[QUOTE="Wickerman777"]Looking at AMD's site it seems like PS4's GPU is pretty similar to a 7970M (The M is important as their desktop GPUs are much more powerful than the mobile ones), just slightly weaker. Meanwhile the next Xbox's looks more like a 7870M but a little better. Basically subtract 2 compute units from the 7970M and you've got the PS4 GPU. Meanwhile add 2 compute units to the 7870M and you've got the Nextbox GPU. Not sure how it will translate to the screen but seems like a fairly large difference on paper.ronvalencia
From http://www.vgleaks.com/world-exclusive-orbis-unveiled/
There are no known GCN with 32 double precision/64bit ALUs per CU.
AMD Tahiti's CU has 64 single precision/32bit ALUs and effectively has 16 ALU double precision/64bit per CU i.e. ratio of 4:1. Even with 18 CUs, PS4's double precision/64bit compute is pretty strong.
7970M has 64 single precision/32bit ALUs and effectively has 4 ALU double precision/64bit per CU i.e. ratio of 16:1. http://www.amd.com/us/products/notebook/graphics/7000m/7900m/Pages/radeon-7900m-series.aspx#/2
For 64bit compute, it looks like Sony is attempting to avoid another repeat of gimped 32bit compute with NVIDIA RSX.
http://forum.beyond3d.com/showthread.php?t=57736&page=5
"I could go on for pages listing the types of things the spu's are used for to make up for the machines aging gpu, which may be 7 series NVidia but that's basically a tweaked 6 series NVidia for the most part. But I'll just type a few off the top of my head:"
...
10) Half floats
You can use full floats just fine on the 360 gpu. On the PS3 gpu they cause performance slowdowns. If you really need/have to use shaders with many full floats then you will want to move such functionality over to the spu's.
"Half floats" refers to 16bit floating point.
Lol, I don't even know what you're talking about. I looked at the tflops and compute modules and that's how I arrived at the conclusion I did.
Log in to comment