![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
PCGH - RX 480 fails at 1400MHz
- Thread starter Ieldra
- Start date
- Status
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,689
This card could have zero OC headroom and it will still be amazing value.
However if these rumors are true, then we have more "Overclocking doesn't matter guise" from the AMD camp.
lets be honest if they wanted the reference card to clock high they would add more cost and sell it to you for higher price. They made reference to make it better performanc/watt compared to their last gen, and gave you pretty basic cooler. Now AIBs are rumored to have higher frequency cards coming out. You can wait for those cards before you doom a reference card that wouldn't be able to cool it anyways. Its cheap thats all there is to it.
TaintedSquirrel
[H]F Junkie
- Joined
- Aug 5, 2013
- Messages
- 12,689
1400 MHz is a 10.5% overclock, that's worse than ref 290 and 290X and only slightly better than Fury X.
Squarely inside "disappointment" territory. Ref cards should be capable of at least 15%.
overclocking never matters, rx 480 best value ever, going nvidia = lighting cash under 300 dollars on fire.
There you go, so many quotes.
It's not as simple as that. You may say that but its not. Reports were partners didn't like the low msrp, but they are allowed to tweak this design as they want. I can pretty much bet aftermarket cards have modified bios to unlock voltage and they are binned better. What you are getting here is a card for 229 that is around gtx 980 and 390/390x performance at stock. Anything higher is better but what would be the point of AIB cards priced at 275-300 if you could get 15% extra performance from reference card. Makes no sense they want to make more profit not throw all AIBs under the bus just because AMD decided to price it at 199/229.
Why? Why at least 15%? Why not 10%? or 20%? Why is 15% the magic number?
Seriously, ref cards should provide good performance with good noise characteristics. That's it. We don't even know why OC fails.
Probably the best thing someone said. Just look at the cooler and design and it performs at the speeds amd rated it at and gets you around 100mhz overclock basically. You can't complain. Obvious they gimped it with a 6pin and the cooler to make AIBs happy and let them have a leg up on reference to probably promote those more as overclockers and get them little more profit. As we all know this is priced pretty damn cheap.
cageymaru
Fully [H]
- Joined
- Apr 10, 2003
- Messages
- 22,077
Is this the 1600+ model?
HD5770?Is this the 1600+ model?
Lame troll attempt
I don't remember any card in recent memory outside of the Fury X and probably GF100 where we didn't have at least 15% headroom for OC on stock voltage. Even GF100 managed close on the stock cooler, though.
So what? Rx 480 is a 200$-230$ card, not a 500$+ card. As long as the card is quiet, not a furnace, and not throttling - it's doing its job. That's it.
Maybe the reference 480 are made to hold their prescribed clocks, but indeed better cooling/more power is needed for better clocks?
Maybe AMD have decided to have the clocks as high as possible, within their desired power requirements.
Maybe that's just what the 480 cards are capable of.
People are just looking for things to take jabs against the Rx480. If it provides ~390x performance, out of the box, for 230$ - that's great. If it has shit OC, then it has shit OC.
I'll ask again - why is that true?
Are you like sleeping with an nvidia card next to your pillow? You seem to not understand the fact that its a fricking 230 dollar card not a enthusiast card. You are expecting too much cuz you simply are looking for it to fail.
It's a historical trend. People's expectations have been set to expect at least that much, due to past releases.
It doesn't even matter that it is a mid-range card, in recent history no reference card I can remember outside of Fury X, GTX 480 and maybe 290(x) had OC headroom less than 15% at stock voltage. If I remember wrong do correct me. That's 3 outliers total.
No I'm not expecting too much. I'm just extrapolating based on historical precedent. The 480 might end up as another outlier, alongside Fury X, GTX 480 and r9 290(x), then.
You are probably right but we can judge the OC when we have after market cooling and custom cards. Reference model doesn't seem to has the capability to cool heavily OC'd card. Looking at the cooler itself.
lolfail9001
[H]ard|Gawd
- Joined
- May 31, 2016
- Messages
- 1,496
And that's the problem, you did make the claim that the card will surpass 1400mhz on reference card. Yet now you do elaborate that voltage is locked (i've had second source on that, so i believe it) and need a voltage adjustment to pass 1400Mhz (we'll keep golde samples out of this for now).
I can see that one of these statements was unverified, and that's not the latter one. Your turn.
cageymaru
Fully [H]
- Joined
- Apr 10, 2003
- Messages
- 22,077
My R9 290 couldn't overclock at all until MSI Afterburner unlocked aux voltage and I put a better cooler on it. With a full custom water loop I can get to 1120. I have a launch reference card I believe since it has the original bugged bios on it still.
If they had a 490 ready in time, sure, but the 8ball says "Not Likely".
harmattan
Supreme [H]ardness
- Joined
- Feb 11, 2008
- Messages
- 5,129
Looks like the reference card from this latest leak and the latest 3dmark analysis at videocardz (the results are pouring in now apparently) we're looking at the following:
- under 1400mhz overclock on ref models
- around 390x performance at stock, just under Fury Pro OC'ed
- the speed to heat/power ratio gets very steep when OC'ed
What we've been seeing for the past week/month and matches with wjat Kyle has said on the clocks.
I've got a 1080 arriving tomorrow. Just can't see a single 480 being enough for me at 1440p and I'm sure as heck not going Crossfire again.
