Robilar

Pardon me?

As I mentioned in the post you quoted, I have two other systems with Vista. After installation, (or reinstallation) I then spend quite some time disabling features in Vista that I have no use for.

I'm not overly thrilled by being called idiotic.

With the exception of the dubious benefits of DX10, for my use there is nothing that Vista provides that is better than XP64.

Last time I checked MS is kicking Vista to the door pending the release of Windows 7 (which I haven't had an opportunity to try yet). XP lasted how many years compared to Vista?

I consider an OS at its best when its in the background and unobtrusive. I don't spend oodles of time tweaking my OS, I don't use widgets, and I make an effort to disable every service that I don't need. With Vista, this takes considerable time and effort.

If that makes me an idiot in your opinion then so be it.

I didn't realize you worked for Microsoft...

Bleached

Robilar speaks truth.

Many businesses and computer labs don't use vista because there is just too much for the system to handle. A prime example would be Aero. Since it is not offered on the lower tiered Home bundle vs Ultimate then that would be a feature considered bloatware. Home has the least amount of bloatware, but business will still shy away from it. One of my coworkers love Vista, but would never ever put it on a lab workstation.

Ravin

Easy there tiger...
I have to agree with you for the most part, Vista provides little improvements over XP- DX10 is somewhat disappointing, but I do think that UAC is a huge benefit- and probably the only one.

Truthfully, (and MS does not want you to know this) Win7 is little more than Vista that has been pre-configured to have many of those "bloatware" services disabled. There are some really nice features in Win7 that could be implemented into Vista VIA SP3 (ie virtual XP), but really- Win7 is a re-skinned optimized Vista, in the same manner that NT, 2K, and XP all sported nearly the same kernel.

That's it for me in this thread...truly sorry for the jack. If someone wants to further continue this discussion start a new thread and PM me.

__________________
Maximus Overhaul *Updated 03/05/08*

CPUz Valid 400x9CPUz Valid 450x8
Ask me about my private DII Battlenet server.

NrGx

Just because you don't use them does not make them bloatware. Bloatware is what you get with your new Acer laptop. These are called features.

I said the presumption was idiotic. No offense to you at all.

Does not mean a thing. Just because the hardware requirements were a little steep and the release was sketchy, the public perception of Vista has been marred and mainly by people that have never used it. It's a brilliant OS and I see no reason why it isn't good for another year or two. This is simply a case of the stupid crowd enforcing a redundant perception; Vista has changed. Do you know just how close Windows 7 is to Windows Vista? They are almost identical.

I hardly think spending ten minutes to turn off sidebar and UAC are that much of a hassle. Perhaps you'd prefer a less functional OS where everything is handed to you like OSX?

Never said that and I wish I did.

__________________
Q9550 @ 4.0GHz

RussianGrimmReaper

http://www.tweaktown.com/articles/14...nce/index.html

There you go guys!

Now let me say one thing, YES, THEY ARE TESTING CF, I KNOW; but it would still stand that if a cards in CF are bottle necked by PCI-E2.0 x8 (same bandwidth as PCI-E1.1 x16) then it's plenty safe to assume that they will be bottle necked individually

AND YES, I did notice that tgains are minimal at best, but hey, the fact that there are gains proves my point.

It's an old article. It was OCN awhile ago (it's dated June 08 ) There was a long thread about it in the news section. I wish I could dig it up for you, but the search feature is useless here

You're free to buy a P45, an X38/48, and two HD4850's for your own testing if you'd like.

Rly rly!



Let me add one last thing. The only time you will see serious FPS differences between PCI-E2.0 x8 and x16 is at higher resolutions. At 1440x900, this scenario doesn't apply. At 2560x1600, it can make a big difference.

No they are not. The PLX chip offers lower latency than two HD4870s. Why? It's because everything is closer together. Electronics are getting to the point where if you want an extra boost in performance, you just have to squeeze stuff together tighter.

