Ati Back on top again

[Tempered81]

Originally posted by: chizow

Originally posted by: jaredpace
chizow i dont think the 8800gtx is twice as powerful as the 1950xtx

+/- 10%?

I just reviewed some of those benchmarks to confirm my initial impressions from over a year ago and they hold true, but feel free to flip through some more of them. Honestly this isn't a point many are going to contest regardless of which side of the fence you're sitting on. G80 was simply that damn good.

Yeah when benching at 2560 res with 4x aa yes.


at 1024 max quality its only 8% faster than an xtx in prey

http://www23.tomshardware.com/...9&model2=722&chart=299

Originally posted by: chizow
I went from a Ti4400 to a 9700pro and I don't remember it being anywhere close to 2x as fast as the Ti4400 (a quick refresher looking over old benchmarks confirms this).

a 9700 pro is about 3 times faster than a ti4600 here:

http://www.tomshardware.com/20...a_charts_iv/page7.html

 

[Munky]

Originally posted by: JACKDRUID

Originally posted by: munky

Originally posted by: JACKDRUID

Originally posted by: munky

Looking for an explanation of two questionable aspects of their FEAR test results.

are you trying to judge the validity of HOCP reviews based on a single bench? non-biased people usually base their judgment on the whole review, which they clearly mention 1900xt as the faster card in both reviews...

don't become an Ackmed

Non-biased people usually base their judgment on several trusted sources. This one source just happens to have results that sometimes contradict what other sources say, while using an unscientific and unreproducible testing methods.

no the result does not contradict other sources.. its just one bench out of many... yet the result still align with other sources.

One bench out of many, which to me looks like it was copied and pasted from another article 2 months earlier. How do you expect me to believe any of their other "results"?

 

[Keysplayr]

Originally posted by: munky

Originally posted by: JACKDRUID

Originally posted by: munky

Originally posted by: JACKDRUID

Originally posted by: munky

Looking for an explanation of two questionable aspects of their FEAR test results.

are you trying to judge the validity of HOCP reviews based on a single bench? non-biased people usually base their judgment on the whole review, which they clearly mention 1900xt as the faster card in both reviews...

don't become an Ackmed

Non-biased people usually base their judgment on several trusted sources. This one source just happens to have results that sometimes contradict what other sources say, while using an unscientific and unreproducible testing methods.

no the result does not contradict other sources.. its just one bench out of many... yet the result still align with other sources.

One bench out of many, which to me looks like it was copied and pasted from another article 2 months earlier. How do you expect me to believe any of their other "results"?

I don't think he expects you to believe them. I believe that he expects you to believe that he believes them. And that's about as far as either of you can take the debate. LOL.

 

[bryanW1995]

Originally posted by: McCartney
No. I didn't expect a job at all who told you that? I had questions about his CAREER. Please read. I don't believe in nepotism at all nor do I practice it. I just think that having an inside look at the industry would be a great thing.
Stop trying to use extrapolation because obviously you weren't very good at logical reasoning.

Dude, I totally understand. I wanted to be bill clinton's national security advisor b/c I lived in arkansas for 3 months and I'm a white male. I even like interns. Did he return my calls? Hell, no!! Maybe monica will, however...hmmm...

 

[bryanW1995]

Originally posted by: McCartney
Um you're not listening again idiot. OPEN YOUR EYES> I said the MEMORY on it is faster. On the 7xxx's the MEMORY was key to speeds.
Go google 7800 512 7900 gtx and come back to me.

You don't program anything as far as I'm concerned. You have 5300 posts in 3 years. Wow that's awesome. So in about 1095 days your rough post average of about 4.9 a day consists of "I know my hardware but I didn't know that hte 512 had some of the best memory on video cards"?

Is that your rubettual? pretty weak imo


Say bye bye to all your nice new friends for about a week. Hey, you were warned.

