Overclocked 290X temperatures, scaling revealed

[MrK6]

Highly doubt a card already at 94c will be allowed to have unlocked voltage. Towards the end of 7xxx series it was much tougher to find an unlocked card than not.

Those are unrelated. Most performance-oriented models have voltage options, while many AIB's elect to bypass voltage control hardware to increase their margins. What we're looking at here is AMD reference, which has included voltage control hardware since the 5xxx series if my memory serves me correctly.

We've seen enough leaks at this point to see the 290X is basically about Titan performance in general and seemingly a little faster than Titan at high resolutions.

All I want to see now is Mantle Battlefield 4 benchmarks. This card is great if you're shopping for a new high end card now, but if you already have one, I think it's only Mantle that can make this deliver appreciably more than what you are running currently.

It will depend on what AMD offers with the card (and if these benchmarks are accurate). If the R9 290X is a $600 part, that's a significant price/performance gain over the 780 that might turn heads. I passed on the 780 because it was barely faster than my overclocked 7970; being several months later, AMD is going to have to work even harder to impress me now. Now if Mantle turns out to be something substantial, that's something worth considering then.

 

[BallaTheFeared]

There's no vcore modding yet since the software isn't updated. These are OC with default vcore, getting above 1.1ghz is pretty good.

While that could be true, we have no idea what the stock volatge is for this card.

Is it 1v, is it 1.25v?

Saying stock voltage doesn't mean much if we don't know what that voltage is.

 

[3DVagabond]

Would have been nice to have found out all the info with a proper review instead of piecing together all the leaks. Amd really bungled this non launch.

If they don't have cards, or they are priced too high, or frame times are still borked, then AMD has bungled this. Nothing so far is bungled. Not exactly a thing of beauty, I'll grant you.

There is heaps of interest in this (both good and bad). I'm sure if they could have released it a month or 2 ago they would have (Back to School). Considering the plan was "nothing new this year", I'd say having new product for the Christmas buying season, instead of allowing nVidia to have free reign in the high performance market, is very good still.

I think the reason for all of the leaks is to play the market until they can get it out. If they had just said nothing up until now, people wouldn't be "holding out" for anything.

 

[Erenhardt]

While that could be true, we have no idea what the stock volatge is for this card.

Is it 1v, is it 1.25v?

Saying stock voltage doesn't mean much if we don't know what that voltage is.

Actually knowing that it is the stock voltage says more than the exact number of volts. If someone would say it is 1.1V, does it tell us something? Other than it is powered by 1.1V, nothing because it is whole new die, with whole new design with different transistor density. On the other hand, knowing that the chip worked at stock voltage without knowing the value, we know that it was operating with the voltage it was designed to work, and there should be some headroom left for overclockers.

 

[BallaTheFeared]

If they said 1.1v than I'd know they had roughly .1 to .2 volts to play with realistically before inferno mode.

We already know 1.35v was noted as the max, that's .5v higher than my cards and 1.3v is unrealistic for me 24/7 if I want to hear.

So yes, knowing the stock voltage would elude to possible OC headroom left with tweaking, it wouldn't be exact but it would be more informative than "stock voltage".

 

[3DVagabond]

Yes, if we're playing 30 second bursts with 10 second pauses between each area.

Sounds like you are talking about your posting style. That's unless you know more about the leaks/leaker than what's been posted?

 

[Erenhardt]

If they said 1.1v than I'd know they had roughly .1 to .2 volts to play with realistically before inferno mode.

We already know 1.35v was noted as the max, that's .5v higher than my cards and 1.3v is unrealistic for me 24/7 if I want to hear.

So yes, knowing the stock voltage would elude to possible OC headroom left with tweaking, it wouldn't be exact but it would be more informative than "stock voltage".

Are you comparing vegetables to fruits, again?

 

[hyrule4927]

Highly doubt a card already at 94c will be allowed to have unlocked voltage. Towards the end of 7xxx series it was much tougher to find an unlocked card than not.

94C in "quiet mode" while running Furmark . . .

Turn up the fan and the temperature goes down. And I don't know about you, but I don't regularly "play" Furmark.

 

[BallaTheFeared]

Sounds like you are talking about your posting style. That's unless you know more about the leaks/leaker than what's been posted?

I'm talking about what has been presented.

I can do 3dMark11 all day with 20% fan, be realistic.

 

[3DVagabond]

Highly doubt a card already at 94c will be allowed to have unlocked voltage. Towards the end of 7xxx series it was much tougher to find an unlocked card than not.

Now you are just trolling.

Member callouts are not allowed. -Moderator Shmee

 

[3DVagabond]

I'm talking about what has been presented.

I can do 3dMark11 all day with 20% fan, be realistic.

Be realistic? You should take your own advice.

 

[BallaTheFeared]

What am I saying that is unrealistic?

Point out the unrealistic for me.

3DMark11

~5 minutes of swtor. I turned it off because I started to get graphical glitches and I didn't feel like destroying my card proving an obvious point.

 

[3DVagabond]

What am I saying that is unrealistic?

Point out the unrealistic for me.

3DMark11

~5 minutes of swtor. I turned it off because I started to get graphical glitches and I didn't feel like destroying my card proving an obvious point.

You are taking something out of context and then making believe you have a point.

 

[Silverforce11]

Whats weirder is how you manage to load CF 7950s 98% in SWTOR... what the heck is that, 300 fps??

 

[BallaTheFeared]

No it's single, 100~ fps with RadeonPro/SweetFX.

You are taking something out of context and then making believe you have a point.

Which thread were you reading?

 

[Attic]

Awsome. Better than the $1k Titan which is a friggen' beast itself.

Lets get the thing in shops and see what Ryan has to say. I'm hoping no later than mid week?!

 

[Grooveriding]

Awsome. Better than the $1k Titan which is a friggen' beast itself.

Lets get the thing in shops and see what Ryan has to say. I'm hoping no later than mid week?!

Retailers already have stock, so my guess is they're coming out this week.

 

[Will Robinson]

Forget power use,temps and performance generally..its all about "Smoothness per watt"
or company financials and market share.()

 

[3DVagabond]

Which thread were you reading?

Maybe I missed where the leaker said he ran the benches the way you described them. I've read it though, and can't find it.

 

[BallaTheFeared]

Maybe I missed where the leaker said he ran the benches the way you described them. I've read it though, and can't find it.

The conversation is about what was said from what was shown, not was what said by the leaker.

You really didn't read the thread, did you?

 

[3DVagabond]

The conversation is about what was said from what was shown, not was what said by the leaker.

You really didn't read the thread, did you?

You said run 30 sec. with 10 sec. in between. Nobody else said that's how any of these tests were run. You set up some fantasy criteria to fit your agenda.

 

[BallaTheFeared]

So 3DMark11 is run another way?

Please explain that one.

 

[3DVagabond]

So 3DMark11 is run another way?

Please explain that one.

It wasn't simply 3D mark that was benched.

 

[BallaTheFeared]

It wasn't simply 3D mark that was benched.

No but it was what was being discussed, because the other extremely short tests, like Metro 2033 weren't properly reporting the values in GPUz. Most of the others were either pointless power virus results, or 100% fan speed yet again short run 3Dmark tests.

Are we clear now?

You done yet?

 

[3DVagabond]

No but it was what was being discussed, because the other extremely short tests, like Metro 2033 weren't properly reporting the values in GPUz. Most of the others were either pointless power virus results, or 100% fan speed yet again short run 3Dmark tests.

Are we clear now?

You done yet?

Why, are you incapable of carrying on a discussion? If you are tired, that's fine.