Vista32- vs. Vista64-bit OS Showdown *Done!*

[nullpointerus]

(Wow! Lots of great comments here! )

Originally posted by: Excelsior

Again, I'm not saying that you guys should install SP1 RC. I myself won't install it on my main rig until the final SP1 comes out. But the performance improvement looks to be undeniable.

Well I've been running SPC RC on my main rig for a few days now with zero issues. I think its an improvement for sure. Can't wait till final comes out.

I switched to SP1 RC Refresh on my main rig after doing a clean installation of Vista x64.

Here's my experience:

--------------------

Being able to choose which partitions to defragment IS a nice feature. I have a 320 GB drive mainly reserved for TV recordings. Vista automatically defragmenting these huge files was a huge pain when I just wanted to optimize file access on my primary partition for some newly installation application. There's no way RTM should have been released without being able to properly configure defragmentation.

Lots of fixes are included. The last count for total SP1 fixes was like 400, many of which are not available for RTM through Windows Update. SP1 RC Refresh took a hefty 123 MB download after all other updates although the SP1 seemed to install fairly quickly (30 minutes?) at stock clocks.

I haven't had *any* problems so far...at least not as many as in RTM.

The UI in SP1 is even faster (i.e. less HDD churning, better responsiveness) than in RTM.

Also, the file copy dialogs are supposed to make...sense.

--------------------

I'll check the 3DMark 06 score difference personally to see what's up with that.

The flipside to benchmarking the SP1 RC Refresh is that, while it's newer than the RTM, it's not final code. The publicly released SP1 may be faster or slower, or have issues/fixes that the SP1 RC Refresh did not have. And it would be a lot of extra work to do four OS's (x86 RTM, x86 SP1 RC Refresh, x64 RTM, x64 SP1 RC Refresh).

I say that if apoppin and I do any serious SP1 RC testing, we should skip the RTM testing altogether. But we haven't really discussed SP1 at all yet, and it is apoppin's "project," so I'm not making promises or dictating anything. Just stating my opinion.

There's nothing to stop others from contributing SP1 results!



Right now I need to verify my overclock. I have a copy of Vista x64 installed on a separate partition (< 35 GB) just for overclocking (so that any errors/BSODs won't affect my primary partition). When it's confirmed stable and everything is installed properly, I'll be able to do some serious benchmarking.

 

[apoppin]

Originally posted by: Aberforth

Originally posted by: BlizzardOne
I wouldn't call 10.9% 'much faster', maybe 'almost noticeably faster'... and those results has nothing to do with gaming, nor 4GB of RAM.

How generic of you...

1. Software has to be optimized to take full advantage of 64bit, any slowdowns are due to 3rd party software not 64bit windows code itself.

2. 32bit OS cannot allocate more than 2 gb per process, there is a feature called video paging in WDD model in vista that does a good job.

3. There is no point in benchmarking 32bit games in Vista 64 mainly because it runs in emulated mode.

4. 64bit cpus have more registers that will allow more information to be processed in a single cycle, so you might see slowdowns if you are running a single threaded application with old EM64T based processors (P4, PD etc).

5. And nvidia drivers are written by janitors who like playing civ4 - incase you dunno.

6. Have fun with this stupid benchmark.

:beer: time to come out of the matrix :beer:

miss you
:roll:
... btw, smith was looking for you ... i sent him to your last known exact location ...
:Q



... most of us are *gamers* and many of US would like to know if there is any practical reason for a GAMER to switch to Vista64. We ALREADY know about the advantages with 64-bit *optimized* applications - we just don't care about them or your silly opinion about our tests.

================

I say that if apoppin and I do any serious SP1 RC testing, we should skip the RTM testing altogether. But we haven't really discussed SP1 at all yet, and it is apoppin's "project," so I'm not making promises or dictating anything. Just stating my opinion.

*My* Project?
--you mean i get the blame

its O-P-E-N ... to anyone interested in serious testing ... for me it's personal
--i don't "do" Beta

 

[Avalon]

I've got an E2140 @ 2.9 and an HD3850 on Vista x64 Ultimate. Can my rig be of service?

