New Xbox Scorpio Console Will Deliver "Really True 4K," Exec Says

  • 99 results
  • 1
  • 2
  • 3
Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#101  Edited By ronvalencia
Member since 2008 • 29612 Posts

@SecretPolice said:
@ronvalencia said:
@SecretPolice said:
@ronvalencia said:

Killer Instinct would be 4K/60 fps with PC's max details.

AMD Vega has double rate 16 bit FP hardware feature. https://www.extremetech.com/gaming/236011-new-rumors-claim-to-shed-light-on-amds-upcoming-vega-10-and-11-forecast-7nm-vega-20

Shader Model 6's upcoming hardware features.

Current NVIDIA/AMD PC GPUs doesn't have double rate 16 FP feature.

GTX 1060/1070/1080/Titan X Pascal are already obsolete on SM6's native 16 bit FP feature.

R9-285/R9-380X/Fury X/Fury Pro/RX-480/RX-470/RX-470D has single rate native 16 bit FP feature.

When Captain Ron says it, I believe. Mighty Scorpio together with the new and improved StormCloud. :o

Freakin Crackdown 3 baby, gonna be nothing like it. Nothin!!! :P

https://devtalk.nvidia.com/default/topic/934562/cuda-programming-and-performance/nvidia-pascal-geforce-gtx-1080-amp-gtx-1070/post/4889687/

"There is no fast fp16 in GP104."

Precisely as I thought. Great minds think alike. :o

I kid, I kid, lololol I'm fairly clueless to your high tech talk and I mean that in a good way. :) I just want a powerhouse console this time around.

Nvidia wasn't aware of MS's SM6.0 16bit floating point feature and it was too late in the final GP104 design (tape out).

SM6.0 = Shader Model 6.0 is the shader language version 6.0

GP104 = GeForce GTX 1070, GeForce GTX 1080 models

Due to Xbox projects, AMD was in sync with Microsoft.

https://forum.beyond3d.com/threads/nvidia-pascal-announcement.57763/page-61#post-1918478

"GP100 has lots of FP16 units (for deep learning training). GP10x does not.

Conversely, GP10x has lots of Int8 dot product units (for deep learning inference). GP100 does not.

The conspiracy theories about throttling are wrong. The chips are just different, with a different balance of functional unit."

At GTC 2016, Jen-Hsun was talking about GPU Inference Engine(GIE) & Tesla M4.

RX-480...

Int8 = quad rate over 32 bit FP

Int16 = single rate over 32 bit FP

Fp16 = single rate over 32 bit FP (Ready for SM6.0)

PS4 Pro (and Vega)

Int8 = quad rate over 32 bit FP

Int16 = ??? rate over 32 bit FP

Fp16 = double rate over 32 bit FP. For Vega, SM6.0's fp16 processing is faster than RX-480.