p.nb11 = (uint32_t)(dst->nb[1] / dst_tsize);
p.nb12 = (uint32_t)(dst->nb[2] / dst_tsize);
p.nb13 = (uint32_t)(dst->nb[3] / dst_tsize);
-
+
return p; // fastdiv values and offsets are initialized later in ggml_vk_op
}
}
static void ggml_vk_cpy(ggml_backend_vk_context * ctx, vk_context& subctx, const ggml_tensor * src0, ggml_tensor * dst, bool dryrun = false) {
- const uint32_t src0_type_size = ggml_type_size(src0->type);
- const uint32_t dst_type_size = ggml_type_size(dst->type);
-
uint32_t ne = (uint32_t)ggml_nelements(src0);
if (ggml_is_quantized(src0->type) && ggml_is_quantized(dst->type)) {
// Convert from number of logical elements to 2- or 4-byte units.
const float v10 = data_a[base + c1.y * p.nb01 + c0.x * p.nb00];
const float v11 = data_a[base + c1.y * p.nb01 + c1.x * p.nb00];
- return
+ return
v00 * (1.0-d.x) * (1.0-d.y) +
v01 * d.x * (1.0-d.y) +
v10 * (1.0-d.x) * d.y +
const vec2 d = c - c0f;
const ivec2 c0 = max(ivec2(c0f), 0);
const ivec2 c1 = min(ivec2(c0f + 1), ne0 - 1);
-
+
return fetch_bilinear(c0, c1, d, i12, i13);
}