Go Back   nV News Forums > Hardware Forums > Benchmarking And Overclocking

Newegg Daily Deals

Reply
 
Thread Tools
Old 09-15-03, 10:49 PM   #157
Sazar
Sayonara !!!
 
Sazar's Avatar
 
Join Date: Jan 2003
Location: Austin, Texas
Posts: 9,297
Default

Quote:
Originally posted by Evan Lieb
These are details that the web sites will pan out themselves. Readers don't need to worry about this agreement if web sites already have everything in place.

If you're wondering whether this will work, I can try and give you a general idea. I almost know for a fact that Kyle at the [H] will be more than willing to cooperate with AT in this agreement if we were to go forward with it today. AT and [H] is a huge part of the puzzle given our respective readership. I haven't met or talked with Dave of B3D, but Anand has, and from what I can gather he'd be willing to come to some sort of an agreement. Scott Damage of Tech Report is also very open minded, I wouldn't be surprised if he were as eager to come to an agreement as Kyle will be. ExtremeTech and THG are probably the biggest question marks.
yes.. the whole point I made about legitimacy is again raised...

you have a whole number of people mentioned here who are of different thoughts it would appear...

what is acceptable to one may not be for the other and who decides which to follow and what information to release to the readership ?

frankly speaking... when it comes to video cards... I don't really care what kyle has to say anymore...

hopefully the inclusion of Dave will add some legitimacy to this 'alliance'
Sazar is offline   Reply With Quote
Old 09-15-03, 11:05 PM   #158
The Baron
Guest
 
Posts: n/a
Default

Honestly, if AT and [H] are The Big Two as you seem to make them out to be, why are probably the most technical people becoming more and more active at Beyond3D? Why are people from ATI, NVIDIA, Beyond3D, Warp2Search, AnandTech, [H], and Bjorn3D, to name a few, active on these boards, rather than [H]'s or AT's?

The big sites have become homogenized--they run Benchmark X, Benchmark Y, and Benchmark Z, tell you those numbers, and that's it. The little sites, where a reviewer does not have to review a card in ten hours, get a feel for a card, know what it does weird, finds out about bugs and problems in the drivers, and basically has the experience that a normal user would have and then reports it to the public.

That's a Good Thing.

And let's see.. if Company X gives money/hardware/special info to Site Y, what company do you think Site Y would push the group of sites to be more lenient towards? And do you honestly think Site Y would tell you that? Get real. The point of multiple hardware sites is to keep each other in check, not to report the same things over and over.
  Reply With Quote
Old 09-16-03, 12:45 AM   #159
StealthHawk
Guest
 
Posts: n/a
Default

Quote:
Originally posted by Evan Lieb
It's tough to say, though off the top of my head the [H], AT, THG, B3D, etc. would have to agree on:

1. What constitutes a cheat and what constitutes an optimization? Is sacrificing a tiny bit of IQ in an First Person Shooter game that only a screenshot would show really a cheat if fps go up by 50% from 30 to 45? What if fps only go up by 10% from 30 to 33 and IQ degradation is noticeable during gameplay in that same FPS game?

2. How to deal with previews, early review samples, etc. in regards to how the product will actually perform once they reach customers with final silicon, final drivers, etc. How to balance exclusives and early review samples in other words.

Among other things. I'll run it by Anand and see what he thinks.
Hi Evan,

I think NVIDIA themselves have done a very good job at defining legitimate optimizations when they revealed their Optimization Guidelines to reviewers and to the public(although they don't seem to follow them to this day, including the Detonator 51.75 )

I am sure that you are well versed with them, but here they are again:
1) An optimization must produce the correct image.
2) An optimization must accelerate more than just a benchmark.
3) An optimization much not contain any pre-computed states.

You will forgive me if this is not verbatim, but it is close. #1 seems very clear, IQ must not be changed or decreased. #2 is a little tricky, I personally believe the intention here is to make sure that any games with benchmark modes do not have just the benchmark mode optimized, but any optimizations must bleed through to the game itself. Other people have interpreted this to mean NVIDIA thinks synthetic benchmarks should not be optimized at all(which I do not think was NVIDIA's intent). #3 seems pretty straight forward as well, things like inserted clip planes should never be acceptable.

