All benches with volari ultra duo only used one chip (gpu)!

[ByteMe]

It just hit me. OMG what would this mean?

 

[xGL]

If it is true, I guess that would explain the "not so good as expected" performance, but I don't really get the point of the question.

 

[bloodbob]

Have you compared it to the top end single chip solutions because unless they are about 1/2 the utlra then you theory is debunked :/

 

[ByteMe]

I will admit that I am disappointed with the "beta" performance. They have 50 million more transistors than the r350 and get alot less performance?

Unless the "beta" drivers have some fatal flaws... I see them having no chance in making it. (or my theory is correct)

 

[Martin Eddy]

If it's true then we can possibly expect an increase of maybe about 50% on it's performance.

The problem is by the time they're drivers mature, ATI and nVidia are going to have there next gen high end cards out.

 

[digitalwanderer]

Have you compared it to the top end single chip solutions because unless they are about 1/2 the utlra then you theory is debunked :/

No, but I haven't seen all that many benchies of anything other than the Duo.

I hope you're right ByteMe, but I have a feeling that XGI would be mentioning their problems with it so they could boast of future performance gains.

 

[X2]

Or maybe there's just too many games that break AFR.

 

[xGL]

I will admit that I am disappointed with the "beta" performance. They have 50 million more transistors than the r350 and get alot less performance?

Really? So the R350 has only 20 to 30 million transistors?????

 

[X2]

I will admit that I am disappointed with the "beta" performance. They have 50 million more transistors than the r350 and get alot less performance?

Really? So the R350 has only 20 to 30 million transistors?????

2x80 = 160

 

[CorwinB]

Re: All benches with volari ultra duo only used one chip (gp

It just hit me. OMG what would this mean?

That they shouldn't have released it in this sorry state, perhaps ?

 

[RussSchultz]

Is it using 1 chip as a fact? Or as a speculative theory to explain its poor performance?

 

[Tim Murray]

Is it using 1 chip as a fact? Or as a speculative theory to explain its poor performance?

A speculative theory. It's not using 1 chip, it's using two, unless the non-Duo cards are only using half a chip. In that case, I would tattoo "OMGWTF" to my forehead.

 

[digitalwanderer]

Is it using 1 chip as a fact? Or as a speculative theory to explain its poor performance?

A speculative theory. It's not using 1 chip, it's using two, unless the non-Duo cards are only using half a chip. In that case, I would tattoo "OMGWTF" to my forehead.

/me fervently prays that the Baron's speculative theory turns out true just to see the picture of Baron with "OMFGTF" tatooed to his forehead!

 

[John Reynolds]

/me fervently prays that the Baron's speculative theory turns out true just to see the picture of Baron with "OMFGTF" tatooed to his forehead!

I'll tell you what, if he (Baron) is wrong than I'll put those words and not their abbreviation on my forehead (because even spelled out they'll still fit).

 

[OpenGL guy]

/me fervently prays that the Baron's speculative theory turns out true just to see the picture of Baron with "OMFGTF" tatooed to his forehead!

I'll tell you what, if he (Baron) is wrong than I'll put those words and not their abbreviation on my forehead (because even spelled out they'll still fit).

(To the tune of Eric Clapton's "Cocaine")
"If she gives you a hug and you lose your rug: Rogaine."



Sorry, I can't recall more of the parody.

 

[StealthHawk]

No, but I haven't seen all that many benchies of anything other than the Duo.

I hope you're right ByteMe, but I have a feeling that XGI would be mentioning their problems with it so they could boast of future performance gains.

If you check out Tom's VGA Charts you will see benchmarks where the VolariV5 beats out the VolariV8Duo

 

[Ailuros]

Why I should judge a chip's or board's capabilities and/or performance from it's transistor count is beyond me.

 

[RussSchultz]

Transistor count (or die area) is an indication of cost, and hence integral to determining "efficiency".

Equivalent die area should result in equivalent performance (in an ideal world, of course).

 

[ByteMe]

Why I should judge a chip's or board's capabilities and/or performance from it's transistor count is beyond me.

Off a few million, no biggie. Off by 40%, then that company has got problems. They would need to go hire some engineer's that knew wtf they are doing.

 

[micron]

/me fervently prays that the Baron's speculative theory turns out true just to see the picture of Baron with "OMFGTF" tatooed to his forehead!

I'll tell you what, if he (Baron) is wrong than I'll put those words and not their abbreviation on my forehead (because even spelled out they'll still fit).

(To the tune of Eric Clapton's "Cocaine")
"If she gives you a hug and you lose your rug: Rogaine."



Sorry, I can't recall more of the parody.

I just spit milk