Slashdot Mirror
32bit Win7 Vs. Vista Vs. XP
An anonymous reader writes "ZDNet's Adrian Kingsley-Hughes tested the latest Win7 build against XP and Vista and came to a surprising conclusion: Win7 performs better than the other 2 OSs in the vast majority of the 23 tasks tested. Even installation. 'Rather than publish a series of benchmark results for the three operating systems (something which Microsoft frowns upon for beta builds, not to mention the fact that the final numbers only really matter for the release candidate and RTM builds), I've decided to put Windows 7, Vista and XP head-to-head in a series of real-world tests...'" This review shows only a 1-2-3 ranking for each test, so there's no sense of the quantitative level of improvement.
This review shows only a 1-2-3 ranking for each test, so there's no sense of the quantitative level of improvement.
In other words a totally subjective opinion with no numbers/statistics to back it up, also known as Totally And Utterly Useless.
Linux can handle 32-bit applications on 64-bit OSes. Surely MS can do the same?
Of course they can, and do. Vista x64 runs 32 bit apps just fine.
Unfortunately MS doesn't have the source for all the devices out there, and can't just recompile all of those to be 64-bit, and the 3rd party vendors that can do it, would rather not spend the effort -- hell, they kicked and screamed and did a half-assed job of updating their drivers to work with Vista in 32 bit (the main source of most real Vista woe).
Linux can handle 32-bit applications on 64-bit OSes. Surely MS can do the same?
Well obviously they can. What they can't support are:
A) 16-bit applications (rare, but some businesses hang onto them)
B) 32-bit drivers
A should be fixable; Microsoft purchased the guys that make Virtual PC, no doubt they could figure out a way to integrate that transparently into 64-bit Windows if they put their noggins to it.
B is a non-issue for the majority of recent computers. That said, it's still a problem for older harder, and especially for some older peripherals (scanners, printers, etc.). Even for the things which the manufacturers still support, the 64-bit drivers may only be available on the manufacturer's site and not the CD it came with, and the non-tech savvy will have issues and blame it on the OS.
1) Netbooks. The Atom processors in most netbooks are 32-bit only. Also consider any other embedded scenario where 64-bit CPUs are not available, practical, or where 64-bit addressing is not necessary.
2) Upgrades. Windows does not support upgrading from a 32-bit OS to a 64-bit OS (you have to choose the "clean install" option). If you want to sell upgrade discs to the vast majority of current customers, you need to sell 32-bit copies.
... but what about gaming? That's the metric I'm concerned about the most. My framerates were down 70% in Vista (HL2EP1). Given my computer is only barely cranking out 30fps on recent games at lowest settings, it's a big deal to me.
If he tested all 3 OSes on the exact same hardware configuration and compared those results, then the tests results are valid.
My major problem with these test results is that he ranked them 1, 2, and 3. He should have put in the actual amount of time these tests took so we could see how much big of a difference it is. 1, 2, 3 tells me nothing. The difference between 1 and 2 could be 0.01% or 5000%.
I really doubt this day was FAIL-free for 17 and a half hours until some retarded FP troll showed up.
64bit x86 gives you 4 times the general purpose register space and twice as many vector registers, which is a huge benefit for an architecture that's so lacking in register space.
He tested things like moving files around, compression, decompression... This is all good and fine, but it's probably not the thing that most people "feel" when they use a computer. What I would like to know is how snappy or sluggish does the operating system "feel" when using it for every-day tasks? Does everything halt while the hard drive cranks away when you click a menu? Do the GUI animations help use the computer or do they simply slow you down? That's the sort of thing that matters to most users. How often do you really have to move 100 MB or 2.5 GB of files around?
Installing kernel level software that isn't certified for the OS you are using isn't the smartest thing in the world to do.
I've said the same when people complained about crappy Vista beta results, so I will say it again: Judge the system's speed when it is done. No nanosecond earlier.
The reason is simple. First, it's plenty possible that there are still parts missing. Parts that can weigh the system down again. In Vista's case, we saw a pretty good improvement in handling, but this can work the other way 'round too if early results are promising (maybe too promising) and optimizing takes a back seat to other matters.
