Overclock.net › Forums › Industry News › Hardware News › [AMD] Polaris based RX 480 and "High performance solutions" announced at computex
New Posts  All Forums:Forum Nav:

[AMD] Polaris based RX 480 and "High performance solutions" announced at computex - Page 59

post #581 of 1394
Quote:
Originally Posted by Slaughterem View Post

If you listen real close to what Raja said (22 mins on edited video 37:45 on unedited) when describing this segment he referred to the dual 480 card He first said the dual 480 and then called it a dual 480 card not cards as in plural. Is this a Freudian slip? He might be describing a Duo 480 card I still believe there is a 2X 480 chip on a single interposer for less than $500.

This is what I am thinking. I don't know how they will get a 480X out of the specs of a single Rx480. The 480x may well be a dual version. How good would that be.
post #582 of 1394
Quote:
Originally Posted by T1A1 View Post

The benchmark is not run at different settings, per se...

The explanation for the way terrain is generated in AotS can be found in Oxide's GDC '16 presentation, specifically page 37-46-47
"Generating the stich map" and "Variable density" or in order not to render offscreen maps, they are using some tricks:

Variable density
"Big trick is to allow each tile to have a unique resolution
On the same screen, some tiles may be only 128x128 texels, while some could be 1024x1024
Terrain system will allocate more texel space to regions of terrain which are vertical, and need more space to avoid stretching artifacts"

Now you can stop blaming AMD - they are using identical settings on the benchmark

The only question that needs to be answered is how much variance can occur between runs.

Did they run the bench a bunch of times for each card and present the two runs with the biggest difference between them?

Or did they present the two runs that were closest to the average.

Nvidia/AMD would probably pick the scenario that presents them in the best light.
Upstairs Rig
(11 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus VI Hero evga 1080 Ti with Hybrid mod Corsair Vengeance Pro 2133 mhz 
Hard DriveHard DriveCoolingOS
Samsung 850 EVO 500gb WD Caviar Black Corsair h100i GTX Windows 8.1 64bit 
MonitorPowerCase
xb280hk EVGA Supernova 1000 G2 Corsair Carbide Air 540 
  hide details  
Reply
Upstairs Rig
(11 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus VI Hero evga 1080 Ti with Hybrid mod Corsair Vengeance Pro 2133 mhz 
Hard DriveHard DriveCoolingOS
Samsung 850 EVO 500gb WD Caviar Black Corsair h100i GTX Windows 8.1 64bit 
MonitorPowerCase
xb280hk EVGA Supernova 1000 G2 Corsair Carbide Air 540 
  hide details  
Reply
post #583 of 1394
Jeez this thread really devolved, eh? rolleyessmileyanim.gif

I'll be picking one up either way. I think it'll be a decent metric to judge a scale up to Vega, the only problem is HBM. Still, worthwhile. I also still think there's an unknown ~$300 part that we'll likely see at E3 which is probably where the official (paper) launch of the entire P10/11 line will happen.

Yesterday wasn't a launch and if people are saying it is, I guess I could see it from that angle, but it was a launch of a single card, the 480 4/8GB card. We have yet to see anything else official besides that single card.
Post-Fire Primary
(20 items)
 
  
CPUMotherboardGraphicsGraphics
FX-8350 Crosshair V Formula (-Z) Diamond R9 290x Diamond R9 290x 
GraphicsRAMHard DriveHard Drive
Sapphire R9 280x Radeon Gamer Series Toshiba HDD Seagate HDD 
Hard DriveCoolingCoolingCooling
Kingston Hyper-X SSD EKWB AMD Waterblock EKWB R9 290x Waterblock EKWB R9 290x Waterblock 
CoolingCoolingCoolingOS
EKWB Pump/Res EKWB 480mm PE Radiator Watercool Heatkiller MB Set Microsoft Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Corsair 1200W Thermaltake Core P5 Razer Ouroboros 
  hide details  
Reply
Post-Fire Primary
(20 items)
 
  
CPUMotherboardGraphicsGraphics
FX-8350 Crosshair V Formula (-Z) Diamond R9 290x Diamond R9 290x 
GraphicsRAMHard DriveHard Drive
Sapphire R9 280x Radeon Gamer Series Toshiba HDD Seagate HDD 
Hard DriveCoolingCoolingCooling
Kingston Hyper-X SSD EKWB AMD Waterblock EKWB R9 290x Waterblock EKWB R9 290x Waterblock 
CoolingCoolingCoolingOS
EKWB Pump/Res EKWB 480mm PE Radiator Watercool Heatkiller MB Set Microsoft Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Corsair 1200W Thermaltake Core P5 Razer Ouroboros 
  hide details  
Reply
post #584 of 1394
Quote:
Originally Posted by mcg75 View Post

The only question that needs to be answered is how much variance can occur between runs.