In addition to these guidelines, I think this absolutely needs to be added.
4) An optimization must respect a developer's wishes. ie, when Futuremark says NVIDIA's optimizations are not legal in 3dmark03 they are not legal optimizations.


In regards to the points you bring up:
1) I think an effort should be made to perform an apples to apples comparison as much as possible. If you also want to benchmark an the game with tradeoffs to reflect "realworld" gaming then so be it.

2) Review samples should use either the shipping driver or the latest available driver. If an IHV hands you a driver that will not ship with the product or will not be immediately available to users of the product, it should not be used.

Case in point, with the HL2 benchmarks just last week NVIDIA wanted reviewers to test with Detonator 51.75. This driver shouldn't be used because it is not publically available, nor is it shipping with any product. A second look at the benchmark can always be done later once the driver has been properly released.

It should also be noted that when testing a driver that is either buggy or known to contain cheats, a previous clean driver should be selected if applicable e.g. 51.75 either cheats or is very buggy with regards to IQ(what seems to be a global IQ reduction of games that use PS2.0 as well as globally applying the UT2003 pseudo-trilinear filtering/trilinear AF to all D3D games).


Just my two cents.
  Reply With Quote
Old 09-16-03, 01:02 AM   #160
incurable
Registered User
 
Join Date: Feb 2003
Posts: 52
Default

Quote:
Originally posted by The Baron
The point of multiple hardware sites is to keep each other in check, not to report the same things over and over.
Exactly my feelings, especially with the behind-closed-doors negotiations Evan mentioned and the influence of certain indiviuals or sites who simply have no deeper understanding of graphics hardware than 'running Q3 at 400 fps is better than running it at 350 fps' yet still could bully the group because of their sheer size and readership.

cu

incurable
incurable is offline   Reply With Quote
Old 09-16-03, 03:40 AM   #161
Evan Lieb
Registered User
 
Join Date: Sep 2003
Posts: 38
Default

Quote:
Originally posted by digitalwanderer
Yeah, that just instills me with confidence.

"Don't worry your silly little heads with the details readers, just let us come to our own decision...I'll go talk to Kyle about it."



I thought you were serious too.
I donít think you fully understand, though I can see how it came off that way. What I was trying to explain was that any type of communication, discussion, problems or whatever between various web sites would be our concern. That most certainly is NOT the same thing as saying we wonít let readers know the terms of the web siteís agreement. Quite frankly, that notion is nonsensical, readers need to know what the agreement is trying to accomplish so they better understand the purpose of a given video card review. Itís in our best interest to reveal the terms of the agreement.

Quote:
Originally posted by incurable
Exactly my feelings, especially with the behind-closed-doors negotiations Evan mentioned and the influence of certain indiviuals or sites who simply have no deeper understanding of graphics hardware than 'running Q3 at 400 fps is better than running it at 350 fps' yet still could bully the group because of their sheer size and readership.
Read above.

Quote:
Originally posted by Sazar
yes.. the whole point I made about legitimacy is again raised...

you have a whole number of people mentioned here who are of different thoughts it would appear...
Of course, but that doesnít mean theyíre right. Iím not here to argue the legitimacy or technical accuracy of any reviews. Quite frankly, if some other web site or person wants to constructively criticize an AT review, Iíll be all ears, but that doesnít mean Iíll necessarily agree with their assessment.

Quote:
hopefully the inclusion of Dave will add some legitimacy to this 'alliance'
Iím sure it will.

Quote:
Originally posted by The Baron
Honestly, if AT and [H] are The Big Two as you seem to make them out to be, why are probably the most technical people becoming more and more active at Beyond3D? Why are people from ATI, NVIDIA, Beyond3D, Warp2Search, AnandTech, [H], and Bjorn3D, to name a few, active on these boards, rather than [H]'s or AT's?
Not sure what you mean. As just one example, we have several extremely active forum members from Intel who post almost everyday; one works in one of Intelís Oregon fabrication plants as a .13Ķ Lithography Technician, and the other was formally part of the first IA64 team as a leading designer, and without a doubt is currently working on another MPU design he canít talk about at the moment. The majority of their posts are very technical in nature; truly excellent people to have around.

