damn...AMDowned nvidia

[QUOTE="ronvalencia"]

[QUOTE="jhcho2"]it's nothing to do with 'heavy' games as you put it. Metro 2033 has always been biased in favour of ATI cards.

jhcho2

Metro 2033 is a Nvidia "The Way It's Meant To Be Played" title and the 7970 didn't exist during the design of Metro 2033.

http://www.nzone.ro/object/nzone_twimtbp_gameslist_uk.html search for metro 2033

7970 didn't win with light weight Blizzard created games, Call of Duty 4 DX9, which skewed the averages.

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

[QUOTE="ronvalencia"]

[QUOTE="jhcho2"]it's nothing to do with 'heavy' games as you put it. Metro 2033 has always been biased in favour of ATI cards.

jhcho2

Metro 2033 is a Nvidia "The Way It's Meant To Be Played" title and the 7970 didn't exist during the design of Metro 2033.

http://www.nzone.ro/object/nzone_twimtbp_gameslist_uk.html search for metro 2033

7970 didn't win with light weight Blizzard created games, Call of Duty 4 DX9, which skewed the averages.

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

hmm did you forget to add tom clancy's HAWX to that list, which for some reason runs better on older gen geforce 500 series cards than the new radeon 7000 series. Also metro 2033 is NOT biased, its simply that the radeon cards have more raw computing power and can push better frames, thats simply the truth. Think about it, metro 2033 and crysis were the most demanding games in their respective time periods. Crysis from 2007-2010, metro from 2010 to 2011 and now currently battlefield 3. ALL of these games radeon does better in then geforce. Its a simple explanation, and that is the number of transistors on the graphics cards. AMD put more on theirs, and nvidia put less. The reason being of course to save money and make more profit. Right now, it is a fact that the 7970 is the FASTEST single card GPU on the market. Until a benchmark is found to prove other wise then my claim stands. Good day folks.

Also the point of my laptop story is that nvidia graphics cards are NOT as reliable as radeon graphics cards. My laptop failure, the geforce 590 catching fire, the millions of geforce 8800's being baked and brought back to life. Hell I just found a link of nvidia's laptop graphic cards failures on cnet right herehttp://www.tomshardware.com/news/Nvidia-GPU-failure,6248.html, and a class action suit against them which WAS awarded, linkie herehttp://www.pcmag.com/article2/0,2817,2370032,00.asp, and also a thread on overclock.net for nvidia gpu baking link herehttp://www.overclock.net/t/748442/graphics-card-baking-sorta-tutorialas well as 100's of other sites. To date none of my radeon cards have failed, that is QUALITY right there.

If only AMD could put in this much effort into their cpus..... Intel is dominating them in that department.

[QUOTE="jhcho2"]

[QUOTE="ronvalencia"]

Metro 2033 is a Nvidia "The Way It's Meant To Be Played" title and the 7970 didn't exist during the design of Metro 2033.

http://www.nzone.ro/object/nzone_twimtbp_gameslist_uk.html search for metro 2033

7970 didn't win with light weight Blizzard created games, Call of Duty 4 DX9, which skewed the averages.

04dcarraher

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

I know that. I just wanted to give him some respite for the techpowerup results. But after years of comparing benchmarks, I know that a mere 5 fps lead means nothing. Changing anything from testing a different area or map in the game, settings or even driver or clockspeed by a little will change the lead. Even things like the CPU used would yield different results.

[QUOTE="jhcho2"]

[QUOTE="ronvalencia"]

Metro 2033 is a Nvidia "The Way It's Meant To Be Played" title and the 7970 didn't exist during the design of Metro 2033.

http://www.nzone.ro/object/nzone_twimtbp_gameslist_uk.html search for metro 2033

7970 didn't win with light weight Blizzard created games, Call of Duty 4 DX9, which skewed the averages.

blaznwiipspman1

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

hmm did you forget to add tom clancy's HAWX to that list, which for some reason runs better on older gen geforce 500 series cards than the new radeon 7000 series. Also metro 2033 is NOT biased, its simply that the radeon cards have more raw computing power and can push better frames, thats simply the truth. Think about it, metro 2033 and crysis were the most demanding games in their respective time periods. Crysis from 2007-2010, metro from 2010 to 2011 and now currently battlefield 3. ALL of these games radeon does better in then geforce. Its a simple explanation, and that is the number of transistors on the graphics cards. AMD put more on theirs, and nvidia put less. The reason being of course to save money and make more profit. Right now, it is a fact that the 7970 is the FASTEST single card GPU on the market. Until a benchmark is found to prove other wise then my claim stands. Good day folks.