 

[apoppin]

Originally posted by: Avalon
I've got an E2140 @ 2.9 and an HD3850 on Vista x64 Ultimate. Can my rig be of service?

i just logged back into say ....

you bet !!!!


:thumbsup:

 

[Cheex]

Originally posted by: apoppin

Originally posted by: Avalon
I've got an E2140 @ 2.9 and an HD3850 on Vista x64 Ultimate. Can my rig be of service?

i just logged back into say ....

you bet !!!!


:thumbsup:

I didn't get an answer about mine...??

Originally posted by: Cheex
I suppose I can help also (if you need me to)

4GB
1280x1024
Core 2 Duo @ 3.0Ghz
8800GTS 320MB
Vista Ultimate 64-Bit

 

[apoppin]

Originally posted by: Cheex

Originally posted by: apoppin

Originally posted by: Avalon
I've got an E2140 @ 2.9 and an HD3850 on Vista x64 Ultimate. Can my rig be of service?

i just logged back into say ....

you bet !!!!


:thumbsup:

I didn't get an answer about mine...??

Originally posted by: Cheex
I suppose I can help also (if you need me to)

4GB
1280x1024
Core 2 Duo @ 3.0Ghz
8800GTS 320MB
Vista Ultimate 64-Bit

I suppose I can help also (if you need me to)

i wasn't sure you wanted to
--so i replied very generally

and thanks for the kind offers.

Yes... Cheex ... you are most welcome ... we'd love to have you join us

... and *anyone else*, you don't have to ask ... JUST DO IT!

Step 1) Have 2 clean formatted partitions and two copies of Vista - one 32 bit and the other 64-bit.

Step 2) Install them identically - with the same programs and games and configure them exactly the same.

Step 3) Run the benchmarks/tests on each OS and give your impressions and results

simple, no ...?
... but a lot of work

 

[Dadofamunky]

Hope this comes together. This could be very useful. I also like what I'm hearing here about SP1. So far, my experiences with Vista 64 have been uniformly positive and I will never look back at XP. Of course, I suspect it behaves best on modern systems like many of ours; however, my new build cost roughly $1000 and the x64 license another $115. And this bloody thing rips. I'd volunteer to help benchmark but I'm wrapped up in transitioning it to be my new main rig and basically don't have the time.

 

[UTFan81]

Originally posted by: Aberforth

Originally posted by: BlizzardOne



3. There is no point in benchmarking 32bit games in Vista 64 mainly because it runs in emulated mode.

How on earth does that make any sense or for that matter almost anything you post?

 

[apoppin]

Hi guys ... just an update ....

i *finally* connected to AT&T's wireless Internet. ... after about 25 hours on a landline to tech support over the last 5 days
-some *bungler* over there stupidly set a switch that did not allow my phone to connect to their network.

i was the one that figured it out with 'error logging' - if you can believe it, not ONE tech asked for an error report. This AM, after i presented the log, the tech said, let me check something ... "opps, we turned off data to your Pantech"

that was it
:Q

the *bad news* is that is isn't much faster than 56K
:roll:
--maybe 3 or 4 times faster as my signal is considered fairly 'weak'
[i am on the Edge network ... when 3G gets here, i will have true broadband ... and of course, i can always connect - anywhere - if i wanted to get a notebook]
But still i will pay $39 a month for 4x faster ... rather than $7 a month for NetZero's [barely] 56k
:thumbsup:

and NO ... i didn't call MS yet
-tomorrow

my back-ups are done and i am blowing away RAID and starting over with 2 new Vista installations tonight [naw] .. or tomorrow

 

[The Boston Dangler]

wow, i'm really looking forward to your results! unfortunately, i'm not in a position to run tests myself, as much as i would love to :brokenheart:. maybe in the future, who knows?

anyway, a couple questions:
will you be including xp sp2 in your tests?
will installing OS's on different partitions produce a relevant difference in HDD speed, or is it not enough to care about?
will you be using an SLI/Crossfire arrangement? plz say no.
will you be using 4GB (3.2GB) in vista 32?
will you be using 2GB in vista 64?

this comes up just days before i intended to kiss vista ultimate 32 goodbye forever. after 6 months, i've finally had enough. but now i'll wait to see if vista 64 is better (performance, reliability, foolishness) or if xp is still where it's at.

good luck!