If HD4850's can saturate PCI-E1.1 x16, then I think some GT200 cores can too

EDIT: Was reading other posts and noticed

End user review if you don't trust single sponsored reviews like TT.

__________________

003

No, this does not mean they are being bottlenecked. A minimal boost in framerate does not mean that it was previously being bottlenecked, in the sense that the 4GB/s bi-directional link is not being saturated. The small framerate boost operates on the same principle of the small gains you get when going from a 4GHz i7 to 4.2GHz and DDR3 1600 to DDR3 2000.

If you actually monitor the bandwidth usage of the PCI-e bus (I believe this is possible with a RivaTuner plugin), you would see the bus comes nowhere near to being saturated.

A properly configured 4870 1GB CF setup is indeed equal to a 4870X2 and this has been demonstrated a number of times. Latency on both configurations is already so small that it makes no difference in terms of performance.

But technically, the 4870X2 actually has more latency than a 4870 1GB CF configuration. Because on the X2, both GPUs must communicate via the onboard PCI-e x16 2.0 chip, and then must again communicate with the system through the PCI-e bus on the motherboard.

In the CF setup, the GPUs communicate directly through the CF bridge, which is significantly lower latency than the PLX chip on the X2. The difference in physical separation does not come into play.

__________________
I called it!

OpTicaL

Yup, Vista is so great businesses are sticking to XP! Heck, I dual-boot and play my games on XP. The world doesn't revolve around enthusiast websites.

Umm, you just proved yourself wrong. Congrats!

Futile attempt to find a article remotely saying the bus is bottlenecked. If your so confident in your proof, why don't you quote us the section of the article instead of the entire review link? Because you have no solid proof!

__________________

RussianGrimmReaper

(New post because the other one is on a slightly different subject)

How does the 8800GTS 640/320 apply in this situation? It uses a different core than the 9800GX2. The G80 is inferior to the G92 in just about every way.



Whether the GTX 295 is faster than two GTX260-192s, I can't say. I would assume yes with the newest drivers.

You have to remember though, the GTX 295 has only 48 more SPs than a pair for GTX260-216s. Also remember, NVidia has ALWAYS been terrible with sandwiched cards. It's a fact, they're just bad at writing the drivers.

A pair of 260-216s can also be clocked higher because of the lower heat.

Technical specs can only make a card so powerful, it's the drivers that unlease the true beast. The GTX 295 is being choked by it's collar, the GTX 260-216 isn't.
Because there is less heat in the casing of the card. Which is why they can be clocked higher, which is why they can be faster.
And how would you explain the correlating results from other end users on OCN?

Using different boards can only impact performance so much.

I'm sure Rob will gladly re-test using identical boards if you're that insisting.

__________________

RussianGrimmReaper

You've confused me now

Are you saying that there is a speed difference between PCI-E2.0 x8 and x16? I always understood it as a difference in bandwidth.

If it isn't bandwidth that is increased by double the amount of lanes of the PCI-E2.0 bus, then yes, you would be correct. An increase in speed would cause a pair of HD4850s on a P45 to be consistently faster than on an X38/48.

What ever happened to reading a linked article?

The PLX chip offers 5GiB/s bi-directional communication between the two RV770 cores at the same time as offering 5GiB/s bi-directional communication with the PCI-E bus. The HD4870x2 also has an onboard CFX bridge that is separate from the PLX chip.

The CrossFireX bridge offers only 900MiB/s and is uni-directional.

Attached Thumbnails

 

__________________

gerikoh

ehhh i don't think it's a bandwidth issue at all. its just that x48 is a better chipset than P45. the best way to test is a 4850 with the bus port covered with a tape to get pure 8x bandwidth, on the same motherboard.

__________________


Ultimate Quad Core Clock for Clock Benchmarks:

**The Prolimatech Club**

Thesis/HTPC Notebook: Puma's Cartwheel