Also the point of my laptop story is that nvidia graphics cards are NOT as reliable as radeon graphics cards. My laptop failure, the geforce 590 catching fire, the millions of geforce 8800's being baked and brought back to life. Hell I just found a link of nvidia's laptop graphic cards failures on cnet right herehttp://www.tomshardware.com/news/Nvidia-GPU-failure,6248.html, and a class action suit against them which WAS awarded, linkie herehttp://www.pcmag.com/article2/0,2817,2370032,00.asp, and also a thread on overclock.net for nvidia gpu baking link herehttp://www.overclock.net/t/748442/graphics-card-baking-sorta-tutorialas well as 100's of other sites. To date none of my radeon cards have failed, that is QUALITY right there.

And when would people learn that transistors, bits, and RAM, despite being essential, are not as important as the overall architecture itself. ATI fans should know this. ATI always has better designs. The 6950s are little more than mid range cards. The 6990 is a monster and beats the 590. So how does two mid range cards beat two high end cards? Its design.

[QUOTE="blaznwiipspman1"]

[QUOTE="jhcho2"]

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

jhcho2

hmm did you forget to add tom clancy's HAWX to that list, which for some reason runs better on older gen geforce 500 series cards than the new radeon 7000 series. Also metro 2033 is NOT biased, its simply that the radeon cards have more raw computing power and can push better frames, thats simply the truth. Think about it, metro 2033 and crysis were the most demanding games in their respective time periods. Crysis from 2007-2010, metro from 2010 to 2011 and now currently battlefield 3. ALL of these games radeon does better in then geforce. Its a simple explanation, and that is the number of transistors on the graphics cards. AMD put more on theirs, and nvidia put less. The reason being of course to save money and make more profit. Right now, it is a fact that the 7970 is the FASTEST single card GPU on the market. Until a benchmark is found to prove other wise then my claim stands. Good day folks.

Also the point of my laptop story is that nvidia graphics cards are NOT as reliable as radeon graphics cards. My laptop failure, the geforce 590 catching fire, the millions of geforce 8800's being baked and brought back to life. Hell I just found a link of nvidia's laptop graphic cards failures on cnet right herehttp://www.tomshardware.com/news/Nvidia-GPU-failure,6248.html, and a class action suit against them which WAS awarded, linkie herehttp://www.pcmag.com/article2/0,2817,2370032,00.asp, and also a thread on overclock.net for nvidia gpu baking link herehttp://www.overclock.net/t/748442/graphics-card-baking-sorta-tutorialas well as 100's of other sites. To date none of my radeon cards have failed, that is QUALITY right there.

And when would people learn that transistors, bits, and RAM, despite being essential, are not as important as the overall architecture itself. ATI fans should know this. ATI always has better designs. The 6950s are little more than mid range cards. The 6990 is a monster and beats the 590. So how does two mid range cards beat two high end cards? Its design.

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

jhcho2

LOL, your "Are you new to benchmarks?" statement is abit rich..

From http://www.techpowerup.com/reviews/Powercolor/HD_6970_Devil_13/17.html

On single GPU, GTX580 beats HD6970.

PS; I have been posting Tom's Hardware benchmarks and some certain NVIDIA fanboy doesn't like Tom's Hardware.

For example

http://www.tomshardware.com/reviews/radeon-hd-7970-ghz-edition-review-benchmark,3232-10.html

Tomshardware's Metro 2033 benchmarks didn't use "Very High"preset settings. Metro 2033 has the following presets i.e. Low, Normal, High (e.g. tomshardware), Very High (e.g. techpowerup).