From what I can see so far, Win7 still has some stability issues. Improving stability often comes at the price of speed. It is entirely possible that MS tried to get a system out for "beta report" tests that is as fast as possible to get these desired effects. Vista's resource hunger and its sluggish handling was one of the core gripes reviewers had, so it was likely the first tests Win7 will be put to will be about speed and handling. Vista had no really crippling stability issues (aside of driver problem which are arguably the hardware supplyer's problem), so this won't be one of the things reviewers will make a big fuss about.
So what did they produce for a beta review? Exactly what we have here. A system that is as fast as it can be, everything else back to the corner there. Yes, it's maybe crashing from time to time, but it's beta, you know, and Vista already was stable, so they'll get that done by release, no worries. Now imagine it was the other way 'round, stable as a rock but sluggish. Yes, it's beta, so the speed issues could be ironed out, but reviewers would have had a field day with it.
Bluntly, I don't give a flying fsck about a beta review of Win7. Wake me when it's ready for release. In other words, when SP1 arrives.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
I'm sure the fact that this build of windows does not have tons of extra bells and whistles installed, thus leaving more system resources for doing benchmarking.
The general feeling around here is that no-one WANTS to believe it is even possible that Windows 7 doesn't suck. Because if that were true, that would sort of devalue everything done to improve Linux the last few years. (because if Windows 7 is fast and stable and lets you play games, that doesn't leave any room for Linux on the desktop)
It could actually be that Microsoft got it right. It may be that the core of Vista is not as terrible as we all think it is. I've seen posts discussing how Vista uses a completely refactored kernel, with more layers of abstraction and cleaning up of many of the quirks of win32.
Then, on top of this decent foundation, they overloaded it with poorly thought out gimmicks in an attempt to compete with Apple. In addition, some of their rewrites introduced new bugs, such as the networking problems where Vista machines are unable to talk to shared file servers.
It's possible that Windows 7 succeeded. If they fixed the bugs, and ripped out some of the bloated, inefficient Vista code then you might have a decent OS after. Microsoft might be a monopoly, but if they sat on their heels for too long, eventually (it might take 10 years) alternatives would overtake them.
The question is, is most of the application level optimization really done? Say I want to deal only with int32s array, I make one on a 32bit processor. Does the 64bit compiler automatically translate this to an instruction fetching two 32-bit words from memory?
Do two 32-bit addition in my program translate to one 64-bit vector addition in my new program?
We could go on and on, but in essence, the compiler has a really tough job to do; I doubt it does that.
This is true(ish), I grant you, however at least there is a clear intent to redress this issue for the client OS within the next few months with the release of snow leopard and that will be the only OS Apple will distribute.
Microsoft will undoubtedly have versions of Windows 7 in 32 and 64-bit and by default distribute the 32-bit version. I completely understand why they might do this, I might even be convinced it is in their interest to do that, but I think it holds back the development of the platform in general.
Finally Leopard itself does support full 64-bitness, for example Apache on OS X Server is running as a proper 64-bit application. I don't mean to dispute your claims that it remains effectively a 32-bit OS, just that it's not as cut and dried as it might appear.
There are some good TechNet articles at Microsoft that would give you specific answers of what happens on Vistax64 with regard to 32bit memory allocation. (The SDK/DDKs will also give you some answers.)
Also check out interviews with NT engineers at channel9.msdn.com.
As for you questions regarding the compiler, yes. If you compile your application for 64bit, optimizations like you describe are handled unless you disable them in the compiler.
However, the things I was talking about in reference to Vistax64 is that running 32bit code on the 64bit OS, gives the OS the ability to make decisions like this on the fly for upper level system RAM (not CPU level optimizations/etc). So on Vista x64, and running your 32bit code, it will execute faster on Vista x64 because the OS is running faster, but also if you are using large chunks of RAM, the 32bit application will get additional boosts by combining 32bit memory chunks into one read/write of 64bit space.
Once you get what you need on what Windows x64 is doing, head over to AMD and read about CPU specific optimizations that happen in the register and cache levels of the CPUs when executing 32bit code.
Even if you stick to 32bit development, your applications get benefits of Vista x64.
---
Side Note for others:
Anyone here that installs Windows for gaming, if you have 2GB of RAM, grab the 64bit version of Vista, you will easily get 15% more performance out of your games over Vista x32 and XP.
And if you play MMOs, your zone and load times in either version of Vista will make you never want to touch XP again as it is often a 10x to 20x difference due to SuperFetch.
