Getting ready for benches. Need a little advice.

[Keysplayr]

Keys ,when is the NDA officially lifted?

Unless I hear otherwise, I guess the date of the launch. March 26th.

 

[MegaWorks]

So you're going to test the 5870 vs GTX285?

 

[CrystalBay]

LOL this is funny ....

 

[Tempered81]

Resident Evil 5
CoD W@W
Far Cry 2
Crysis
Crysis Warhead
Dark VOID
Cryostasis
Batman: Arkham Asylum (Need to buy 2nd copy of this due to single reg)
Unigine Heaven Benchmark
3D Mark Vantage

List any other games you would like to see benched and I'll see if I can get them.

Modern Warfare 2
Bad Company 2
Unreal Tournament III
3d Mark 2006
Stalker
Clear Sky
Call of Pripryat
Borderlands
Bioshock 1 / 2
Wolfenstein
Mass Effect 2
Aliens vs. Predator
Dirt 2
Arma II
Anno 1404
Doom III
Left for Dead 1 / 2
Dawn of War II
HAWX

oh, And Warcraft & Shift

 

[CrystalBay]

I will say this it will be awhile before I replace a last generation product !

 

[Keysplayr]

So you're going to test the 5870 vs GTX285?

Yup.

 

[Kakkoii]

Should use Ati Tray Tools. Can do benchmarks with EATM on instead of Anti-Aliasing to compare performance. Can be enabled in ATT's advanced tweaks.

 

[bryanW1995]

Not that I expect you to comment, but if you're benching a GTX 480 against a 5870 you've already admitted it's not competing against the 5970. Just like you did a 2900XT against an 8800GTS because it clearly didn't compete with the 8800GTX.

If I'm not mistaken, that's more concrete than anything I've read about the underwhelming performance.

I assume that since he said "similar to my 2900xt vs 8800gts 640 comparo" he will be comparing it to a 470 gtx, but I could be wrong.

edit: nm, looks like it will be 5870 vs gtx 285. I would have waited 21 days and gone gtx 470/480 vs 5870 or 5970, but this will still be cool. Keys does great benches.

 

[BFG10K]

Yes, the 5870 vs GTX285 will be interesting because it represents the two flagship single GPUs from each company. If I had a 5870, I’d compare the two cards too.

The first game I’d test would be Doom 3.

 

[dug777]

Yes, the 5870 vs GTX285 will be interesting because it represents the two flagship single GPUs from each company. If I had a 5870, I’d compare the two cards too.

The first game I’d test would be Doom 3.

No doubt good to get the benchmarks from you and Keys, but it's rather an out of date comparison these days, it has been done many, many times out there, and the 5870 appears to utterly crush the GTX285.

AT's latest review (5830) contains both cards, and the GTX285 performs much closer to the 5850 which still generally outclasses it.

http://www.anandtech.com/video/showdoc.aspx?i=3750&p=4

Looking at newegg now, the cheapest GTX 285 is $355 and the cheapest HD 5850 is $310, while the cheapest HD 5870 is $380.

Let us hope that Fermi shakes that up, but the GTX 285 seems like a pretty hard sell in those circumstances...

 

[akugami]

Keys, I'll assume you were just being snide to the 5870 vx 285 comment.

Anyways, I'll suggest that you run your tests through three resolutions in your tests. We've seen before how various GPU's can scale up in resolution relatively well, while other GPU's will suck going over a certain threshold. I'll also suggest that since benchmarking can be time intensive that you limit some of the games. For instance, no need for five different games utilizing the Unreal engine or whatever unless they have been heavily modified. Finally, in narrowing down the games, choose some that have traditionally favored ATI, and some that have favored nVidia.

 

[3DVagabond]

Unless I hear otherwise, I guess the date of the launch. March 26th.

Can you please confirm for us that it's not a gtx285. (Sorry for being pedantic )

I'm hoping that it means AT got supplied a 480 for testing. It means at least some do really exist in the wild and we'll get an accurate, unbiased review.

 

[Madcatatlas]

Not AT, but Keys. Atleast so i figure.

Read his sig.

Looking forward to your bench of 5870 vs 480!