Did they run the bench a bunch of times for each card and present the two runs with the biggest difference between them?

Or did they present the two runs that were closest to the average.

Nvidia/AMD would probably pick the scenario that presents them in the best light.

I was thinking the same. To be honest, since I don't have AotS to test, I was looking for benchmarks on YouTube to see whether discrepancies can be found (still watching though).
Looking at the profiles of people that run the benchmark multiple times, there are small variations in FPS, unfortunately no detailed settings are shown and no definite conclusion can be made - but I think that what is happening - run several times and chose the best one.
Edited by C2H5OH - 6/1/16 at 4:48pm
post #585 of 1394
Quote:
Originally Posted by BulletBait View Post

Jeez this thread really devolved, eh? rolleyessmileyanim.gif

I'll be picking one up either way. I think it'll be a decent metric to judge a scale up to Vega, the only problem is HBM. Still, worthwhile. I also still think there's an unknown ~$300 part that we'll likely see at E3 which is probably where the official (paper) launch of the entire P10/11 line will happen.

Yesterday wasn't a launch and if people are saying it is, I guess I could see it from that angle, but it was a launch of a single card, the 480 4/8GB card. We have yet to see anything else official besides that single card.

Thats the million dollar question. Possibly:

  • RX480 4g @ $200
  • RX480 8g @ $250
  • RX480X 8g @ $300
R7 1700
(8 items)
 
  
CPUMotherboardGraphicsRAM
RYZEN 1700 ASROCK AB350 GAMING K4 ZOTAC 1080TI REFERENCE MODEL G SKILL 8 GIG (2X) 
Hard DriveOSPowerCase
WESTERN DIGITAL 250 M.2 2280 WINDOWS 10 64 CORSAIR SF600 LIAN LI PC-O10 
  hide details  
Reply
R7 1700
(8 items)
 
  
CPUMotherboardGraphicsRAM
RYZEN 1700 ASROCK AB350 GAMING K4 ZOTAC 1080TI REFERENCE MODEL G SKILL 8 GIG (2X) 
Hard DriveOSPowerCase
WESTERN DIGITAL 250 M.2 2280 WINDOWS 10 64 CORSAIR SF600 LIAN LI PC-O10 
  hide details  
Reply
post #586 of 1394
Quote:
Originally Posted by keikei View Post

Thats the million dollar question. Possibly:

  • RX480 4g @ $200
  • RX480 8g @ $250
  • RX480X 8g @ $300

Total DominATion from low end to high end.
the 1080 will feel the pressure from vega later due to the rest belongs to the 480.
Bold move from amd and the right one.
Thyslexia
(16 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 ga B350 gaming 3 Nitro sapphire 390 16gb corsair lpx 2933mhz 
Hard DriveOptical DriveCoolingOS
samsung 850 pro  sony water w10 64bit Pro 
MonitorKeyboardPowerCase
acer xg270hu 1440p 144hz! steelseries 6gv2 Corsair 750w Cm 840 
MouseMouse PadAudioAudio
corsair deathadder (G400 cable went bad) sennheiser HD 600 asus essence stx 
  hide details  
Reply
Thyslexia
(16 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1600 ga B350 gaming 3 Nitro sapphire 390 16gb corsair lpx 2933mhz 
Hard DriveOptical DriveCoolingOS
samsung 850 pro  sony water w10 64bit Pro 
MonitorKeyboardPowerCase
acer xg270hu 1440p 144hz! steelseries 6gv2 Corsair 750w Cm 840 
MouseMouse PadAudioAudio
corsair deathadder (G400 cable went bad) sennheiser HD 600 asus essence stx 
  hide details  
Reply
post #587 of 1394
Quote:
Originally Posted by WanWhiteWolf View Post

As far as I know, it doesn't work that way. Even in crossfire you - normally - use your GPUs to 100% of the GPU clock (at least according to afterburner info). Your performance is not double - compared to a single GPU - . The performance is usually 70-90 % extra considering one card. That's called scalling. Now....what does that 51% in the video mean is what people are wondering - some assuming - GPU clocks turned down - for both cards.