Driver signing requirement is NOT for rootkit prevention.
It's here to make pro-DRM studios happy. Plain and simple.
"Win7 performs better than the other 2 OSs" In other words, it only crashes once a month, instead of once a week.
The article wasn't about Windows 2000 vs. 98 and 95.
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
Scheduler in Vista also performs worse than on XP (so MS had to resort to such hacks: http://blogs.technet.com/markrussinovich/archive/2007/08/27/1833290.aspx [technet.com] ).
Saying this with the link you provide pretty much discredits anything you continue to say.
You have no idea what you are talking about...
Here:
http://www.extremetech.com/article2/0,2845,2302499,00.asp
Make sure you read the PCMark, then click Next to go to the Gaming Page. Vista outperforms XP in every test. (The only test it is a couple of points behind is the synthetic 3DMark.)
And this is SP3 - the fastest XP compared to Vista.
So go on again about how horrible the scheduler is in Vista, I am guessing you don't even know what a scheduler does and especially I know you don't know how it works in NT.
If you want to put your hands over your ears and eyes and keep screaming, "Vista is slower", try clicking your heels together too, it is as likely to make it true and take you to Kansas.
The Vista is slower myths need to stop and the idiocy behind them is really getting annoying.
So proprietary code stifles the advancement of hardware....
Meanwhile, open source drivers for linux and bsd have been ported to 64bit hardware years ago.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Now that is bullshit. If Microsoft got their act together and made somethign as fantastic as Longhorn was supposed to be it really doesn't have anything to do with linux, and many of us that use a variety of things would rejoice. Gnome, KDE etc might be furthur motivated to add new features and improve old ones at the desktop level. I use linux because I want a cheap unix - a vastly improved Microsoft operating system isn't going to change that, plus it fills so many niches that are completely unprofitable for Microsoft so they will never go there. So many of the improvements are in things that Microsoft just does not care about and why should they - such as embedded devices, NFS, decreasing boot speed etc etc.
Since Vista even has problems networking with NT4 machines that some people still need for legacy apps I really am not as optimistic as you are. When something from Microsoft is not Microsoft compatable it is a sign that you have a product with problems.
I'll hold my opinion until they implement things like the poorly integrated DRM that caused so many problems in Vista - if they have to have it (and I think they will decide they do), they should implement it in such a way that it doesn't cause a lot of other problems in the system. To be frank, they are the company that has a long list of very stupid mistakes the latest of which is forgetting leap years exist. I doubt that the 32 bit version of MS Windows 7 will even support Intels Pentium Pro from 1995 and be able to address more than 4GB.
Yes, I read that article.
Microsoft developers are actually TRYING. Remember, there are a LOT of these people : I know quantity isn't everything, but microsoft has thousands and thousands of these men and women. Also, while standards have gradually slipped as the company grew, they have always tried to hire degreed computer science graduates with the top grades. Meaning, on average, microsoft developers are at least competent.
Sure, brilliance and freedom and various efficiencies (as well as IBM having a huge team of their own) have let Linux sort of keep up, but the evil empire of Microsoft does have a lot of firepower.
Windows is working with them to develop security software. So, what actually happened is your friend installed a Beta test of an OS, and then installed a technical preview beta of antivirus software and on his hardware something goes wrong.
As long as he reported the issues to MS, then it all seems like standard operating procedure for using test software running on a test OS.
1 (short ton / firkin) = 89.1432354 slugs / keg
No need to if your Windows DVD happen to have all drivers either I guess, or atleast not many times.
Do you still have to rebuild/reinstall modules for Linux for each version of the kernel? That's always awesome ..
Atleast you don't have to reinstall every driver in Windows each time you've ran Windows update ...
And before you moderate me flamebait, be sure to understand that it's NOT needed for all other oses.
The problem is you just don't understand what the audio system in Vista does. It is by far the most advanced personal computer audio system available on any platform. Which is the reason that it needs a more consistent stream of data. Because adjusting the timing to the computer's various speakers so that the audio arrives at your head at the same time rather than leaves the speakers at the same time isn't free.
Good point. However, I didn't see anything in TFA that said all operating systems were on one drive. Though it also did not say the contrary.
This is NOT a benchmark or any kind. It is a paid-for-MS-Win-7 advert.
Seriously, no real performance tester would grade results as 1,2,3. WTF?
