[gpu/NVDRV] Finalize, improve AllocObjCtx #333

SDK-Chan · 2025-08-27T15:29:51+02:00

SDK-Chan commented

2025-08-27 15:29:51 +02:00

Improves object allocation per channel, only allowing max amount of 6 objects contexts per channel.
Previously objects were stored in a heap allocated vector which is sub-optimal for performance reasons.
The new implementation instead uses a stack based array with a O(1) approach.
This should boost performance in games which heavily rely on object context creation.

Improves object allocation per channel, only allowing max amount of 6 objects contexts per channel. Previously objects were stored in a heap allocated vector which is sub-optimal for performance reasons. The new implementation instead uses a stack based array with a O(1) approach. This should boost performance in games which heavily rely on object context creation.

crueter approved these changes 2025-08-28 01:33:54 +02:00

Dismissed

SDK-Chan dismissed crueter's review 2025-08-28 17:24:16 +02:00

Reason:

New commits pushed, approval review dismissed automatically according to repository settings

Lizzie reviewed 2025-08-28 21:57:30 +02:00

src/core/hle/service/nvdrv/devices/nvhost_gpu.cpp Outdated

					
				@ -132,6 +132,7 @@ void nvhost_gpu::OnOpen(NvCore::SessionId session_id, DeviceFD fd) {

				void nvhost_gpu::OnClose(DeviceFD fd) {

				    sessions.erase(fd);

				    std::scoped_lock lk(channel_mutex);