Anandtech Moderator - Keysplayr2003

damnit, I was waiting for his/her/its next comment!!! way to go, keys!

 

[BFG10K]

I believe the 9700 Pro was such a huge leap over the 8500 and Ti4600 (5800U was late), and was overall such a superior card in every way.

Sure, but exactly the same applies to the 8800 GTX which was pretty much twice the speed of the previous generation while offering superior AF, AA and DX10.

It gave you more than just DX9 in terms of new tech. DX9 caught on faster in games than DX10 too. But thats not fault of the GTX.

DX10 has largely been irrelevant on the G80; it's because it shines in DX9 is why it?s so good. That and because it allows 4xAA as a mandatory minimum standard in pretty much any game out there.

SLI lowered the impact the GTX gave. Because you could use a 7950 or SLI'd 7900's. The GTX was still faster than those still, but if not for those, the impact of speed would have been even greater.

I disagree and if anything it highlighted how much better the card was because it could beat a pair of previous generation cards while generating much higher IQ.

Obviously using the 9700 Pro for the first time was not like the first time I put in a Voodoo card, but in a second behind that.

Had my experience not been marred by driver issues I would have been blown away just as much as the 9700 Pro, if not more. From a performance and IQ standpoint the leap was huge, I just couldn?t play many games for a long time due to driver issues.

 

[BFG10K]

at 1024 max quality its only 8% faster than an xtx in prey

1024 is irrelevant to a 8800 GTX given you're not even testing the GPU but are instead testing the CPU/platform.

Your benchmark doesn't even use AF or AA which is even more ludicrous. If you want to test CPUs go and run a ?CPUMark? or something instead of wasting time with a 8800 GTX.

This is why for years I?ve campaigned against benchmarks without AA/AF and against benchmarks run at low resolutions like 1280x1024. It?s disingenuous to even claim you?re testing a GPU at such settings.

Given both vendors support 16xAF and 4xAA/8xAA we shouldn?t see any benchmarks without them unless it?s physically impossible to enable one or more said features.

 

[chizow]

Originally posted by: Ackmed
Im not sure where I said it was 2x as fast in my post you quoted. You may want to take a look at AT's review though. I said the 9700 Pro gave you performance at settings before unthinkable. And said that it gave you playable frames with AA/AF, far above the Ti4600. If you look at the AT review I linked, and the AA/AF benchmarks, its pretty clear I was right. Especially in "higher resolutions", of 1600x1200. Heck, in one test, the 9700 Pro got 100fps and the Ti4600 got 20fps in Jedi Knight 2 at 1600x1200 with AA/AF. Thats 5 times as fast. Generally it is easily 2-4 times faster with AA/AF at a higher resolution in the games they tested. Thats pretty impressive. AA/AF killed the Ti series. The 9700 Pro could even use 6xAA in a lot of games.

I had a Ti4200, one of my best cards Ive ever owned. It was a very healthy upgrade for me, my biggest yet. Aside from my Rage3D to my Voodoo, which doesnt really count. At least till next time its the biggest upgrade. Im on a old X1900XT, got it when it first came out. Actually had two in CF. Couldnt upgrade because I was gone for 7 months.. when newer cards just came out. And now getting ready to be gone for 13 months. So Im suffering with this card till I get back, June 09'. Then Ill buy whatevers the best, and catch up on the games I missed. Which sucks because my card is having a great deal of difficulty playing current games at my native res of 1920x1200. Im resorting to gaming more on my consoles.

But as I said, its all opinion. The GTX certainly is no slouch, I just dont think it was the break thru of the 9700 Pro. You do, and thats fine, dont have any problems with that. Its our opinions. It really doesnt matter much, I think we can both agree that this 3870x2 launch is not like the 9700 Pro launch. Which was the topic of the post.

