Shamino's Full Review of The X1950XTX

[Your_Boss]

Here are some benches from HardwareZone before they pulled it.











I don't know why this review looks so much slower performance wise for the 1950XTX than vr-zone.

Maybe they realized they screwed up their benchmarks and are redoing them right now.(maybe vise-versa )

Anyway these are the pics of their results BEFORE they pulled their review article, enjoy.

[pitata]

http://resources.vr-zone.com.sg/Shamino/1950/

Thank You

[ahmad]

ok forget the game benchmark numbers for the time being, I got flamed pretty bad all around by readers, I gotta figure out what went wrong with my 7950GX2 runs.

Since the date i can publish the review is pushed back quite a bit so i'll redo it and redo the setup.

the rest is still pretty relevant though

Thats BS man. If you can run all the 3DMarks and get a higher score, the games should work just the same.

[Vengure]

Thats BS man. If you can run all the 3DMarks and get a higher score, the games should work just the same.

I think games and GFX benchies are two diff things. In a game you'll have alot of area's that require little of the card but when alot of action is on the card its stressing it. Benchies are full load high res trying to render large maps and intricate designs. When your getting shot and so much comotion is going on in game I think its a bit different from 3dmark my self

[cantankerous]

thanks ever so much for those graphs. I missed the original article and was searching everywhere for the results it produced.

EDIT... just gave a quick glance at the results and I must say I am dissapointed, especially after reading the VR-Zone results which painted a much prettier picture of the card. I am upset to see the 3DMark scores actually go down while any gains seen elsewhere were extremely minimal. I am glad to see the temps are a little lower, but without info on system tested on, fan speed setting, overclock (if any), as well as ambient temps during testing, makes this review hard to go by as well. Shamino just hinted the card may be pushed back until the middle of September before release, so that should give more time for other reviews, hopefully end users, to put in their commends and results for comparison purposes. I was stoked about getting this card and selling my current but perhaps it really isn't worth it after all.

[ahmad]

Here are some benches from HardwareZone before they pulled it.
..............
I don't know why this review looks so much slower performance wise for the 1950XTX than vr-zone.

Maybe they realized they screwed up their benchmarks and are redoing them right now.(maybe vise-versa )

Anyway these are the pics of their results BEFORE they pulled their review article, enjoy.

Thats because:

1: Crappy AF settings. These are high end cards that can handle it
2: Chronicles, Quake4, FEAR (was even done with no AA) are all games we know that Nvidia is better at (also, FEAR takes advantage of the raw power of 2 cards). Who cares about OpenGL performance?

Sad. Why dont they show some HL2 Episode 1, or serious sam, or farcry. Hardwarezone goes on the list of sites not believe anything from

[rpg711]

Thats BS man. If you can run all the 3DMarks and get a higher score, the games should work just the same.

u dont have a clue what ur talking about
ddr4 memory has higher latencies, hence the lower 3d mark scores
latencies matter only in benchies and software stuff
latencies matter none in games, where the x1950xtx's memory bandwidth is used to higher potential, the x1950xtx's memory bandwidth is used more efficiently than the 7950gx2... and the memory clock also matters alot in games

[Eastcoasthandle]

If they pulled it, they screwed up the benchmarks bottom line, there is no mystery there.

[mdzcpa]

Actually Hardware Zones look pretty accurate all around compared to other reviews....except for the new X1950XTX results of course. The GX2 and X1900XT results look right on the money.

[smith_6079]

ok forget the game benchmark numbers for the time being, I got flamed pretty bad all around by readers, I gotta figure out what went wrong with my 7950GX2 runs.

Since the date i can publish the review is pushed back quite a bit so i'll redo it and redo the setup.

the rest is still pretty relevant though

Just because you got flamed by people who didn't believe your results doesn't mean you should scrap your findings as irrelevant. Some do believe your results to be correct.

Time will tell whether you were acurate or not. Until then, don't concern yourself with what other angry 7950GX2 owners are yelling at you for. You did what you could as accurately as you could with the time frame you were given. Thanks.