 

[3DVagabond]

Maybe I'm assuming too much because he's a mod here? I won't hold his "friends" against him. (speaking of his sig)

 

[shaynoa]

looking forward to then to but you know not one of these games will really give you a true benchmark
Modern Warfare 2
Bad Company 2
Unreal Tournament III
3d Mark 2006
Stalker
Clear Sky
Call of Pripryat
Borderlands
Bioshock 1 / 2
Wolfenstein
Mass Effect 2
Aliens vs. Predator
Dirt 2
Arma II
Anno 1404
Doom III
Left for Dead 1 / 2
Dawn of War II
HAWX
not one has the depth that trainz 2010 has in a big map, contact auran .com to get the game and a big map then you will see which card can do the hard work,well
time for bed 3 am start at sydney markets,dirty job but someone has to do it

 

[Soleron]

Don't you think it would be most prudent to actually WAIT for the benchmarks first? Just a thought.

Well, yes, but since there is zero good information on this (Nvidia being completely quiet) I have to take what I can see.

If you're only testing against a 5870 because you don't have a 5970, then wouldn't it be better to say that rather than give that impression?

 

[Keysplayr]

Well, yes, but since there is zero good information on this (Nvidia being completely quiet) I have to take what I can see.

If you're only testing against a 5870 because you don't have a 5970, then wouldn't it be better to say that rather than give that impression?

There is a difference between taking what you see for what it is, and taking what you see
and murdering it's context.

Your second sentence is an example of the latter. Enough of this please. I know everyone is hungry for information, but all I'm doing right now is getting ready, as the thread title states, for my benches. I'm using a 5870 because it's probably what everyone wants to see. The fastest single AMD GPU compared to the fastest Nvidia GPU. I have asked for assistance here in accomplishing this so when my benches are completed and I post them, I'll know that I did the best job I could've done. Reliable, sensical, games people want to see with particular settings.

All too often we see threads commenting about various reviews with things like, "I wish they would have......" or "They never do this......."

I like doing things this way to eliminate any room for debate, or at least absolutely minimize it. Nobody is perfect, but we can get close where most are happy with the results.

I'm going to start working on getting as many of these games as possible. I may have to go out and get a pair of 1TB drives!! LOL, a lot of games!

 

[Soleron]

There is a difference between taking what you see for what it is, and taking what you see
and murdering it's context.

Your second sentence is an example of the latter. Enough of this please.

Ban me.

 

[Keysplayr]

Keys, I'll assume you were just being snide to the 5870 vx 285 comment.

Anyways, I'll suggest that you run your tests through three resolutions in your tests. We've seen before how various GPU's can scale up in resolution relatively well, while other GPU's will suck going over a certain threshold. I'll also suggest that since benchmarking can be time intensive that you limit some of the games. For instance, no need for five different games utilizing the Unreal engine or whatever unless they have been heavily modified. Finally, in narrowing down the games, choose some that have traditionally favored ATI, and some that have favored nVidia.

Sounds excellent. Can you make some suggestions? Much appreciated.

As far as 3 resolutions go, keep in mind that the highest I could run right now is 19x12. I'm pretty sure that would accomodate most gamers.

So, I would run:

1920x1200
1680x1050
1440x900

I'd run out and get a 3007 Dell, but I'm not THAT wealthy.

 

[Keysplayr]

Ban me.

Hmm. Not what I expected, but if you wish to be banned, please contact the Anandtech Moderator account and explain to them why you wish to be banned.

I don't know what else to tell you. Now, may we get on with our nice thread here?

 

[Soleron]

Hmm. Not what I expected, but if you wish to be banned, please contact the Anandtech Moderator account and explain to them why you wish to be banned.

Well played. I wish it was clearer whether it was worth waiting for the 480, that is all. I have no further need to post.

 

[SlowSpyder]

You should also do some overclocking of the 5870 and any Nvidia card you might test.

 

[Keysplayr]

Well played. I wish it was clearer whether it was worth waiting for the 480, that is all. I have no further need to post.

I understand. But here it is, March 7th already. If you waited this long, no need to sweat 2.5 weeks.

 

[Keysplayr]

You should also do some overclocking of the 5870 and any Nvidia card you might test.

I could do that for a couple of benches. It would be an additional monumental undertaking to run all benches over again at o/c'd speeds. Have a little mercy on me.

 

[Voo]

Sounds excellent. Can you make some suggestions? Much appreciated.

As far as 3 resolutions go, keep in mind that the highest I could run right now is 19x12. I'm pretty sure that would accomodate most gamers.

So, I would run:

1920x1200
1680x1050
1440x900

I'd run out and get a 3007 Dell, but I'm not THAT wealthy.

I'd say if you can't do 2560x1600 just use 1680 and 1920 - I don't think the results for 1440x900 would be that interesting (who'd buy a 5870 or 470 if he had such a monitor? Also you run into CPU limited territory with some games I assume) and it would be a lot less work for you. I'd prefer more benchmarking results or maybe different AA levels in some games.