 

[apoppin]

Originally posted by: The Boston Dangler
wow, i'm really looking forward to your results! unfortunately, i'm not in a position to run tests myself, as much as i would love to :brokenheart:. maybe in the future, who knows?

anyway, a couple questions:
will you be including xp sp2 in your tests?
will installing OS's on different partitions produce a relevant difference in HDD speed, or is it not enough to care about?
will you be using an SLI/Crossfire arrangement? plz say no.
will you be using 4GB (3.2GB) in vista 32?
will you be using 2GB in vista 64?

this comes up just days before i intended to kiss vista ultimate 32 goodbye forever. after 6 months, i've finally had enough. but now i'll wait to see if vista 64 is better (performance, reliability, foolishness) or if xp is still where it's at.

good luck!

first of all, i just noticed that i said i spent "25 hours with Tech support"
- it FELT like 25 hours ... certainly close to 15; well over 4 today.
-- i mention it because the tech people HATE Vista ... there isn't one out of at least 10 who would say anything nice about it. In fact, i had to show most of them where and how to find hidden goodies. i actually like Vista - after having migrated directly from Win2K to it. i just think MS should have released it when it was really ready ... about right now.
:roll:

ANYWAY, i got the equivalent of a 220K connection ... at least 3X faster then what i had ... probably 4X faster. So that does save time; just not as much as i hoped.

The tests are *open* .... anyone can test XP ... but to be truly a valid test of the differences between a 32 and a 64 bit OS, they need to be the same OS [vista 32 vs 64 OR xp32 vs 64].

imo, XP is the past ... there already is a forum across the Pond that is testing XP 32 vs. XP 64

http://forums.overclockers.co....wthread.php?t=17823231

as to installing OSes on different parts of the same HD partition, the performance differences due to that arrangement should be pretty minuscule ... i guess i could return my identical HDs to a non-RAIDed configuration ... but that is a LOT of work as i intend to continue running RAID after the tests are over.

i don't have a 2nd GPU, but i'd *love* to test X-fire ... so no multi-gpus from me for a while.
[i'm too broke to buy one now as i *still* am not back to full-time work after my accident ]


We are going to test 4GB vs 4GB to see if the 'extra' 1/2 GB RAM that 64-bit addresses makes any performance difference in games
-it appears that 2 GB isn't enough RAM for some new games; since RAM is SO cheap, many gamers just get 4GB - "anyway" ... i prefer to stick with just 4GB for all the tests for simplicity sake. Anyone else is free to remove 2 sticks and test on.

 

[Avalon]

I'm trying to acquire a good, spare HDD to install a 32bit copy of Ultimate to. My current HDDs are already partitioned.

 

[apoppin]

There is a celebration in the apoppin family -
well, *me*

:wine: :beer: ?
nope ...
*C-A-F-F-E-I-N-E*
- i finished at 2 AM and was back to the installs at 6AM
:roll:

then i go to work

seriously ... i installed Vista 64 and Vista 32 successfully on separate identical ACHI SATA drives with my IDE drive also functioning
no hitches with either install. Of course, i only have the *basic* installs done .. no bluetooth or aps installed ... just the drivers

will installing OS's on different partitions produce a relevant difference in HDD speed, or is it not enough to care about?

i thought about it more ... so i blew away my RAID-0 [which is overrated anyway] and installed each OS on the identical 2x250GB Seagate 10.2 Barracuda SATA HDs.

my BIOS is set up this way:

SATA RAID/ACHI Mode [ACHI]
SATA Port0-3 Native Mode [Enabled]
Onboard SATA/IDE Device [Enabled]
Onboard SATA/IDE Ctrl Mode [ACHI]

-got it

and i have a reminder to change settings when i change from 32>64-bit

[HPET -> 64-bit] (for Vista64 and
[HPET -> 32-bit] for Vista32)

so tonight and this weekend, i will install my applications and try to configure bluetooth and DUN
- then the real tests of 64-bit compatibility with my HW/SW and then ... hopefully
--the Showdown

 

[apoppin]

1/25 -Friday Night Update: so ... how is everybody doing?

i have some progress to report ... and it is all really a big surprise to me .... Vista 64 installed just as easily as Vista 32.

