From 3a3598cc247afc9074c4cb36b43873e915e741e5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Johannes=20G=C3=A4=C3=9Fler?= Date: Thu, 1 Oct 2026 21:52:40 +0200 Subject: [PATCH] llama: refer to segment documentation [no ci] (llama/29074) --- ggml/include/ggml-backend.h | 1 + 1 file changed, 1 insertion(+) diff --git a/ggml/include/ggml-backend.h b/ggml/include/ggml-backend.h index cc3f8cd36..2a0957436 100644 --- a/ggml/include/ggml-backend.h +++ b/ggml/include/ggml-backend.h @@ -383,6 +383,7 @@ extern "C" { // - most tensors have n_segments == 1 and a contiguous slice of the tensor data // - some tensors have an inhomogenenous data layout along the split axis, // those tensors are divided into segments which are each individually split across devices + // (this usually happens when multiple tensors are fused into a single one) // - ne has one entry per segment and device and that segment repeats nr times, // in total when accounting for repetitions the segments add up to ggml_tensor::ne for that axis, // the outer/inner loops are over segments/devices like [seg0_dev0_r0, seg0_dev1_r0, seg0_dev0_r1, seg0_dev1_r1, seg1_dev0_r0, seg1_dev1_r0],