By using this site, you agree to our Privacy Policy and our Terms of Use. Close
starworld on 18 February 2014
Edit Thread
bonzobanana said:
Turkish said:
Not sure. gpu+cpu:

PS4: ~1.95/2TFlop (32ROPs, 1152 stream processors,
Xbone: 1.4TFlop (only 16 ROPs on gpu so 720p will be very common, 768 stream processors)
WiiU: estimates range from 300 to 600TFlop (8Rops, 320 stream processors)

There are as many ROPs and stream processors on the PS4 vs Xbone gap as the Xbone vs WiiU gap. The gap between PS4 and Xbone is ~500Gflop, Wii vs Xbone can be 800Gflop at minimum


wii u estimates start at 176 gflops and go up to 352 gflops, however the fact is the wii u gpu likely has a wii gpu integrated into it and this gpu may run at higher speeds in wii u mode to generate the gamepad screen when independent of the main screen. At best it probably adds 30 gflops performance to the main gflops performance of the wii u. The performance level we have seen from wii u makes it highly likely the performance is at the lower level 176gflops. However architecture differences means this will perform probably 30% plus minimum over the older architecture of the 360 gpu. Roughly speaking the wii u gpu is about the same performance level as 360/PS3 but has an enhanced feature set and has 32MB of ultra fast embedded memory. So its is bottleneck free almost. However still it performs below 360 and PS3 on most ocassions due to a slow CPU and low main memory bandwidth.

LIkely wii u architecture is something like a mobile Radion 6400M. Mobility architecture is extremely likely due to less heat issues even when fabricated at a low cost 40nm process as used in the wii u.

For the wii u to take so little power and produce so little heat that only needs a small fan it has to be a fairly low performance gpu. Not only that but you have to factor in the low bandwidth memory chips used which are less than 360 and PS3. It would be pointless fitting a higher performance gpu with such limited memory bandwidth. Lets not forget in the wii u that memory bandwidth has to be shared by both gpu and cpu. The 32MB of embedded memory certainly helps but its clear there would be huge restrictions on a more powerful gpu.

 

  • Engine clock speed: 480-800 MHz (wii u gpu is at about 550mhz so 176 gflops)
  • Processing power (single precision): 153-256 GigaFLOPS
  • Polygon throughput: 120-200M polygons/sec
  • Data fetch rate (32-bit): 15.36-25.6 billion fetches/sec
  • Texel fill rate (bilinear filtered): 3.84-6.4 Gigatexels/sec
  • Pixel fill rate: 1.92-3.2 Gigapixels/sec
  • Anti-aliased pixel fill rate: 7.68-12.8 Gigasamples/sec
  • Memory clock speed: 800-900 MHz GDDR5 or DDR3
  • Memory data rate: 3.2 Gbps GDDR5 or 1.6-1.8 Gbps DDR3
  • Memory bandwidth: 25.6 GB/sec (GDDR5) or 12.8-14.4 GB/sec (DDR3)

12.8GB/s is the DDR3 memory speed fitted to the wii u

  • TeraScale 2 Unified Processing Architecture
    • 160 Stream Processing Units
    • 8 Texture Units
    • 16 Z/Stencil ROP Units
    • 4 Color ROP Units
  • GDDR5/DDR3 memory interface
  • PCI Express 2.1 x16 bus interface
  • DirectX® 11 Evolved technology
    • DirectX® 11 support
      • Shader Model 5.0
      • DirectCompute 11
      • Programmable hardware tessellation unit
      • Accelerated multi-threading
      • HDR texture compression
      • Order-independent transparency
    • OpenGL 4.1 support
    • Image quality enhancement technology
      • Up to 12x multi-sample and super-sample anti-aliasing modes
      • Adaptive anti-aliasing
      • Morphological anti-aliasing (MLAA)
      • 16x angle independent anisotropic texture filtering
      • 128-bit floating point HDR rendering
  • AMD Eyefinity multi-display technology5
    • Support for up to 4 simultaneous displays
      • Independent resolutions, refresh rates, color controls, and video overlays
    • Display grouping
      • Combine multiple displays to behave like a single large display
  • AMD App Acceleration2
    • OpenCL™ 1.1 support6
    • DirectCompute 11
    • Accelerated video encoding, transcoding, and upscaling7,8,9
    • UVD 3 dedicated video playback accelerator
      • MPEG-4 AVC/H.264
      • VC-1
      • MPEG-2 (SD & HD)
      • Multi-View Codec (MVC)3
      • MPEG-4 part 2 (DivX®, Xvid)
      • Adobe® Flash®10
    • Enhanced video quality features
      • Advanced post-processing and scaling9
      • Dynamic contrast enhancement and color correction
      • Brighter whites processing (blue stretch)
      • Independent video gamma control
      • Dynamic video range control
    • Dual-stream 1080p playback support11
    • DXVA 1.0 & 2.0 support
  • AMD HD3D technology1
    • Stereoscopic 3D display/glasses support
    • Blu-ray 3D support3
    • Stereoscopic 3D gaming support
    • 3rd party Stereoscopic 3D middleware software support

the 176 gflops is confirmed, the project car dev/NFSWU dev flat out confirmed it.