Besides that, the AT forums have active members from tech companies like IBM, Microsoft, AMD, ATI, NVIDIA, DFI, OCZ, and Mushkin, to name a few. We also have members from web sites like the [H] (Brent, not Kyle), 2CPU, B3D (Daveís not very active, but heís posted several dozen times) and Storage Review, to name a few. Iím not an active member of the [H] forums but I lurk there occasionally, so I donít know their situation. But just as I lurk at the [H], Iím sure many other companies and web sites lurk at the AT forums. Quite frankly though, a web site doesnít need to be as technical as is absolutely possible to be valuable to end users.

Quote:
The big sites have become homogenized--they run Benchmark X, Benchmark Y, and Benchmark Z, tell you those numbers, and that's it. The little sites, where a reviewer does not have to review a card in ten hours, get a feel for a card, know what it does weird, finds out about bugs and problems in the drivers, and basically has the experience that a normal user would have and then reports it to the public.
Iíd like to think thatís what we do. There will always be someone to criticize a review, but I believe we do a great job. Besides, there are disadvantages to spending too much time with a product. Remember, we have to balance the frequency of reviews with the details contained in those reviews. For example, in the last few weeks we've released an average of six articles per week. We wouldnít be able to do that if our editors spent three or four weeks working on one card. If anything, itís much more advantageous to readers that we pack in as much information as possible in addition to as many reviews a week as possible. That way, they get information on a wide range of companies and technologies, not just a handful of technical reviews that don't really help readers find what they're looking for.

Quote:
And let's see.. if Company X gives money/hardware/special info to Site Y, what company do you think Site Y would push the group of sites to be more lenient towards? And do you honestly think Site Y would tell you that? Get real. The point of multiple hardware sites is to keep each other in check, not to report the same things over and over.
This is a common misconception among readers. I canít speak for other web sites, but AnandTech editors have absolutely no affiliation with our advertising agency. I have no clue which banner ads are going to be in my next motherboard or video card review. What I literally do for a review is 1) contact the manufacturer via email or phone, 2) introduce myself if I havenít already done so and politely ask if I can review their product, and 3) the majority of the time the manufacturer is kind enough to send that product via overnight or 2-day shipping. That is literally how it works. Sometimes I go out to dinner and talk with manufacturers about their products, but thatís about it. Anand is the only person that has to deal with any sort of advertising, which of course is unavoidable because he owns AnandTech.

Quote:
Originally posted by StealthHawk
[b]Hi Evan,

I think NVIDIA themselves have done a very good job at defining legitimate optimizations when they revealed their Optimization Guidelines to reviewers and to the public(although they don't seem to follow them to this day, including the Detonator 51.75 )
Yeah, itís an unfortunate problem, which is exactly why this agreement may be a good solution.

Quote:
I am sure that you are well versed with them, but here they are again:
1) An optimization must produce the correct image.
2) An optimization must accelerate more than just a benchmark.
3) An optimization much not contain any pre-computed states.

You will forgive me if this is not verbatim, but it is close. #1 seems very clear, IQ must not be changed or decreased. #2 is a little tricky, I personally believe the intention here is to make sure that any games with benchmark modes do not have just the benchmark mode optimized, but any optimizations must bleed through to the game itself. Other people have interpreted this to mean NVIDIA thinks synthetic benchmarks should not be optimized at all(which I do not think was NVIDIA's intent). #3 seems pretty straight forward as well, things like inserted clip planes should never be acceptable.
These rules are actually pretty straight forward. Obviously clipping planes and scenes that are distorted for no other purpose but to accelerate a benchmark and not a game is cheating and is unacceptable. What is of debate are the grey areas I mentioned previously.

Quote:
In addition to these guidelines, I think this absolutely needs to be added.
4) An optimization must respect a developer's wishes. ie, when Futuremark says NVIDIA's optimizations are not legal in 3dmark03 they are not legal optimizations.
I wonít address the Futuremark comment for now, but this is where disagreement ensues. There are people at NVIDIA and ATI that will actually disagree with developers like Valve on how a scene should be rendered faster or with higher quality. Itís not necessarily because NVIDIA or ATI want to cheat either, a lot of time itís simply because they see a way of optimizing their cards performance.

Quote:
In regards to the points you bring up:
1) I think an effort should be made to perform an apples to apples comparison as much as possible. If you also want to benchmark an the game with tradeoffs to reflect "realworld" gaming then so be it.
Agreed.