As others pointed out the diff of performance between Vista and Win7 could have been %0.001. or something negligible.
The only thing I can think of is that this was given the olde "wink-wink" from MS to this "reporter".
Great reporting work there Lou!
full 64bit drivers that can shove data to devices oh like Video cards much faster
How do 64bit drivers speed up DMA?
This includes not only the OS's operation, but even 32bit applications running on the OS.
My understanding is that 32bit application would run slightly slower if the CPU was in 64bit mode. Presumably 15% would be the overall system performance, including legacy 32-bit applications?
You see when you have a 64bit memory addressing and can optimize for this in the memory manager you no longer have FS and pagefile lookkup tables for extended amounts of RAM.
What is a FS (filesystem?) lookup table?
You also can do like Vista x64 does and shove two 32bit memory writes into on 64bit address space, so when it can, you get double the read/write performance out of the memory chip because you are pulling two 32bit chunks in one read cycle.
By "64 bit address space" I presume you mean 64 bit register (you fit 2^32 32bit address spaces in a 64bit address space). But even in 32bit mode Intel CPUs can access 128bit registers via SSE. Anyway, this presumably has more to do with your compiler than your OS, so I don't know what Vista has to do with this.
Everything else being equal, 64bit software would run slower than 32 bit, because you need twice as many bits to represent a pointer. Essentially, unless you need an address space larger than 4GB, you are wasting 32bits on every pointer. This would waste memory, cache and memory bandwidth etc. The standard answer as to why 64bit software runs faster on Intel/AMD CPUs is that on these CPUs everything else is *not* equal.
The biggest bonus to running in 64bit mode on Intel/AMD chips is that since 64bit is essentially a whole new arch, we can throw out all the backward compatibility. In 64bit mode we actually have a decent number of registers. Also since 64bit code won't run on old processors anyway, there is no point in compiling code to be backward compatible with the old i586.
Understand yet?
Not really. Not any better than I understand this paper anyway :P. Could you give links explaining your claims above?
If he tested all 3 OSes on the exact same hardware configuration and compared those results, then the tests results are valid.
Kindly disagree. There are billion different ways to make or tune benchmarks so that your favourite (OS, language, whatever) will look better.
The results shown are 100% meaningless.
This guy -- an employee of ZDNet -- is running a pirate copy of Windows 7 (only 32-bit, the 64-bit version is not on the pirate sites). Why would a "EULA" matter.
It's disgusting that ZDNet condones this.
I don't often defend Microsoft, but in this case I think they are right. Running an unsigned driver is something that can totally compromise your system and only needs to be done by a small number of people. Making it difficult is exactly the right thing to do. UAC makes it far too easy to click 'allow' without reading the dialog for this kind of thing.
I am TheRaven on Soylent News
I've got one thing to say:
for the love of god use <br>
Granted. But to the average user, logging out and back in (with or without restarting the X server) is the same as rebooting. As far as Aunt Minnie is concerned, it takes at most two reboots to get her Linux box installed and running.
Good, inexpensive web hosting
So what you're saying then, if I got this right, is that the best audio system on the planet is the one that you have to write yourself? Awesome.
Didn't say it was hard. Said no other OS is doing it. Your argument seems to be that it's possible to do something, therefore it's already been done. To go for the car analogy, you are saying that since you can melt a Honda Civic down and recycle the materials into some of the parts needed to custom build a super car, a Honda Civic is better than a Bentley Continental.
Show me another OS that, out of the box, has the same feature set that Vista has. Any linux distro. Any kernel. And I'll concede that Vista doesn't have the most advanced audio system. Until them blather on, but you're still wrong.
I'm not sure what your point is.
So proprietary code stifles the advancement of hardware....
No, proprietary code stifles the support of legacy hardware.
New hardware, conversely, tends to be supported by proprietary software first, and the FOSS comes later.
And niche hardware tends to be exclusively supported by proprietary software, and FOSS never gets around to adding support. (Think medical instruments, etc...)
Meanwhile, open source drivers for linux and bsd have been ported to 64bit hardware years ago.
Yup, exactly. Your old 10 megabit network card will have 64-bit latest kernel support on OSS while the windows drivers haven't been updated since windows 98. But the latest nvidia graphics card? Proprietary drivers are way ahead of the OSS stuff, and it will be years before the OSS stuff catches up.