Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [chiphell] RX480 in disguise ? (Update)
New Posts  All Forums:Forum Nav:

[chiphell] RX480 in disguise ? (Update) - Page 23

post #221 of 339
Thread Starter 
Updated pic



Ghetto mod is strong in this one...
Edited by BigTree - 6/12/16 at 5:47am
post #222 of 339
Quote:
Originally Posted by zealord View Post

Fury(X), R9 390(X), R9 290(X) etc. all have low clock speeds compared to Maxwell too.

AMD cards, atleast as of late, scale really well with higher clocks though. Going from 1000 to 1100 mhz is basically a 10% increase in performance, but 100mhz on a Pascal card is way less than 10%.

I'd have to find the post but again but someone was comparing overclock scaling and Hawaii was ~90-95% while Maxwell was ~55-60%
Atlas
(20 items)
 
M17x R4
(8 items)
 
 
CPUMotherboardGraphicsRAM
i7 3740qm @ 3.9Ghz M17x R4 7970m @ 1Ghz Corsair 1600mhz DDR3 4x4GB 
Hard DriveHard DriveOptical DriveOS
480gb seagate 600 SSD 500GB Seagate 2.5 Liton Slot-Load Blu-Ray Windows 10 x64 Tech Preview 
  hide details  
Reply
Atlas
(20 items)
 
M17x R4
(8 items)
 
 
CPUMotherboardGraphicsRAM
i7 3740qm @ 3.9Ghz M17x R4 7970m @ 1Ghz Corsair 1600mhz DDR3 4x4GB 
Hard DriveHard DriveOptical DriveOS
480gb seagate 600 SSD 500GB Seagate 2.5 Liton Slot-Load Blu-Ray Windows 10 x64 Tech Preview 
  hide details  
Reply
post #223 of 339
Quote:
Originally Posted by BigTree View Post

Updated pic



Ghetto mod is strong in this one...
No wonder about the temps lmao
post #224 of 339
Quote:
Originally Posted by guttheslayer View Post

Huh, isnt that score close to a GTX 1080 performance?

Quote:
Originally Posted by lolfail9001 View Post

Warning: Spoiler! (Click to show)
Nice 295x2
That is pretty much your stereotypical 295x2 score.

Stock standard 980ti lol. Just used cheat engine to search for the Nvidia 980ti string and replace it with 480X lol.
Never believe pictures of benchmarks,
post #225 of 339
Quote:
Originally Posted by bmgjet View Post


Stock standard 980ti lol. Just used cheat engine to search for the Nvidia 980ti string and replace it with 480X lol.
Never believe pictures of benchmarks,
It's not pictures anymore though, we got a video.

And the original thread got a removal request from AMD.
post #226 of 339
Quote:
Originally Posted by Drake87 View Post

Last I time I checked MSI didn't allow RMAs for the broken fans. I know what brand of card I'm not getting next.
Really? I know one person broke their ASUS DC2 fan because they didn't know how to mount CF and the card sag onto the lower one and broke the fan by hitting a pin on the back of the PCB biggrin.gif Warranty, blah blah, got a new card or same with replaced fan. I mean the fan was busted to pieces, not just won't spin, like shattered and an obvious mistake of his.
In EU under the 2 year warranty they would have to replace it IMHO, there is no way a fan dies on a GPU and they don't replace the fan or the card within warranty.

prznar1: clocks have nothing to do with 14nm or 300nm, it has to do with how much IPC there is and how the architecture is created and optimized. Nvidia pretty much made the next evolution of their architecture Maxwell it couldn't push it on 16nm so they released it on 28nm with higher clocks due to optimizations, now with Pascal they evolved it further and finally could release it on 16nm with further optimizations to push clocks by shortening the execution path. So Nvidia is pushing higher clocks with probable lower IPC. AMD has had previous GCN cards on sub 1GHz, they are said to increase the clock with Polaris but dunno how much especially how much because of optimizations to shorten the execution path.

As you can see from Maxwell vs Pascal IPC comparisons Pascal isn't that great.
Quote:
When we compare GM200 versus GP100 in clock-per-clock, Pascal (slightly) lags behind Maxwell.
They optimized Pascal for clock and reduced the work it does per clock (IPC). It's like with CPUs when they couldn't get more performance by architecture update they started raising the frequencies until it was not possible practically anymore.

A high IPC low clock can perform the same as low IPC high clock. It's just a different approach to solving a problem.
post #227 of 339
Thread Starter 
What is this?
post #228 of 339
post #229 of 339
Quote:
Originally Posted by zealord View Post

Fury(X), R9 390(X), R9 290(X) etc. all have low clock speeds compared to Maxwell too.

AMD cards, atleast as of late, scale really well with higher clocks though. Going from 1000 to 1100 mhz is basically a 10% increase in performance, but 100mhz on a Pascal card is way less than 10%.
noticed that. That is why im always saying that polaris is better architecture than pascal
post #230 of 339
Quote:
Originally Posted by BigTree View Post

What is this?

a picture of a reference R9 290/290X/390/390X board.
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 4770K @ 4.7ghz Gigabyte Z87X-UD4H EVGA GTX 980TI ACX 2.0  8GB G.Skill Trident X 2400 mhz 
Hard DriveHard DriveHard DriveCooling
Samsung 830 SSD 128GB 3TB Toshiba  3TB Toshiba Custom Water 
OSMonitorPowerCase
Window 7 Ultimate 64 bit HP 23xi Corsair HX750i Phanteks Enthoo Luxe 
Mouse
Logitech G600 
  hide details  
Reply
My System
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 4770K @ 4.7ghz Gigabyte Z87X-UD4H EVGA GTX 980TI ACX 2.0  8GB G.Skill Trident X 2400 mhz 
Hard DriveHard DriveHard DriveCooling
Samsung 830 SSD 128GB 3TB Toshiba  3TB Toshiba Custom Water 
OSMonitorPowerCase
Window 7 Ultimate 64 bit HP 23xi Corsair HX750i Phanteks Enthoo Luxe 
Mouse
Logitech G600 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [chiphell] RX480 in disguise ? (Update)