[Hardwareluxxx]Breaking news: AMD Fury X will be slower than GX 980Ti

[n0x1ous]

They obviously know everything about the performance. But they couldnt give hardwareluxx exact numbers due to NDA stuff.

exactly

 

[Headfoot]

You don't think an AIB partner of AMD's knows the performance?

Lol really? If the BIOS isn't finalized it is UNKNOWABLE. I dont see how this is hard to grasp. In case this is news for anyone: clockspeed affects performance... a lot. So do power tune targets. And fan speeds. All of which are set in BIOS.

Remember when AMD dropped the 47% boost in fan speed for 290s when it came out which ended up boosting performance quite a bit?

 

[LTC8K6]

I guess this is the reason why we havent seen AMD launching these cards yet...
They should have launched it before Nvidia laubched 980Ti. Atleast against the expensive Titan X it would sell.

Now its like the worst timing if its slower and facing a much cheaper GM200 GPU

We're just guessing, but I think it's something like this:

AMD was going to release Fury soon.

NV released 980ti first.

980ti was faster than anyone thought.

Fury was then not fast enough to release right after the 980ti.

Fury had to be reset, for release a little later with faster clocks.

 

[n0x1ous]

Lol really? If the BIOS isn't finalized it is UNKNOWABLE. I dont see how this is hard to grasp. In case this is news for anyone: clockspeed affects performance... a lot. So do power tune targets. And fan speeds. All of which are set in BIOS.

Remember when AMD dropped the 47% boost in fan speed for 290s when it came out which ended up boosting performance quite a bit?

LOL UNKNOWABLE. YES! even AMD doesn't have a hint of a clue how it performs yet its getting released this month /s

just because clocks aren't finalized doesn't mean AMD and its partners don't have a strong idea of how it will end up.

 

[Headfoot]

We're just guessing, but I think it's something like this:

AMD was going to release Fury soon.

NV released 980ti first.

980ti was faster than anyone thought.

Fury was then not fast enough to release right after the 980ti.

Fury had to be reset, for release a little later with faster clocks.

Yup. That's why its pointless to pretend like you know what the exact performance is until its packaged in a box.

Dollars to donuts, AMD almost certainly is revising the bios to bump power use and clocks in order to meet/beat 980 Ti now that nVidia's played its hand.

AIBs will obviously know the general contours of performance (better than 980 and 290x) but the last 10-20% is far too dependent on BIOS definitions to know if you don't have the BIOS...

 

[Headfoot]

LOL UNKNOWABLE. YES! even AMD doesn't have a hint of a clue how it performs yet its getting released this month /s

just because clocks aren't finalized doesn't mean AMD and its partners don't have a strong idea of how it will end up.

I never said that "they don't have a strong idea". You're pulling that from thin air. The claim is, Fury slower than 980 Ti. They can't know that without a BIOS, period. Too many factors depend on it. They can have a strong idea of the performance generally -- but they cant make the claim one is faster than the other if the BIOS is still being built. Which is abundantly obvious to anyone without an agenda.

Logic: Everyone including AIBs know Fury is in the general ballpark of 980 Ti. (Of course I acknowledge AIBs know the ballpark performance, and I have NEVER claimed otherwise). That means the last 10-20% of performance that depends on BIOS. If Fury and 980 Ti are within 10-20% performance, BIOS is the determinant factor. I believe this to be the case.

 

[Makaveli]

Which is it cloudfire you create this thread saying its slower then you have another thread saying it will be priced at $850.

How can it be priced at $850 if they know it will be slower than a $649 980 Ti?

 

[Genx87]

I'd be more worried they are still tweaking clock speeds at this point.

 

[exar333]

Yup. That's why its pointless to pretend like you know what the exact performance is until its packaged in a box.

Dollars to donuts, AMD almost certainly is revising the bios to bump power use and clocks in order to meet/beat 980 Ti now that nVidia's played its hand.

AIBs will obviously know the general contours of performance (better than 980 and 290x) but the last 10-20% is far too dependent on BIOS definitions to know if you don't have the BIOS...

It is troubling though because if AMD is tweaking (raising) clocks and power-usage, they are 'borrowing' OC headroom and incorporating that into the core product. Think 7970Ghz edition released right off the bat. It's fine, but it is telling that the raw performance wasn't really there.

IMHO the Fury branding was counting on the fact that it was competing against a $999 TX. A $649 980Ti is a very different benchmark if they are trying to compete against NV's best. A WCE Fury XT at $849 (just for argument's sake) that is 99% of TX, but ~ 980Ti will be a non-starter....

Edit: Anyone suggesting the AIBs don't know the performance, or at least very close to it, are just delusional. They know exactly what it will be. That is how they determine their own SKUs and decide what models to release (WCE, triple cooler, etc.) and how to price them.

 

[coercitiv]

Which is it cloudfire you create this thread saying its slower then you have another thread saying it will be priced at $850.

He'll be right, one way or the other

 

[n0x1ous]

I never said that "they don't have a strong idea". You're pulling that from thin air. The claim is, Fury slower than 980 Ti. THEY CANT KNOW THAT YET WITHOUT A BIOS, PERIOD. Too many factors depend on it. They can have a strong idea of the performance generally -- but they cant make the claim one is faster than the other if the BIOS is still being built.

Do you deny that CLOCKSPEED, POWER TARGET (TDP) AND FAN SPEED rely on the BIOS? Yes or no. If you do not answer, or answer other than yes or no I will take that as an admission you know I'm right and have chose not to play because of it

Assumptions and logic: Everyone including AIBs know Fury is in the general ballpark of 980 Ti. (Of course I acknowledge AIBs know the ballpark performance, and I have NEVER claimed otherwise). That means the last 10-20% of performance that depends on BIOS. If Fury and 980 Ti are within 10-20% performance, BIOS is the determinant factor. I believe this to be the case.

