diff --git a/docs/developer/am.md b/docs/developer/am.md index 67699422fb..30843618fd 100644 --- a/docs/developer/am.md +++ b/docs/developer/am.md @@ -1,6 +1,6 @@ # AM Driver -AM driver is a userspace driver targeting AMD's 7900XTX. You only need tinygrad to send compute tasks to your GPU! +AM driver is a userspace driver targeting AMD's RDNA3/RDNA4. You only need tinygrad to send compute tasks to your GPU! ## How to run? Make sure that amdgpu module is unloaded and just run tinygrad with `AMD=1`! diff --git a/docs/runtime.md b/docs/runtime.md index d109b156cf..b38824f336 100644 --- a/docs/runtime.md +++ b/docs/runtime.md @@ -5,7 +5,7 @@ tinygrad supports various runtimes, enabling your code to scale across a wide ra | Runtime | Description | Requirements | |---------|-------------|--------------| | [NV](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_nv.py) | Provides acceleration for NVIDIA GPUs | Ampere/Ada series GPUs | -| [AMD](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_amd.py) | Provides acceleration for AMD GPUs | RDNA2/RDNA3 series GPUs | +| [AMD](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_amd.py) | Provides acceleration for AMD GPUs | RDNA2/RDNA3/RDNA4 series GPUs | | [QCOM](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_qcom.py) | Provides acceleration for QCOM GPUs | 6xx series GPUs | | [METAL](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_metal.py) | Utilizes Metal for acceleration on Apple devices | M1+ Macs; Metal 3.0+ for `bfloat` support | | [CUDA](https://github.com/tinygrad/tinygrad/tree/master/tinygrad/runtime/ops_cuda.py) | Utilizes CUDA for acceleration on NVIDIA GPUs | NVIDIA GPU with CUDA support | diff --git a/tinygrad/runtime/support/am/ip.py b/tinygrad/runtime/support/am/ip.py index f72bd7ae93..583cf701ba 100644 --- a/tinygrad/runtime/support/am/ip.py +++ b/tinygrad/runtime/support/am/ip.py @@ -35,7 +35,7 @@ class AM_GMC(AM_IP): self.vm_base = self.adev.mm.va_allocator.base self.vm_end = self.vm_base + self.adev.mm.va_allocator.size - 1 - # GFX11 has 44-bit address space + # GFX11/GFX12 has 44-bit address space self.address_space_mask = (1 << 44) - 1 self.memscratch_paddr = self.adev.mm.palloc(0x1000, zero=not self.adev.partial_boot, boot=True)