[sabrewolf732]

I think games and GFX benchies are two diff things. In a game you'll have alot of area's that require little of the card but when alot of action is on the card its stressing it. Benchies are full load high res trying to render large maps and intricate designs. When your getting shot and so much comotion is going on in game I think its a bit different from 3dmark my self

No, you have it backwards. Since games are run at high res and AA + AF the video card is getting stressed hence the small differances in perf. between cpu's and other small things in gaming environments. Only the gpu matters (for the most part) in gaming environments. Since benchmarks are run at lower res and settings and are made to evaluate the entire system's performance cpu and other sublities such as ram timings etc. play a much bigger role in the results.

[zakelwe]

I think they were pulled after people had words rather than any questions on the results.

Regards

Andy

[[cTx]Philosophy]

No offence to you shamino, we all know your top in the overclocking community and well respected bro..

Im glad to see that it was pulled ^^^ I actually soaked my pillow in tears after accepting my stepup from evga thinking I was finally gonna have a top performing card, and then read your article and got shot down hard..Hopefully after they allow you another shot at the card itll be a different story..

[gr8golf]

Actually Hardware Zones look pretty accurate all around compared to other reviews....except for the new X1950XTX results of course. The GX2 and X1900XT results look right on the money.

I think you're right Mike. Will be interesting to see how it looks after a few other reviews / tests have been run using some other drivers on both sides. I'd expect the Xfire with the X1950 to compete with the 7950GX2, but not the single card. I'm not a fanboy for either camp, if one is substantially faster than the other for the same money - I'll buy it. I also consider the noise and heat levels too - and the NVidia options have been better in those two areas lately. I hope the X1950 changes that.

Does LardArse = Shamino?

[kenofstephen]

No offence to you shamino, we all know your top in the overclocking community and well respected bro..

Im glad to see that it was pulled ^^^ I actually soaked my pillow in tears after accepting my stepup from evga thinking I was finally gonna have a top performing card, and then read your article and got shot down hard..Hopefully after they allow you another shot at the card itll be a different story..

what do you mean?

[Okda]

what do you mean?

he means that he got his 7950GX2 from evga and then he finds according to this review that X1950XTX AND X1900XTX is better than his card which is not true

[Origin_Unknown]

will there be a need for a master card for cf? or can you just use two x1950xtx's?

[Kinc]

will there be a need for a master card for cf? or can you just use two x1950xtx's?

Yeah you will need a X1950 master card. It will cost the same as the slave and have the exact same specs.

[don_vercetti]

That's a pleasant change. Do they all clock +- equally as well?

[Origin_Unknown]

Yeah you will need a X1950 master card. It will cost the same as the slave and have the exact same specs.

hmm, thought they were getting rid of that and having it done internally (sli style)

[Phase]

After reading this review (Before it was pulled) it seems quite obvious that the 7950, was not running both cards, but only a single card was functioning properly. (In the gaming tests)

Whether or not it was software or hardware related i'm sure Shamino will figure it out and run a new test.

[JuanFlaiter]

Shamino did not pull out the review because the benchmarks where not correct, he pulled it out because the red guys asked him to do it.

If Peter says that are the results then those are the results, bottom line.

[Phase]

Hey take it easy now
I am not questioning his integrity.

He knows there was something wrong with the 7950 results and has said as much.

[rpg711]

That's a pleasant change. Do they all clock +- equally as well?

all cards clock differently, because all cores are different
if u get two cards from the same batch, one may overclock to 900/1500 and one might not at all... silicon is weird

[don_vercetti]

all cards clock differently, because all cores are different
if u get two cards from the same batch, one may overclock to 900/1500 and one might not at all... silicon is weird

Helpful answer. Read my post again, think about it, and then try and come up with a constructive reply.

I said +- well. I know all silicone is not equal, everyone does. It's just that the X1900 Crossfire was based off the X1900XT rather than XTX, so it didn't clock as well as an X1900 Crossfire might have done had it been based off XTX, hence limiting the OC potential of Crossfire somewhat. However if the X1950 Crossfire clocks as well as the X1950 XTX...then we're very happy!