So far, there is only one unsigned driver - my bluetooth adapter - but i still cannot get a connection to AT&T wireless on any version of Vista. So it will be slow going d/ling MS updates and a few others that have to go over 56K dialup. i can get a good connection to AT&T wireless only on XP - but it is still pretty slow - like a 224k connection


Anyway, there were 135.6 MB of Vista 64 updates for me to D/L and all of them were ultimately successful. i am not having the problems with CCC that Nullpointerus is having [thankfully, for a change]. Just for 'kicks' i installed MS Word 97 to read Word docs - and it works flawlessly ! [!!]
:Q

i installed, patched and played The Witcher - again with everything in-game totally maxed at 16x10 - and it is as smooth as i remember it in Vista32 [2 weeks ago]; i even loaded the saves from somewhere in Chapter One and continued on. However, i *know* i had some issues with my rig before and there were occasional crashes - not so now; i played for about an hour [for purely "testing" purposes, of course] and it was very smooth and excellent on the load/saves times

i am installing *identical* programs, applications and games on each HD and will compare the Witcher Load/Save times and give general impressions of any differences pretty soon [while checking the system tools]. My installs are going to be pretty bare - more like a "gaming rig". So far, there is my NetZero program, non functional bluetooth connection, Avira AntiVir Personal edition and the Witcher ... the next few days will be spent loading up identical games and benchmarking tools

i expect to install

1. 3DMark05
2. 3DMark 06
3. HL2 - Lost Coast
4. Prey
5. Lost Planet
6. FarCry
7. Call of Juarez
8. Painkiller [Overdose IS current]
9. S.T.A.L.K.E.R
10. Crysis demo
11. F.E.A.R.
and possibly
12. Hellgate: London [again - i will make a custom bench - or better, Link me to one]
13. The Witcher [i will make a custom bench - or better, Link me to one]

How is that for a dozen? Perhaps it is too ambitious. i can also run Oblivion or BioShock but likely won't unless you can link my lazy ass to a good prepared time-demo - it is pretty time consuming making your own .. and i may not even get to Hg:L or the Witcher - perhaps just concentrating on Load/save times

LMK if there is anything else you want me to compare - please remember that i *still* have a slow connect, so no 1GB d/ls, please.


so there IS progress ... and i have that great feeling that comes with a smoothly functioning rig [again ]

 

[sutahz]

First post on 1/20 and still not one edit to the 'reserved' fields.. not even 1 or 2 benchmarks/tests to show if there is any difference?

My point being, when is there going to be action instead of just words.

Just saw the post above mine. 5 days into the testing and you've only just installed the 64-bit OS. Ehh, everyone's a critic right? And I am a cynic. You should definatly use the exact same save point if your timing level load times though (but level load times should be the same, same hdd, slight differnt performance due to differnet partition or different hdd. Playing a game is highly subjective, also some levels are more taxing on the graphics system then others.).
http://www.youtube.com/watch?v=BSkDQYe2FYw

 

[apoppin]

Originally posted by: sutahz
First post on 1/20 and still not one edit to the 'reserved' fields.. not even 1 or 2 benchmarks/tests to show if there is any difference??

so like ... you have a point?

 

[nullpointerus]

Originally posted by: sutahz
First post on 1/20 and still not one edit to the 'reserved' fields.. not even 1 or 2 benchmarks/tests to show if there is any difference?

My point being, when is there going to be action instead of just words.

There *has* been plenty of action. It just hasn't been *pretty*.

We entered into this thing with rigs that *appeared* to be working fine, according to all the traditional stability tests. Then we hit some very annoying problems. What good are test results from rigs that aren't at 100%...right?

We have *real life* to deal with, too. But practically the only things I've been doing with my PC the past few days are swapping cables, swapping parts, flashing my BIOS, typing in my help thread, etc. and trying to grab a bite to eat amidst the chaos. I'd like to get my primary partition up and running.

