geordiemp
Member
"This continuous improvement in AMD technology means it's dangerous to rely on teraflops as an absolute indicator of performance.
And CU count should be avoided as well. "
I'm sure that both TF and CU count being handwaved and 'avoided as an absolute indicator of performance' has nothing to do with the fact that they're both at a higher number in the XSX.
PR stunt. Accept it for what it is.
It isn't any different than Spencer saying they're not interested in selling more XSXs than Sony will sell PS5s because he knows the PS5 will utterly destroy the XSX in sales and tries to downplay that fact.
Its not a PR stunt by Cerny, TF is a the theoretical max number and only applies if the technologies doing the shading are the same.
They are not the same ps5 is different to standard RDNA2 and XSX as eluded to by many.....
Simplified GPU is in 2 stages, vertices shader and then the FP32 data is sent to local caches and then processed by pixel shader. This creates bottlenecks and slowdown of the wavefronts between the 2 parts. Ps5 patent by Cerny and Naughty dog compresses the data between the vertex and pixel parts of the GPU.
It overcomes this

So, faster caches due to 2.23 Ghz, Cache scrubbers, GPU compression between two major function blocks...
Dont be so sure ps5 shaders is not XSX shaders. How do games run, can ps5 run games look good in 4k....remember its only 15 % difference and you have DX12 abstract apis, faster ps5 caches and faster data between GPU blocks.
Not so straight forward is it ?
Last edited: