Skip to content

armgit / janas

Public Imported from GitHub
0

37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cpp on all 13 models compared in Q4_K_M. LLMs bigger than your RAM, pure C, CPU or GPU.

638 B
Raw Blame
/* SPDX-License-Identifier: GPL-3.0-or-later
   Copyright (C) 2026 Maurizio Cammalleri */
#version 450
// vkcheck's zero-copy test: reads x and writes y, both in the host's own
// memory imported by the GPU (VK_EXT_external_memory_host): y = 3x + 1.
#extension GL_EXT_buffer_reference : require
#extension GL_EXT_buffer_reference_uvec2 : require

layout(local_size_x = 64) in;
layout(buffer_reference, std430, buffer_reference_align = 4) buffer U { uint v[]; };
layout(push_constant) uniform P { uvec2 x, y; uint n; } p;

void main()
{
    uint i = gl_GlobalInvocationID.x;
    if (i < p.n)
        U(p.y).v[i] = U(p.x).v[i] * 3u + 1u;
}