------------------

If I seem a bit prickly, it's probably because I've been doing ~25 clean installations of Vista x86 in the past few days to isolate a problem no one else seems to know anything about.

Here's the one that has me totally stumped:
http://forums.anandtech.com/me...=2146506&enterthread=y

We already solved one wierd problem, yet the error messages remain. WTF?!

I may have to solve that one with $$$ ...
... meaning new Vista DVDs and possibly some HW.

------------------

That's not counting my problems with Vista x64.

I attached an old 17" LCD as a secondary monitor so that I could load performance monitoring apps on it and watch them in real-time to track down stuttering issues and such. But my HD2900XT and CCC go ballistic -- in a seemingly random and unpredictable manner -- whenever I run it with multiple monitors attached.

:|

I've now got a huge support incident with ATI, which is now on hold because the Vista x86 problems have, at various times, called into question the stability of my core components (CPU, MB, RAM, PSU).

Sure, I can forgo this multi-monitor idea and just get some hard numbers in, but without Vista x86, there wouldn't be anything to compare them to, would there? For this reason, I've been working like crazy on Vista x86, trying to get the damned thing to install.

I ran it for several months last year off the very same DVD, which has no visible scuffs, scratches, or any other problems that my eyes can see. And now it won't install *properly* even after swapping HDDs, DVD drives, RAM sticks, SATA cables, SATA controllers (i.e. different ports), etc....?

My rig passes all common stability tests [memtest86+, Prime95 (all three torture tests!), ATITool's artifact tester, etc.) when it's overclocked, not to mention what it does at stock speeds, which is where I've kept it during this troubleshooting fiasco!

:Q


Hey, I don't mind a little healthy skepticism -- this thread would be pointless if no one questioned what they read -- but please don't assume we've just been sitting around lazily and talking endlessly about what a great idea this would be.

------------------

At least apoppin's rig is back up and running smooth...in both 32-bit and 64-bit. :thumbsup:

He lost his RAID, some data, and had a bunch of IDE/SATA errors we eventually tracked down to an incomplete BIOS flash. For some reason, his board requires *two* flashes to update to the latest version. After all this, he's been reinstalling three operating systems plus dl-ing most of the Windows updates on dialup.

I *know* that's frustrating!

Heck, I'm up hours late explaining how/why I've spent the past few days accomplishing nothing.

Eh...power failure. At least my UPS is functioning!

This is SO not my day!

 

[soccerballtux]

Ya installing OS's sucks; and I for one would never want to be a professional benchmarker installing OS's over and over again (or restoring images over and over) lol.

In for the subscription.

 

[n7]

Originally posted by: apoppin
i installed, patched and played The Witcher - again with everything in-game totally maxed at 16x10 - and it is as smooth as i remember it in Vista32 [2 weeks ago]; i even loaded the saves from somewhere in Chapter One and continued on. However, i *know* i had some issues with my rig before and there were occasional crashes - not so now; i played for about an hour [for purely "testing" purposes, of course] and it was very smooth and excellent on the load/saves times

No surprise here.

I've seen a number of people claiming better stability in the Witcher w/ x64 OSes vs. x86.


If i wasn't lazy, i could install one of my x86 Vista Ultimates & bench/test both, but i honestly have no desire to, since i'm not interested in going back to using an x86 system anyway.

I do respect the amount of hell you guys will go thru to bench this...i love the idea of benching, but the actual work involved is annoying to say the least.

 

[Paladin]

Thanks apoppin and nullpointerus. most of us appreciate your efforts. keep us posted.

 

[apoppin]

anyway, everything on my end is up and running .. i have the 3Dmark06 comparisons ...