Yep, the 9700pro did definitely standardize IQ settings like AF and raised expectations with running AA. But the key difference is that it wasn't a sink or swim situation when compared to the GF4 series. Without AA or AF, you could still run similar resolutions and similar framerates with a GF4, the 9700pro just pulled way ahead once you enabled AA or AF. In many cases you could enable both and still maintain playable framerates. With the GF4 you could still play at higher resolutions and high frame rates, you just had to sacrifice IQ.

With the G80, it doubled performance across the board with or without AA. That was often the difference between playable or non-playable at higher resolutions like 1920+. AF was free and AA was realistic even at higher resolutions that bordered unplayable with older cards. That's the main distinction imo, and why the G80 had a greater impact on my overall gaming experience.

 

[chizow]

Originally posted by: jaredpace

Originally posted by: chizow

Originally posted by: jaredpace
chizow i dont think the 8800gtx is twice as powerful as the 1950xtx

+/- 10%?

I just reviewed some of those benchmarks to confirm my initial impressions from over a year ago and they hold true, but feel free to flip through some more of them. Honestly this isn't a point many are going to contest regardless of which side of the fence you're sitting on. G80 was simply that damn good.

Yeah when benching at 2560 res with 4x aa yes.


at 1024 max quality its only 8% faster than an xtx in prey

http://www23.tomshardware.com/...9&model2=722&chart=299

Originally posted by: chizow
I went from a Ti4400 to a 9700pro and I don't remember it being anywhere close to 2x as fast as the Ti4400 (a quick refresher looking over old benchmarks confirms this).

a 9700 pro is about 3 times faster than a ti4600 here:

http://www.tomshardware.com/20...a_charts_iv/page7.html

G80 No AA
It doesn't need high resolution or AA to maintain its healthy 2x lead over predecessor parts. Unlike the GF4 and 9700pro, where performance was comparable until you enabled AF and AA (which weren't expected or standard back then). Linking CPU limited resolutions is kinda pointless when arguing the merits and performance of a GPU.

 

[JACKDRUID]

Originally posted by: chizow

G80 No AA
It doesn't need high resolution or AA to maintain its healthy 2x lead over predecessor parts. Unlike the GF4 and 9700pro, where performance was comparable until you enabled AF and AA (which weren't expected or standard back then). Linking CPU limited resolutions is kinda pointless when arguing the merits and performance of a GPU.

Agreed...

 

[Tempered81]

Originally posted by: chizow

G80 No AA
It doesn't need high resolution

If 2560 is not a high resolution to you, then what is? What does your monitor run?

At <1600 the cards are closer. Yes, at >1600 with any AA or AF, the 8800gtx annihilates at about 190% the speed of the xtx.

 

[chizow]

Originally posted by: jaredpace

Originally posted by: chizow

G80 No AA
It doesn't need high resolution

If 2560 is not a high resolution to you, then what is? What does your monitor run?

At <1600 the cards are closer. Yes, at >1600 with any AA or AF, the 8800gtx annihilates at about 190% the speed of the xtx.

The link also includes 1600 and 1920 resolutions below the 2560 graphs. I run 1920 native but I can scale down at 1:1 if I need to. Keep in mind these are for older games, in newer games you would be able to see similar differences at lower resolutions as the shift went from CPU to GPU limited (Crysis, WiC, CoD4 etc). Which brings us to another by-product of G80's dominance and impact on the GPU market. You'd be hard pressed to find newer benchmarks with X1900s and GF7 parts because the G80 and all its variants made all predecessor parts irrelevant more than any card before it. Many wonder what NV was thinking by releasing so many similar parts that cannibilzed their other parts. I think its pretty obvious what they were trying to accomplish looking at recent benchmarks, and I'd say it worked as you'll often see 4-5 G80/G92 variants with 1-2 ATI representatives.

 

[NinjaJedi]

No one brings up the UVD factor. Would 2 GPUs enhance HD/Blu-ray playback?

 

[BFG10K]

No one brings up the UVD factor. Would 2 GPUs enhance HD/Blu-ray playback?

Unlikely.