I think we are actually in agreement. Yes, final performance won't be known until those things are finalized, but the fact that they are still messing with clocks this late in the game is telling.....

 

[Azix]

cloudfire777 did you keep "leaking" information to build rep so you could start spreading rumours counter AMD?

No bios and use of the word ought. Well, we'll know soon enough. I am more interested in price since nvidia can't get a cent from me till they change culture and management. Not again. Can't shot myself in the foot supporting a company actively working to ruin PC gaming

 

[poofyhairguy]

Assumptions and logic: Everyone including AIBs know Fury is in the general ballpark of 980 Ti. .

Isn't that bad news in itself though? Doesn't AMD need to blow the doors off the 980 Ti OCs thanks to Gameworks? Like 20%+?

 

[Headfoot]

I think we are actually in agreement. Yes, final performance won't be known until those things are finalized, but the fact that they are still messing with clocks this late in the game is telling.....

Agreed.

Messing with clocks at this point probably means either 1) they're way behind, or 2) they're so close that messing with clocks/tdp/fan speed can make the difference.

 

[Udgnim]

We're just guessing, but I think it's something like this:

AMD was going to release Fury soon.

NV released 980ti first.

980ti was faster than anyone thought.

Fury was then not fast enough to release right after the 980ti.

Fury had to be reset, for release a little later with faster clocks.

initial rumor on 3/8/2015 was Fiji launch around Computex

http://www.vrworld.com/2015/03/08/amd-r9-300-series-said-launch-computex-2015/

AMD might still end up paper launching during Computex. who really knows

later rumor from 3 weeks ago, was for a 6/18/15 to 6/24/15 launch which makes sense considering AMD is going to be holding a PC gaming conference during E3

http://wccftech.com/amd-radeon-r9-300-series-fiji-xt-release-date/

rumor of GTX 980Ti release during Computex was reported 2 weeks ago

http://wccftech.com/nvidia-gtx-980-ti-gm200-310-maxwell-launch/

so assuming AMD actually launches Fiji around 6/18/15 to 6/24/15, then their launch time frame has more to do with their marketing plans during E3 than the GTX 980Ti getting released

 

[exar333]

Which is it cloudfire you create this thread saying its slower then you have another thread saying it will be priced at $850.

How can it be priced at $850 if they know it will be slower than a $649 980 Ti?

I think $849 sounded good before the 980Ti....

Also, I HIGHLY doubt BIOS clock tweaks will provide anywhere close to a 20% bump. Generally, you are talking 5-7% bumps. MAYBE 10% if they are being very aggressive, but over 10% would be a considerable change in the overall design. Remember that these clock adjustments need to work for all GPUs and they don't want to need to bump-up the default voltage too much to keep them all stable.

The issue above is exactly what has been happening to AMD on the APU front. They pushed clockspeed too much, we see throttling, and the default voltage is much higher for 70% of samples...

 

[exar333]

Agreed.

Messing with clocks at this point probably means either 1) they're way behind, or 2) they're so close that messing with clocks/tdp/fan speed can make the difference.

Good point.

If they are within 1-2%, it might be worth-it to sacrifice a few rpm to bump-up the speeds a bit.

if #1, then yeah, its bad...

 

[96Firebird]

I don't see how they can change clocks much from what they found during testing to be the optimal speed for their target parameters...

Wouldn't quality have to retest at the new clockspeeds to ensure stability over time? Or is it a one day process that the have a wide range of acceptable results?

 

[EightySix Four]

I don't really understand how this is a "BREAKING NEWS" thread. It's a rumor thread like the 4 other threads floating around about Fiji. Why can't we just wait until it launches before making a decision about how good it is?

 

[Noctifer616]

I don't see how they can change clocks much from what they found during testing to be the optimal speed for their target parameters...

Wouldn't quality have to retest at the new clockspeeds to ensure stability over time? Or is it a one day process that the have a wide range of acceptable results?

If the cards are supposed to be released this month then yea, I don't see them changing much considering they need to package and send the cards to retailers.

 

[cbrunny]

I don't really understand how this is a "BREAKING NEWS" thread. It's a rumor thread like the 4 other threads floating around about Fiji. Why can't we just wait until it launches before making a decision about how good it is?

Because that isn't fun.

 

[Skurge]

Agreed.

Messing with clocks at this point probably means either 1) they're way behind, or 2) they're so close that messing with clocks/tdp/fan speed can make the difference.

Or they just want to release a polished product. Launch it with freesync crossfire support, new drivers, or whatever else they need to make sure they have all their bases covered.

 

[Kenmitch]

I don't really understand how this is a "BREAKING NEWS" thread. It's a rumor thread like the 4 other threads floating around about Fiji. Why can't we just wait until it launches before making a decision about how good it is?

It's not. Wait and see guessing the truth will be told soon enough.

 

[cbrunny]

Or they just want to release a polished product. Launch it with freesync crossfire support, new drivers, or whatever else they need to make sure they have all their bases covered.

If that were the case (and it SHOULD be the case - that it is not is telling) I think we'd be seeing a ton of leaks right now about how great the performance is.

 

[exar333]

Well, in all fairness, this 'rumor' is at least information (allegedly) received from a AMD partner present at the conference. A lot more founded than random screenshots that have appeared over the previous months.....

In any event, looking forward to some concrete news.