Quote:
2) Review samples should use either the shipping driver or the latest available driver. If an IHV hands you a driver that will not ship with the product or will not be immediately available to users of the product, it should not be used.
Of course. What some people are currently debating is:

1) Should reviewers use non WHQL drivers?
2) Should reviewers use early stepping silicon when they donít know for sure if itíll perform identically to final silicon?

Quote:
Case in point, with the HL2 benchmarks just last week NVIDIA wanted reviewers to test with Detonator 51.75. This driver shouldn't be used because it is not publically available, nor is it shipping with any product. A second look at the benchmark can always be done later once the driver has been properly released. It should also be noted that when testing a driver that is either buggy or known to contain cheats, a previous clean driver should be selected if applicable e.g. 51.75 either cheats or is very buggy with regards to IQ(what seems to be a global IQ reduction of games that use PS2.0 as well as globally applying the UT2003 pseudo-trilinear filtering/trilinear AF to all D3D games).
Correct, which is one reason we donít use BETA drivers.

Last edited by Evan Lieb; 09-16-03 at 03:52 AM.
Evan Lieb is offline   Reply With Quote
Old 09-16-03, 06:00 AM   #162
Bert
a+d(L.N)+s(N.H)^n
 
Bert's Avatar
 
Join Date: Apr 2003
Location: Blue Circle
Posts: 69
Default

Quote:
Originally posted by StealthHawk

1) An optimization must produce the correct image.
2) An optimization must accelerate more than just a benchmark.
3) An optimization much not contain any pre-computed states.
In addition to these guidelines, I think this absolutely needs to be added.
4) An optimization must respect a developer's wishes.
Well, I'd say #4 is already covered by #1. The problem is to define what the correct image is. I'm not sure if there is, for example, a DirectX specification that exactly states what trilinear filtering is. The situation is a bit better on the OpenGL side, although in some places the OpenGL spec allows for implementation-dependent detail.

I'm pretty sure that nowhere is it stated that texture filtering may only be done on the first tex unit. ATI started this. Now, unfortunately, NVIDIA is doing the same and more. This is why developers are pissed: If I write code that switches on tri-linear, the graphics card has to render it, or it is a bug. Since I do not want to work with buggy cards, I'll switch to a board that does. Too bad there are non left
Bert is offline   Reply With Quote
Old 09-16-03, 06:27 AM   #163
StealthHawk
Guest
 
Posts: n/a
Default

Quote:
Originally posted by Bert
Well, I'd say #4 is already covered by #1. The problem is to define what the correct image is. I'm not sure if there is, for example, a DirectX specification that exactly states what trilinear filtering is. The situation is a bit better on the OpenGL side, although in some places the OpenGL spec allows for implementation-dependent detail.

I'm pretty sure that nowhere is it stated that texture filtering may only be done on the first tex unit. ATI started this. Now, unfortunately, NVIDIA is doing the same and more. This is why developers are pissed: If I write code that switches on tri-linear, the graphics card has to render it, or it is a bug. Since I do not want to work with buggy cards, I'll switch to a board that does. Too bad there are non left
Well, no. #1 and #4 are not the same. For example, look at ATI's optimization in GT4 of 3dmark03 which supposedly was mathematically and functionally the same as the original code. However, Futuremark at that point still deemed code replacement unacceptable. And in the case of synthetic benchmarks, I agree. And from a poll I conducted earlier, the majority of people agree with me.

re: trilinear filtering, I don't know if there is anything in the DX spec that says you have to do trilinear on all texture stages. However, IQ certainly IS degraded in UT2003. And there is no way to get normal IQ, therein lies the problem. Even though NVIDIA added in an "Application" setting in 45.23, selecting "Application" does nothing. You still get pseudo-trilinear and pseudo-trilinear AF. 51.75 still is like this.
  Reply With Quote
Old 09-16-03, 06:41 AM   #164
goblin
 
Join Date: Sep 2003
Posts: 60
Default

Official Nvidia news:

"Due to popular demand by the GFFX users, we have decided that performace is more important than image quality. As are result our image quality from now on will decrease to give the best performance possible. Our cards were, and always will, be the best card to play benchmarks with." -Invidia
goblin is offline   Reply With Quote

Old 09-16-03, 06:55 AM   #165
StealthHawk
Guest
 
Posts: n/a
Default

Quote:
Originally posted by Evan Lieb
These rules are actually pretty straight forward. Obviously clipping planes and scenes that are distorted for no other purpose but to accelerate a benchmark and not a game is cheating and is unacceptable. What is of debate are the grey areas I mentioned previously.
Well, addressing the grey areas, I agree. Based on NVIDIA's own guidelines though, any IQ drop would seem to be an illegitimate optimization. Offering users a tradeoff is not a bad idea. Forcing it on them is.

