Tags

gpu ¹³ compiler ¹¹ spirv ⁹ mlir ⁶ vulkan ⁸ linalg ⁴ android ⁵ dialect ³ graphics ⁵ layout ⁵ triton ⁵ vectorization ³ compute ⁴ distribution ² dxc ⁴ fusion ² hlsl ⁴ ir ³ llvm ² ml ⁴ performance ⁴ shader ⁴ tiling ² vector ³ arith ² driver ³ edge ³ hierarchy ³ hoisting ² kernel ³ linear-layout ³ mali ³ math ² mobile ³ pattern ² resource ³ structure ² transformation ² unrolling ² abstraction ² adreno ² amd ² canonicalization ¹ cf ¹ cloud ² code ² codegen ² cuda ² gluon ² inference ² iree ² linux ² loop ¹ matrix ² memory ² memref ¹ mfma ² mhlo ¹ nvidia ² pipeline ² scf ¹ semantic ² standard ² synchronization ² tensor ¹ tf ¹ torch ¹ tosa ¹ agent ¹ amdgcn ¹ api ¹ architecture ¹ arm ¹ benchmarking ¹ blocked-layout ¹ buildkite ¹ business ¹ career ¹ challenge ¹ compatibility ¹ composition ¹ concept ¹ constraint ¹ conversion ¹ convolution ¹ correctness ¹ counter ¹ cpu ¹ dana ¹ debugging ¹ deployment ¹ descriptor ¹ design ¹ development ¹ device ¹ direct3d ¹ directx ¹ distributed-layout ¹ divide-left ¹ doc ¹ download ¹ ecosystem ¹ examples ¹ extensibility ¹ fragmentation ¹ github ¹ glsl ¹ google ¹ gpgpu ¹ heterogeneity ¹ host ¹ infrastructure ¹ interface ¹ internal ¹ inversion ¹ jit ¹ landscape ¹ location ¹ lowering ¹ metal ¹ mma-layout ¹ modularization ¹ opencl ¹ operation ¹ optimization ¹ oss ¹ padding ¹ pass ¹ philosophy ¹ portability ¹ product ¹ productivity ¹ pytorch ¹ requirement ¹ runtime ¹ shared-layout ¹ single-node ¹ software-pipelining ¹ summary ¹ swizzle ¹ swizzling ¹ syscall ¹ talk ¹ technology ¹ tips ¹ toolchain ¹ training ¹ tutorial ¹ type ¹ uvkcompute ¹ vm ¹ walkthrough ¹ wheel ¹