Vista 32 wins 10442 to 10094 Vista 64
The reason it is lower than my score in sig, is that i am running my CPU at 3.0Ghz and my RAM at stock [as is everything else].

and i have now have Crysis benchmarked in Vista 32 enough to get a good picture ... but i am having a big problem with running it in Vista 64 ... more to follow

Everything is *identical* ... the installed programs, the running background processes, the HW including the HDs, the same games, demos and benching tools

i figure i do NOT want a "pure gaming rig" like Derek Wilson uses for AnandTech as a reviewer - i want to have a all-purpose gaming rig that has an internet connection, antiVirus and a working firewall, Defender is ON, i will have my printer installed and the 'basics' to also do day-to-day tasks. i will not be installing everything from an "image" each time i make a change of graphics driver. i will NOT reboot before benching each game - you don't do you - before you play? The only concession i'll make is to defrag fairly often.

i will update the 2nd post as soon as i make my Crysis findings ... and i am going to call them "preliminary" as i am not 100% certain i have every last driver in my rig absolutely up-to-date [my BIOS is Jan 2, '08] ... as soon as i AM certain, it will go from 'preliminary' to 'final' ... and it gives me a chance to run more tests.

i think i will play The Witcher now on Vista32 ... since i played on Vista64 last night
- for research purposes ... of course

i'm STILL having trouble with the Crysis Demo benchmark on Vista64 :|
[i'll get it if i have to BEAT it into submission :evil:]

but here are the results of the Crysis Demo on Vista32 on my Rig [e4300@3.0Ghz; everything else stock]

- 11.7 FPS average GPU
- 11.09 average CPU

--everything 'very high' and 'blood' @ 16x10

i won't be playing at these settings

detailed results to follow with the Vista 64 benches

 

[apoppin]

Originally posted by: The Boston Dangler
wow, i'm really looking forward to your results! unfortunately, i'm not in a position to run tests myself, as much as i would love to :brokenheart:. maybe in the future, who knows?

anyway, a couple questions:

(1) will installing OS's on different partitions produce a relevant difference in HDD speed, or is it

(2)will you be using an SLI/Crossfire arrangement? plz say no.

good luck!

(1) ok ... i'm using separate identical HDs, ACHI .... not in RAID

(2) Yes ... CrossFire is in the cards

:Q

i just got this:

http://www.newegg.com/Product/...x?Item=N82E16814102717




Price is at $159 ... for a total of $177 shipped [minus a possible] $15 MIR .. there is *nothing* i can get for that price for my rig that will give me a significant performance increase .. i am thinking +33% overall - enough to give some solid AA+ and a shot a Crysis at 10x7 [] ... i doubt i can get $200 for my 2900xt and it is over $450 for a 3870x2 ... --i am thinking mine will be a "poor mans" 3870x2

i guess i can only pair my 2900xt with a 2900p or another XT according to the charts:

http://ati.amd.com/technology/crossfire/charts.html

So ... i have a MB with a 16x and a 4x PCIe slot ... the 4X slot definitely 'holds back' a 2900xt but shouldn't with the 10% slower 2900pro. Also, if i keep it stock, my 850w OCZ should be plenty for a X-fire setup

so i ordered it ... i expect to have it Thursday
-i'll update my sig

OK, both HDs have Lost Coast set up ... and i should also have Crysis benched by ... later tonight [with luck]

btw, i have aero enabled and system restore is on ... as most users would do for Vista

 

[happy medium]

Appopin do not try to crossfire with a 4x slot. It will dramaticly decrease performance.
Theres no 2900 pro in this article but there is a 3850 and 2900xt. I think you should cancel that order.

http://www.tomshardware.co.uk/...fire,review-29948.html

You should google more stuff on this. Theres plenty out there.

 

[lopri]

Even x8 slot is a hindrance to G92 and RV670. Tested on 975X which auto-negotiates PCI-E bandwidth. My lowly HD 3850 loses approx. 5~10% of performance just by going from x16 to x8. BTW I'm loving this card with Vista. Wish AMD had this card with 512MB and single-slot..

 

[The Boston Dangler]

i'm just saying, raid and multiple cards will add a lot of variables.
for video alone:
[(driver issues + OS issues) *2 for multi-card issues] *2 different OS's
or
(driver issues + OS issues) *2

i for one am only interested in single card tests, because i think multi-card setups are just wasteful. if i had a second pc, i'd have some nvidia scores for comparison.