If NVIDIA really wants to offer users a tradeoff they should do so through the Intellisample slider in the drivers. For example, adding an "Application" setting was supposed to solve the UT2003 trilinear debates. But "Application" does not work, there is still no way to get pre-44.03 UT2003 quality. This situation becomes compounded with comments like this coming from NVIDIA:
Quote:
http://www.3dgpu.com/modules/wfsection/article.php?articleid=70

I spoke with Brian Burke, PR Director for Desktop Products at NVIDIA regarding the UT2003 situation as was told that the drop in trilinear filtering levels does not degrade IQ, and therefore does not violate their new optimization methods.
Clearly it has been demonstrated that IQ is lowered. But NVIDIA says otherwise. Websites definitely need to voice outrage at this blatant deception and hold NVIDIA accountable for statements it makes. Either by bringing the issues to the public individually, or by banding together in the proposed alliance and doing so with a shared voice. A question of interest is whether or not all these "optimizations" undertaken by NVIDIA are really benign. If ATI hardware performed lower would NVIDIA really push their aggressive optimizations? Look at "optimized" UT2003 scores which are coveniently at the same level as ATI. Also look at AquaMark3 with Det50. Again, conveniently scoring the same as ATI.

Quote:
I wonít address the Futuremark comment for now, but this is where disagreement ensues. There are people at NVIDIA and ATI that will actually disagree with developers like Valve on how a scene should be rendered faster or with higher quality. Itís not necessarily because NVIDIA or ATI want to cheat either, a lot of time itís simply because they see a way of optimizing their cards performance.
Ideally IHVs would negotiate with the developers, and the developers would implement lower IQ options straight into the game. IHVs creating an artificial image quality ceiling just does not seem right.

Quote:
Of course. What some people are currently debating is:

1) Should reviewers use non WHQL drivers?
2) Should reviewers use early stepping silicon when they donít know for sure if itíll perform identically to final silicon?
These of course, are much more difficult questions to answer.

1) Since not all IHVs release only WHQL drivers, I would say using only WHQL drivers is impossible. For example, there was no official WHQL driver released for NV30. 43.51 was a beta, unofficial driver and was the first driver to pass WHQL certification for NV3x cards. 44.03 was the first official driver from NVIDIA to pass WHQL, and was launched with NV35. Since users in the past have used official drivers that are not WHQL I don't see a good reason why you shouldn't test non-WHQL drivers. Unless all IHVs make a commitment in the future to only release official drivers that are WHQLed, testing only WHQL drivers might be hard on your part for reasons described above.

Also of interest is the testing methodology required for WHQL. From my understanding DX9 spec adherence is not part of WHQL testing. Which further weakens the necessity of only using WHQL drivers. For example, high dynamic range effects are missing from HL2 on NVIDIA cards because NVIDIA's drivers do not expose float buffers(although the hardware is supposed to support them).

2) This should be rephrased as "should reviewers use early stepping silicon at all." There is no way to know whether or nor final silicon will perform the same or not for sure. You may have assurances from a company that it will, but nothing concrete. I don't see a problem testing early/pre-production silicon. It provides an interesting glimpse of the future. As always, anything that is not final should have a disclaimer indicating such.
  Reply With Quote
Old 09-16-03, 07:40 AM   #166
Skuzzy
Bit Bumper
 
Join Date: Aug 2002
Location: Here
Posts: 782
Default

Whether or not tri-linear or bi-linear filtering is used is up to the video driver/hardware. In DX, the app just asks for some filtering to be done. The purpose of the design was to allow the video card companies to promote what worked best for the hardware and thus removing the responsibility of the application to determine what should be used and when.
Quite frankly, while it certainly can be done, it would make for a significant amount of code to make the determination for each drawprim and slow the app down.
Doing tri-linear on the first texture stage only is not a bad methodology. Normally the remaining texture stages are various types of maps (cubic, environment, bump....). Very rarely are these later texture stages high detail textures, which is where tri-linear helps the most.
I cannot fault NVidia for disabling tri-linear. If the hardware does not have the power to do full tri-linear, then it is a good call to disable it in order to keep the game playable. I would expect ATI to have some thresholds and disable tri-linear if the frame rates drop too low as well. These are perfectly acceptable things to do and devs expect it.

