> Actually, CSF stands in the way of re-allocating memory to other > contexts, because once we've allocated memory to a tiler heap, the FW > manages this pool of chunks, and recycles them. Mesa can intercept > the "returned chunks" and collect those chunks instead of re-assiging > then to the tiler heap through a CS instruction (which goes thought > the FW internallu), but that involves extra collaboration between the > UMD, KMD and FW which we don't have at the moment. Not saying never, > but I'd rather fix things gradually (first the blocking alloc in the > fence-signalling path, then the optimization to share the extra mem > reservation cost among contexts by returning the chunks to the global > kernel pool rather than directly to the heap). > > This approach should work fine with JM GPUs where the tiler heap is > entirely managed by the KMD though.
I really think CSF should be relying on the simple heuristics with incremental-rendering, unless you can prove that's actually a performance issue in practice. (On Imagination/Apple parts, it almost never is and we rely entirely on this approach. It's ok - it really is. For simple 2D workloads, the initial heap allocation is fine. For 3D scenes, we need very few frames to get the right size. this doesn't cause stutters in practice.) For JM .. yes, this discussion remains relevant of course.