Ok, thanks for clearing this up for me. I don't deal with SLI/CF at all so never really learned about it. I think I get it now. Basically in a SLI/CF setup, 100% =/= the total load each GPU can assume, basically in a SLI/CF setup only a certain percent of normal GPU load is utilized between the 2 cards for use in rendering the game onto the display? Again, want to make sure I get it right.
Edited by Coydog - 6/1/16 at 4:53pm
Timberwolf
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-6500 Asus Hero VIII EVGA 650 TI Boost Corsair PC4 21300 DDR4 
Hard DriveHard DriveHard DriveCooling
Crucial MX200  Intel 320 Western Digital Caviar SE16 WD2500KS Cooler Master Hyper 212 Evo 
OSMonitorKeyboardPower
Windows 10 Pro Asus VS228 Generic EVGA G2 650 Watt 
CaseMouse
Phanteks Evolv ATX Logitech G502 
  hide details  
Reply
Timberwolf
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-6500 Asus Hero VIII EVGA 650 TI Boost Corsair PC4 21300 DDR4 
Hard DriveHard DriveHard DriveCooling
Crucial MX200  Intel 320 Western Digital Caviar SE16 WD2500KS Cooler Master Hyper 212 Evo 
OSMonitorKeyboardPower
Windows 10 Pro Asus VS228 Generic EVGA G2 650 Watt 
CaseMouse
Phanteks Evolv ATX Logitech G502 
  hide details  
Reply
post #588 of 1394
Quote:
Originally Posted by Slaughterem View Post

If you listen real close to what Raja said (22 mins on edited video 37:45 on unedited) when describing this segment he referred to the dual 480 card He first said the dual 480 and then called it a dual 480 card not cards as in plural. Is this a Freudian slip? He might be describing a Duo 480 card I still believe there is a 2X 480 chip on a single interposer for less than $500.

That would be really nice, but unlikely. Leaving asside the technical challenges, there were people in the audience who noticed shutter on the Polaris benchmark.

Shuttering , or better said , microshuttering, is a common problem for SLI/ CF setups but doesn't affect - from what I've read - the dual GPU setups. So it this were a dual GPU card , one would expect no shuttering (?). Then again, I never had a R929X or a ProDuo; maybe someone who has one can comment.
post #589 of 1394
Quote:
Originally Posted by BulletBait View Post

Jeez this thread really devolved, eh? rolleyessmileyanim.gif

I'll be picking one up either way. I think it'll be a decent metric to judge a scale up to Vega, the only problem is HBM. Still, worthwhile. I also still think there's an unknown ~$300 part that we'll likely see at E3 which is probably where the official (paper) launch of the entire P10/11 line will happen.

Yesterday wasn't a launch and if people are saying it is, I guess I could see it from that angle, but it was a launch of a single card, the 480 4/8GB card. We have yet to see anything else official besides that single card.

Honeslty, if it was a launch I think we would see LOT more info on the card as far as pricing (they still haven't released the 8GB version and I feel the 4GB and 8GB version would released at the same time).

Now finally looking at the dates for E3 I could really see them doing the paper launch there for more buzz etc.
Timberwolf
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-6500 Asus Hero VIII EVGA 650 TI Boost Corsair PC4 21300 DDR4 
Hard DriveHard DriveHard DriveCooling
Crucial MX200  Intel 320 Western Digital Caviar SE16 WD2500KS Cooler Master Hyper 212 Evo 
OSMonitorKeyboardPower
Windows 10 Pro Asus VS228 Generic EVGA G2 650 Watt 
CaseMouse
Phanteks Evolv ATX Logitech G502 
  hide details  
Reply
Timberwolf
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel i5-6500 Asus Hero VIII EVGA 650 TI Boost Corsair PC4 21300 DDR4 
Hard DriveHard DriveHard DriveCooling
Crucial MX200  Intel 320 Western Digital Caviar SE16 WD2500KS Cooler Master Hyper 212 Evo 
OSMonitorKeyboardPower
Windows 10 Pro Asus VS228 Generic EVGA G2 650 Watt 
CaseMouse
Phanteks Evolv ATX Logitech G502 
  hide details  
Reply
post #590 of 1394
Why would people expect tonga with memory compression to enhance performance closer to hawaii or even tahiti? They put in a smaller bus to save $$ ie good performance for less money. Now, if you compare 256bit bus to 256bit bus, one with compression, one without, that's where you should see a memory performance increase.

And iirc 1.1 just added true audio and updated powertune. 1.2 was an actual update to GCN and had the lossless color compression, but this new one is supposed to be a major revision.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [AMD] Polaris based RX 480 and "High performance solutions" announced at computex