Onward....

Problems:
1) Most every site I have seen appears to be using a frame grab of the front buffer to get an idea of the IQ of the app. Well, there is a problem here. The video card driver is responsible for getting that frame. The frame it CAN grab is not neccessarily the frame buffer, or what is actually being displayed on the screen. NVidia has said that they do some post-processing of the buffer, so the screen shots may not reflect what is really on the screen.
This simply means they are not actually returning the true frame buffer. By defination the frame buffer is the frame being displayed.
Basically what I am getting at, is you have no way of knowing if the actual frame buffer is being returned when a screen shot is taken.

2) The public nature of the benchmarks used by reviewers has allowed video card companies to put in specific hacks/cheats/optimizations (use them interchangeably) to get better results in the benchmark. No amount of posturing by the public/consumer is going to change that. That point should be very clear by know.

3) People are buying into the 'it's ok to optimize,..but/if' argument. Here are some of the problems with that, and it has been very well brought out in this thread. Where is the line? Who determines where the line is? Who checks and verifies the line is not being violated? Who has enough insight to be able to make that call? Does anyone have the right to make that call?
Video cards are big business and while the high end cards are not the lion share of the business, they generate more free press than any other video card product. The board of directors of these companies would be in violation of their fudiciary responsibility, to the investors, if they did not seek out any way possible, to put the company in the best light possible, in terms of what it takes to be recognized as "King of the Hill".

4) Consumers have no idea what should be showing up on the display in an application. Only the developer actually knows what it should look like. This leads reviewers down a path which is to take your "best guess". Just because one card may look better does not mean it is correct.


Well, the above are just covering the issues in a broad sense, but to address a problem, you have to identify it first. Then pick a starting point and get to work.
Solutions could be available to the above issues, but it will take a major mindset change for everyone involved to pursue those solutions. Talking about it is just so much noise. Doing something about it will take time.
I am working on a solution, as I have alluded to in a few posts over the last couple of months. I am not ready to announce what I am doing, but the above is what I am attacking. The problems can be resolved.
__________________
Stuff Happenz!
Skuzzy is offline   Reply With Quote
Old 09-16-03, 07:41 AM   #167
digitalwanderer
 
digitalwanderer's Avatar
 
Join Date: Jul 2002
Location: Highland, IN USA
Posts: 4,944
Default You're right, I was misunderstanding.

Yup, I thought you were actually here trying to raise the standards of reviewing as a whole Evan...my mistake, you're here because you realize your credibility is tanked and now you need to add some to your site before it becomes a joke.

Don't try and dictate, try and listen; it could save you.
__________________
[SIZE=1][I]"It was very important to us that NVIDIA did not know exactly where to aim. As a result they seem to have over-engineered in some aspects creating a power-hungry monster which is going to be very expensive for them to manufacture. We have a beautifully balanced piece of hardware that beats them on pure performance, cost, scalability, future mobile relevance, etc. That's all because they didn't know what to aim at."
-R.Huddy[/I] [/SIZE]
digitalwanderer is offline   Reply With Quote
Old 09-16-03, 08:16 AM   #168
Evan Lieb
Registered User
 
Join Date: Sep 2003
Posts: 38
Default Re: You're right, I was misunderstanding.

Quote:
Originally posted by digitalwanderer Yup, I thought you were actually here trying to raise the standards of reviewing as a whole Evan...my mistake, you're here because you realize your credibility is tanked and now you need to add some to your site before it becomes a joke.
My personal credibility or AT in general? Either way, you are in the minority and are being overly critical and sensational.

Quote:
Don't try and dictate, try and listen; it could save you.
I haven't done otherwise. Reread what I've written for clarification.
Evan Lieb is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 01:26 AM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright ©1998 - 2014, nV News.