[QUOTE="jhcho2"]

[QUOTE="ronvalencia"]

Metro 2033 is a Nvidia "The Way It's Meant To Be Played" title and the 7970 didn't exist during the design of Metro 2033.

http://www.nzone.ro/object/nzone_twimtbp_gameslist_uk.html search for metro 2033

7970 didn't win with light weight Blizzard created games, Call of Duty 4 DX9, which skewed the averages.

04dcarraher

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

Why targeting lower resolution for high end GPUs?

Lower resolution wouldn't maximize horizontal performance. My statement is in-line Anandtech's statement on shader related driver (i.e. JIT re-complier) improvements i.e. "AMDs optimizations are on the shader/texture side rather than ROP/memory"

As FirePro W8000 has shown, 256bit bus has miminial impact on HD 7970 type GPU @900Mhz. Note that W8000's GDDR5 (1375Mhz) is slower than GTX680's GDDR5(~1502Mhz) memory.

[QUOTE="04dcarraher"][QUOTE="jhcho2"]

Are you new to benchmarks? Metro 2033 has always yielded unfavourable results to Nvidia cards. ATI cards that would lose to certain Nvidia cards in every benchmark would suddenly win for Metro 2033. So far, I have identified two games which I would never use as a fair comparison between Nvidia and ATI. One is Enemy Territories: Quake Wars. The other is Metro 2033.

Second thing, next time post benchmarks from other sites like Guru3D or Tom's Hardware as well. Their results seems to show something different about Crysis 2.

ronvalencia

Why targeting lower resolution for high end GPUs?

Lower resolution wouldn't maximize horizontal performance. My statement is in-line Anandtech's statement on shader related driver (i.e. JIT re-complier) improvements i.e. "AMDs optimizations are on the shader/texture side rather than ROP/memory"

As FirePro W8000 has shown, 256bit bus has miminial impact on HD 7970 type GPU @900Mhz. Note that W8000's GDDR5 (1375Mhz) is slower than GTX680's GDDR5(~1502Mhz) memory.

you ignore the point or dont understand, and your statement proves it ""AMDs optimizations are on the shader/texture side rather than ROP/memory" if a 7970 is suppose to be stronger in processing shader workloads then a 680 then why dont it actually outperform by a good margin in all at 1920x1200 benchmarks ? The fact that at resolutions beyond 1080/1200 memory bus and the amount of memory plays a big role in performance

04dcarraher

"techpowerup results with the 12.11 drivers, it shows that the 7970 ghz is not clearly a faster gpu at resolutions that the gtx 670/680's memory bus can handle aka under 1440p"

Depends on the title. For this topic, Battlefield 3 can't be use skew the averages. I recall graphics cards includes memory chips and trace lines in the PCB.

you ignore the point or dont understand, and your statement proves it

The point doesn't exist since "AMD's optimizations are on the shader/texture side rather than ROP/memory" statement was about shaders/textures not your memory issue.

"if a 7970 is suppose to be stronger in processing shader workloads then a 680 then why dont it actually outperform by a good margin in all at 1920x1200 benchmarks"

You're being ignorant with 680's registers/cache vs ALU count limitations.

"The fact that at resolutions beyond 1080/1200 memory bus and the amount of memory plays a big role in performance"

I recall graphics cards includes memory chips and trace lines in the PCB.

Again, why target lower resolutions for high end GPUs? Why lower resolution(thus pixels/values) bar?

http://www.techpowerup.com/reviews/AMD/Catalyst_12.11_Performance/1.html

Basically, the 7970 is on par or stronger than the 680 gtx. Supposedly AMD was figuring out how to optimize drivers for a new architecture which is why theres alot of performance increases coming so late. It makes sense because the 7970 has more transistors than the 680 gtx meaning more raw power. AMD really pulled the wool over nvidias eyes with this one. On a side note, why didn't ionusx post this??

blaznwiipspman1

From http://pcpartpicker.com/parts/video-card/#c=101,70&sort=a5

On price, 7970 competes against GTX 670

red = 7970 win.

green = 670 win.