- under 1400mhz overclock on ref models
- around 390x performance at stock, just under Fury Pro OC'ed
- the speed to heat/power ratio gets very steep when OC'ed
What we've been seeing for the past week/month and matches with wjat Kyle has said on the clocks.
I've got a 1080 arriving tomorrow. Just can't see a single 480 being enough for me at 1440p and I'm sure as heck not going Crossfire again.
crazycrave
[H]ard|Gawd
- Joined
- Mar 31, 2016
- Messages
- 1,878
I don't see the issue many of you are bitching about because how do fix your company image ?
Well you start with building bullet proof cards that people can afford and if you think this ref RX480 is a bad buy at $199 because it can't overclock well ( your insane) , now this card is geared so it would run on your Antec 480 ECO Green watt power supply and offer performance of the 390x without the need to buy a($100 to $200) bigger power supply as something AMD had a bad rap for with Hawaii.
This is the 750Ti/950/960/970/980GTX killer in one card. Now you have to wait to see what the AIB's have in store for us as AMD had to keep the cost down to be able to offer it at that $199 price.
Well you start with building bullet proof cards that people can afford and if you think this ref RX480 is a bad buy at $199 because it can't overclock well ( your insane) , now this card is geared so it would run on your Antec 480 ECO Green watt power supply and offer performance of the 390x without the need to buy a($100 to $200) bigger power supply as something AMD had a bad rap for with Hawaii.
This is the 750Ti/950/960/970/980GTX killer in one card. Now you have to wait to see what the AIB's have in store for us as AMD had to keep the cost down to be able to offer it at that $199 price.
D
Deleted member 93354
Guest
Everything I have been claiming...
"AMD would push it near it's limits at launch with little in terms of overclock headroom"
"Overclocking with voltage increases lead to large thermal runaway"
doesn't matter too much what the AIB's have in store, their cards will cost more and use more wattage with limited clock rate rate to performance increases, If the leaks are really about 110-120 watts at stock, these are showing a small over clock is sapping up a ton of wattage. we are looking at a 5% overclock is 20% increase in wattage, what do you think will happen with AIB cards? 15% overclock with 60% increase in wattage? That is Fury Pro level performance at 225 watts. Nano is better to get than this at the end of you are looking to overclock AIB cards.
This card seems to be great for non overclockers. Now it looks clear why AMD did what it did.
D
Deleted member 93354
Guest
For the upteenth time, the majority of us are claiming it's a good buy for $200. But with the characterisitics it's putting out, it pretty much shows AMD screwed the pooch on the design target. Way too many transistors for the performance. Way too much heat for the performance.
harmattan
Supreme [H]ardness
- Joined
- Feb 11, 2008
- Messages
- 5,129
No kidding. I'm not sure who has been claiming 390x performance at much reduced power is bad for $200 - it's great, actually. The point that most of us are making is that it looks like there is steep power/heat needed to go above stock which points to the design issues we've been hearing about.
anything i've said stayed within the boundaries of what has already been leaked, as to not get in trouble. All I will say just because AMD included voltage support in their new driver doesn't mean it fully unlocks the cards potential. Being limited to 1.15 volts sucks.
Tup3x
[H]ard|Gawd
- Joined
- Jun 8, 2011
- Messages
- 1,942
IMO if the card doesn't overclock without touching the voltage it will not overclock even after touching the voltage. Increasing the voltage only helps tiny bit while causing massive increase in power usage and temperatures. Not worth it at all in my opinion.
Sounds like a better execution of the Nano situation.
Controlled thermals/power consumption at off-the-shelf speeds => should guarantee glowing reviews.
This card carries a positive narrative on price, performance, and power consumption - at stock clocks anyway
AIB models will address the noise levels/thermals when OCed, and hopefully not pricing significantly higher (still <$300 to meet Lisa's keynote).
ecmaster76
[H]ard|Gawd
- Joined
- Feb 6, 2007
- Messages
- 1,150
How is that a design issue?
Its like complaining that a naturally aspirated V8 has a lower red line than a turbo charged four cylinder engine.
If it has good price/performance and performance/watt for a $199-239 card AND AMD actually profits on the sales then they have met their design target.
If a part is guaranteed to OC 10-20% then they are just sandbagging. The old Intel Celerons that could be OC'd 50% only existed because Intel had no strong competition and wanted to gouge each market segment for all it was worth.
Way too much heat? It's 150w TDP Max.....
Tup3x
[H]ard|Gawd
- Joined
- Jun 8, 2011
- Messages
- 1,942
GTX 1070's TDP is 150W. If it shoots to ~150W with small overclock... I guess GF's 14nm process isn't as good as TSMC's.
Ieldra
I Promise to RTFM
- Joined
- Mar 28, 2016
- Messages
- 3,539
What?
Not the first finfet gpus... Yes we already know how they react to voltage increases, some people here care
I'm not an expert by any means, but that is just normal with finfets, its threshold voltage has a limited range.
So that puts a lot of strain on the design of the GPU when it comes to voltage.
D
Deleted member 93354
Guest
I meant for the performance based on the die size, especially when overclocked.
D
Deleted member 93354
Guest
It's the law of physics and how FinFETs work. Intel has improved on how the 3D transistor is built up so it isn't so much a pyramid it once was with Ivy Bridge, but that really didn't help them that much.
- Status