Getting ready for benches. Need a little advice.

[3DVagabond]

I doubt Ares will be $1200.00. I hope not anyhow, but I suppose we'll see.

There are other reasons to buy a high quality video card. It's not all about fun and games, you know. Nvidia could probably sell all they can make just for F@H, as an example. If it's twice as good in OpenGL than a 5870 it would be worth $700.00, easily...

 

[Shmee]

^ that is a good point. But then arent the GTX 4xx supposed to be very hot and power hungry? Kinda would be a pain to have it folding all the time.

I certainly hope the GTX 400 series is nice and cheap, but who knows. Time will tell.

 

[BTRY B 529th FA BN]

Can't wait for these to come out!

 

[Keysplayr]

Anyone have a problem with me using the 10.3a's for the benches? They seem to have a nice improvement over the 10.2's. And I don't want to wait until March 24th when the WHQL drivers are available from amd.com. I have the 10.3a's here from amd underground. Think those will be good enough for the benchmarking?

 

[SlowSpyder]

Anyone have a problem with me using the 10.3a's for the benches? They seem to have a nice improvement over the 10.2's. And I don't want to wait until March 24th when the WHQL drivers are available from amd.com. I have the 10.3a's here from amd underground. Think those will be good enough for the benchmarking?

I'd use whatever is out that fits into your schedule. If there was something specific in 10.4 that you wanted, wait. But if you plan on benching on a certain date, I'd just use whatever is available at that time. If it's 10.3a, then that's what I'd use. Good luck with it, I'm really happy with my 5870 so far.

*edit - And just out of curiosity, what is the purpose of you benching the 5870? Just something you want to do? Are you getting a GTX480 you want to compare it to or something? I'm just curious, with you being part of the Focus Group, I'm just curious. Anyway, good luck with the benches, I hear 10.3a brings about some nice performance improvements (I'm still on 10.2 for the time being).

 

[Will Robinson]

So you need some advice?
Here's some:Imagine the other posters in this forum are actually adults that can reason and think.
That way the idea of an NVidia Focus Group member providing unbiased and believable benchmarks may have more than a snowflake's chances in hell of being accepted.

 

[tviceman]

Keys: use the best available drivers at the time you benchmark.

snip

Leave and never come back.

 

[nemesismk2]

So you need some advice?
Here's some:Imagine the other posters in this forum are actually adults that can reason and think.
That way the idea of an NVidia Focus Group member providing unbiased and believable benchmarks may have more than a snowflake's chances in hell of being accepted.

I agree with you 100% because I think if you are a member of the nvidia focus group then you are an nvidia marketer recruited to promote nvidia products so unbiased goes out the window. I am not having a go at keysplayr because he seems to be one of the more honest nvidia focus group members but I wouldn't trust the results.

 

[Keysplayr]

I agree with you 100% because I think if you are a member of the nvidia focus group then you are an nvidia marketer recruited to promote nvidia products so unbiased goes out the window. I am not having a go at keysplayr because he seems to be one of the more honest nvidia focus group members but I wouldn't trust the results.

The benches you will see from me are 100% legit on the money. Anything other than that would be a disservice to all of my fellow members here and anyone on the web who views them. That is not my style guys. It is what it is. If a 5870 kicks the snot out of a GTX480, so be it. If a GTX480 just barely beats out a 5870, so be it. I don't care. We all deserve to know REAL results. You should also note the last time I did a bench like this was the 2900XT vs 8800GTS 640. And in that bench session, if the 2900XT did something better than the 8800GTS 640, it in there.

That's how my benches roll, and that's how they will always roll.
If you're not convinced, there is nothing I can do about that. Many folks will appreciate the effort though, and those are the ones I do this for, as well as for myself because I truly enjoy it.

Those that don't appreciate it and condemn it even before it exists, sadly it's their loss.

P.S. Thanks, Will, let me know how those snowflakes are doing in hell. Is it warm down there?

P.S.S. Ok, I'm going to use the 10.3a's for the benches. It would feel like a waste of time using 10.2's when the 10.3a's appear to offer much more all around.

/peace

 

[nemesismk2]

My lack of trust comes from when the Nvidia Focus Group was just about Stealth Marketing and their members didn't let anyone know their advice was biased. As I said I am more likely to trust you because I think you are honest.

 

[Keysplayr]

My lack of trust comes from when the Nvidia Focus Group was just about Stealth Marketing and their members didn't let anyone know their advice was biased. As I said I am more likely to trust you because I think you are honest.

Thanks!

 

[1h4x4s3x]

The question is, what's the point of this exercise?
We will have lots of independent reviews by the time NDA falls, so why are you bothering? My guess, you're doing it for yourself, for the joy of benchmarking, right?

