, because screenshots would be of more utility, anyone can make a graph say anything after all. I know screenshots are not full-proof but they're more credible imo.
 
Conker said:
Windows Xp sp3 32bit is not the same as xp 64bit since windows xp 32bit was based on windows 2000 and windows xp x64 was based on windows server 2003. Completely different guts with a similar gui though.
Click to expand...

And windows 2003 is based on Windows XP which is based on Windows 2000. The guts are not all that different, not as different as Vista is to all of them, with it's different driver model and so on.
 
Meow said:
The XP64 install by far outperforms both others in almost everything I throw at it, all I'd like to do with this machine can be done best in XP64. The only REAL reason one might favor Vista is because it has some specific software written for it that you fancy (UAC or the likes), or because you like the ridiculous shiny bullshit on your screen nobody really needs in actual production use of daily life.
Click to expand...
:rolleyes: Sure.

What benchmark software did you use? I'd like to repeat the same benchmarks on my hardware with an unmodified copy of Vista when I get the chance.
 
Conker said:
lol your delusional in reality. This is probably the only test i see that goes faster than xp x64. That first link was done on directx9 on xp64 and vista64 and xp64 was faster. The guy had more games to base off his results and had averages and peaks more. I already know the person trying to sway everyones thinking making this thread probably downclocked his cpu or something. Atleast the guy in the other forum posted up his cpuid snapshot while taking some of these tests. As it is not many games were tested either. Some of these tests have to be taken more than once also to know the actual average. Not even sure if he installed the amd dualcore optimization. He should have some screenshots open of his cpu and gpu settings in windows on the side to show he wasn't cheating.

I bet some of the guys here bashing xp64 haven't ever even installed it. I had both and xp64 is just better performance wise.
Click to expand...

lol
 
Nice accumulation of data.

Benchmarks aside, an OS is a tool that's useless without support to do the things you want to do. When I got my EM64T system at the end of 2006, I wanted to run a 64-bit OS and I checked out XP x64. Pathetic support for anything. No drivers, BSODs, etc... The OS was great, but no one who made any of my peripherals, addons, plug ins, or software wanted to write anything for it.

I gave up after a couple of weeks and went XPSP2 (and hence I could print, calibrate my monitor, use my tablet, use programmable keys on my keyboard, use my joystick and force-feedback wheel, have working AV software, watch TV, have working context menu entries for editors and decompressors etc...etc...etc... again)

I've gone Vista x64 just in the last 3 months or so. With the rewrite of a lot of drivers and hooks for Vista, I've found that many companies chose to write 32/64 bit drivers and support both flavours. The x64 drivers may still be lower quality and have less testing than the 32-bit, but at least they exist. So, if a company supports "Vista", you have a decent chance of it working under x64 too. Of course there are exceptions (like my TV tuner card).
 
Surly73 said:
I've gone Vista x64 just in the last 3 months or so. With the rewrite of a lot of drivers and hooks for Vista, I've found that many companies chose to write 32/64 bit drivers and support both flavours. The x64 drivers may still be lower quality and have less testing than the 32-bit, but at least they exist. So, if a company supports "Vista", you have a decent chance of it working under x64 too. Of course there are exceptions (like my TV tuner card).
Click to expand...

I believe Microsoft requires hardware manufacturers to produce both 32 and 64-bit Vista drivers if they want to be certified as Vista compatible. I feel your pain on the TV tuner, although getting my Hauppauge cards to work is exponentially easier on Vista than it ever was with XP MCE.
 
devil22 said:
And windows 2003 is based on Windows XP which is based on Windows 2000. The guts are not all that different, not as different as Vista is to all of them, with it's different driver model and so on.
Click to expand...
They are all built on top of each other. Plenty of XP shit left in Vista and plenty of Vista shit left in 7.

2000 - 5.0.3700
XP - 5.1.2600
XP x64 - 5.2.3790
Vista - 6.0.6001
7 - 6.1.7000

Just keeps adding/subtracting from the same code.
 
bigdogchris said:
They are all built on top of each other. Plenty of XP shit left in Vista and plenty of Vista shit left in 7.

2000 - 5.0.3700
XP - 5.1.2600
XP x64 - 5.2.3790
Vista - 6.0.6001
7 - 6.1.7000

Just keeps adding/subtracting from the same code.
Click to expand...

yup, pretty much

still bugs me that they call windows 7, windows 7 instead of windows 6.1 (which is what it is), or windows XP part 2 (which is what it is- spiritually)
 
bigdogchris said:
They are all built on top of each other. Plenty of XP shit left in Vista and plenty of Vista shit left in 7.

2000 - 5.0.3700
XP - 5.1.2600
XP x64 - 5.2.3790
Vista - 6.0.6001
7 - 6.1.7000

Just keeps adding/subtracting from the same code.
Click to expand...

Yea I know, but my point was that Vista has many more differences from xp, w2k3, xp-64, and win2k than those non-vista OSes have from each other. I would be surprised if any of them performed substantially different than any other, of the non-vista OSes. Hell I'd be surprised if Vista performed any differently for the most part. It's all the same OS line (NT) after all.
 
bigdogchris said:
2000 - 5.0.3700
XP - 5.1.2600
XP x64 - 5.2.3790
Vista - 6.0.6001
7 - 6.1.7000
Click to expand...

Do you think Windows 7 will ship with the 6.1 version number? Or are they just using it now for testing? Then will Windows 8 be version 7.0?
 
You must log in or register to reply here.
Back
Top