Recent content by Love_In_Rio

FWIW:

It seems so, in Tom Henderson's leaked docs extended info article: In addition: 30 WGPs running specialised BVH8 traversal shaders vs 18 WGPs running BVH4 tranversal shaders on the standard PlayStation 5.

AMD so far has no traversal hardware, only intersection resolve units. Traversal is software based and thats why is more cache sensitive than in Nvidia solution, but thats where supposely infinity cache helps for RT. AMD RDNA 3.5 ray tracing unit with its intersection resolve unit can resolve if...

RDNA 3.5 is BVH4x1, like 7700XT. PS4 pro has double the processing capability of ray intersects with BVH8x2 (8 BVH boxes and 2 triangles intersection resolve per cycle) per RT unit. So RDNA 4 at least. In the leaked doc only is mentioned the word custom (fully) for the ML hardware and its 300 tops.

Who designed Tempest engine?. Is a CU without cache and with an DMA controller added based on the SPU design, made with AMD blocks with a configuration asked from Sony. But Cernys team is also capable of design low level things on their own and ask AMD to make them, the Kraken decompressor was...

They are getting 2 cells inside with 5 integer units per SPU instead of 4 floating point ones+1 integer one. ;) Well, seriously, they could have modified several CUs in the way they did for the DMA tempest engine (that is 35% faster in PS5 PRO, so must be clocked higher that the main GPU mode)...

Going from a RT unit capable of BVH4 to BVH8 plus going from 36 CUs to 60 explains the 2x-3x improvement and allows for wider BVH structures with less cache stress.

Your lamp in the maze does not cast any shadows, Blade of darkness with a 486 and a TNT Riva casted shadows from torchs from every object everywhere.

Independently it seems all PS5s have a certain degree of coil whine:

That´s in the dashboard or with a game running?. Doesn´t sound like coil whine. In fact it doesn´t sound too strong. If something i would say it could have the famous sticker in the fan.

And here we have a clear example of Heisenberg principle. You can´t measure it without modifying its behaviour.

It abruptly disappears going the home menu right?.

There is not secret sauce because Cerny told us all. There is a great custom coherency and cache system and there is a geometry engine that will make polygons sing.

We are lucky this gen coincided with the best AMD architecture since R300. Getting 200 watts of Nvidia like post-Maxwell efficiency is something nobody would have believed a year ago.

It seems to be a 512MB chip, the other 512 will be in the southbridge, for a 1GB total, like in the PS4 PRO.