 

[nyker96]

Let me get this straight, you are benching 5870 vs the new GTX480 right? If that's true that's great. good luck with the project!

 

[WelshBloke]

The question is, what's the point of this exercise?
We will have lots of independent reviews by the time NDA falls, so why are you bothering? My guess, you're doing it for yourself, for the joy of benchmarking, right?

Its the only way he's allowed to buy an AMD card, he's wanted one since they were released. ()

(sorry Keys, couldnt resist:twisted

Seriously though, Keys is not going to fudge any benchmarks. He's an honest bloke.

 

[Keysplayr]

Its the only way he's allowed to buy an AMD card, he's wanted one since they were released. ()

(sorry Keys, couldnt resist:twisted

Seriously though, Keys is not going to fudge any benchmarks. He's an honest bloke.

The real problem was convincing my wife to let me buy one.

 

[Rhoxed]

Looking forward to the results.

Thanks for taking the time to do this Keys, its always nice to see personal results along with the review sites.

 

[Keysplayr]

The question is, what's the point of this exercise?
We will have lots of independent reviews by the time NDA falls, so why are you bothering? My guess, you're doing it for yourself, for the joy of benchmarking, right?

See post #58 & #59.

 

[1h4x4s3x]

See post #58 & #59.

Yes, but I don't get it. Don't you think people are more interested in 'pro' benchmarks from all the major - and hopefully independent - review sites than from you? (no offence)
Anyway, it's a nice add-on I guess.

 

[Keysplayr]

Yes, but I don't get it. Don't you think people are more interested in 'pro' benchmarks from all the major - and hopefully independent - review sites than from you? (no offence)
Anyway, it's a nice add-on I guess.

Why do I do it? Oh, jeez. That's easy. It's to give you guys another opportunity to tear it down. Call it lies. You know, the standard stuff. Status Quo.

Just take some notes from Will Robinson above. He's got it goin on!

 

[Kenmitch]

Why do I do it? Oh, jeez. That's easy. It's to give you guys another opportunity to tear it down. Call it lies. You know, the standard stuff. Status Quo.

Just take some notes from Will Robinson above. He's got it goin on!

Hmm....I figured you wanted to do it for the fun of it

Are you just gonna do runs at stock speeds? Or do you plan on doing some overclocking on the chips also? (as in a average overclock) <----Might be hard on fermi as average would be unkown still.

 

[SunnyD]

Please make sure to use SANE and CONSUMER FRIENDLY benchmark settings? I mean seriously, why do reviewers feel the need to benchmark with the absolute top end system overclocked to hell and back and feel those benchmarks actually mean something to the rest of the world?

 

[aigomorla]

I mean seriously, why do reviewers feel the need to benchmark with the absolute top end system overclocked to hell and back and feel those benchmarks actually mean something to the rest of the world?

*goes to hide my computer inside my closet*

 

[Keysplayr]

Please make sure to use SANE and CONSUMER FRIENDLY benchmark settings? I mean seriously, why do reviewers feel the need to benchmark with the absolute top end system overclocked to hell and back and feel those benchmarks actually mean something to the rest of the world?

Because if I don't overclock the CPU, I'll hear a bunch of "Why didn't you o/c?!?!, you know that GPU is being held back by your stock processor!!!!!"

Safer to O/C the CPU to lessen the possibility of a CPU bottleneck in a GPU benchmark suite.

I'm running a Core i7 860 at a pretty rock solid 3.4GHz, HT and Turbo enabled, which I think is probably adequate. Sure, higher the better, but I want a stable system and I haven't had the greatest of luck o/c'ing CPUs.

As for the overclocking benchmarks mentioned above, I may do a few, but don't count on it being any kind of extensive. As it is, I have over 10 titles to bench in a system configured once for Nvidia, and once again for ATI (all else the same save the graphics card and hard drive) at two resolutions.

 

[Qbah]

Can't wait for the results! You've included all the games I'm interested in so nothing more to add from me. Perhaps Lost Planet ? I remember this being kinda hard on hardware in DX10, especially with DX10 and AA.

And to the bashers - seriously, the review's not even out yet and you're already questioning Keys' motives? Heh

 

[Keysplayr]

Thank you Qbah.
Ok, I've updated my game library quite a bit.

So far, for DX11 titles and bench utilities, I now have:
Metro 2033
BattleForge
Aliens vs. Predator
Stalker CoP bench
Unigine Heaven Bench

Still need to DL Dirt2 Demo
Downloading BF Bad Co 2 Beta RC1. (Don't know how good this is going to be.)

We don't care about 3DMark Vantage do we? Just the games, yes?
Besides, I think FutureMark is working on a DX11 benchmark.