mirror of
https://github.com/hanzoai/engine.git
synced 2026-11-20 14:17:52 +00:00
301 lines
9.7 KiB
Rust
301 lines
9.7 KiB
Rust
// Numeric gate for the IQ4_NL native ROCm decode (qmatvec_core<DW_IQ4_NL> in quant.hip). IQ4_NL is
|
|
// the 32-element nonlinear-codebook 4-bit type: y = d * KVALUES_IQ4NL[nibble] (the same 16-entry int8
|
|
// LUT IQ4_XS rides, but a single per-block f16 scale and a 32-elem block like Q4_0). The reference is
|
|
// the REAL CPU `BlockIQ4nl::to_float`, reinterpreting the exact GGML bytes the GPU reads. Codebook
|
|
// quants are integer-exact, so the only error is f16/bf16 activation rounding + f32 reorder; the gate
|
|
// requires nbad == 0 for the f16 AND bf16 matvec, single-expert and indexed-MoE. Lives in hanzo-cli so
|
|
// it links the HIP runtime via hanzo-cli's build.rs.
|
|
#![cfg(feature = "rocm")]
|
|
|
|
use half::{bf16, f16};
|
|
use hanzo_ml::backend::{BackendDevice, BackendStorage};
|
|
use hanzo_ml::quantized::k_quants::BlockIQ4nl;
|
|
use hanzo_ml::quantized::GgmlType;
|
|
use hanzo_ml::{RocmDevice, RocmQuantType, RocmStorage};
|
|
|
|
fn val(i: usize) -> f32 {
|
|
let x = ((i.wrapping_mul(2654435761) >> 8) & 0xffff) as f32 / 65535.0;
|
|
(x - 0.5) * 6.0
|
|
}
|
|
|
|
fn byte(seed: usize) -> u8 {
|
|
(seed.wrapping_mul(2654435761).wrapping_add(seed >> 7) & 0xff) as u8
|
|
}
|
|
|
|
fn to_f32(s: &RocmStorage) -> Vec<f32> {
|
|
match s.to_cpu_storage().expect("to cpu") {
|
|
hanzo_ml::CpuStorage::F16(v) => v.iter().map(|x| x.to_f32()).collect(),
|
|
hanzo_ml::CpuStorage::BF16(v) => v.iter().map(|x| x.to_f32()).collect(),
|
|
hanzo_ml::CpuStorage::F32(v) => v,
|
|
other => panic!("unexpected dtype {:?}", other.dtype()),
|
|
}
|
|
}
|
|
|
|
fn as_blocks<T: Clone>(bytes: &[u8]) -> Vec<T> {
|
|
let sz = std::mem::size_of::<T>();
|
|
assert_eq!(
|
|
bytes.len() % sz,
|
|
0,
|
|
"byte len {} not a multiple of {}",
|
|
bytes.len(),
|
|
sz
|
|
);
|
|
let n = bytes.len() / sz;
|
|
let mut v: Vec<T> = Vec::with_capacity(n);
|
|
unsafe {
|
|
std::ptr::copy_nonoverlapping(bytes.as_ptr(), v.as_mut_ptr() as *mut u8, n * sz);
|
|
v.set_len(n);
|
|
}
|
|
v
|
|
}
|
|
|
|
fn build_bytes<F: Fn(&mut [u8], usize, usize)>(
|
|
n: usize,
|
|
k: usize,
|
|
blk: usize,
|
|
tsz: usize,
|
|
fill: F,
|
|
) -> Vec<u8> {
|
|
assert_eq!(k % blk, 0);
|
|
let nblk = k / blk;
|
|
let mut out = vec![0u8; n * nblk * tsz];
|
|
for r in 0..n {
|
|
for b in 0..nblk {
|
|
let base = (r * nblk + b) * tsz;
|
|
let block = &mut out[base..base + tsz];
|
|
for (i, by) in block.iter_mut().enumerate() {
|
|
*by = byte(r * 7919 + b * 131 + i * 17 + 3);
|
|
}
|
|
fill(block, r, b);
|
|
}
|
|
}
|
|
out
|
|
}
|
|
|
|
// Small exact-f16 scale at byte `off` so the reference and kernel f16 scales agree to the last bit.
|
|
fn put_d(block: &mut [u8], off: usize, r: usize, b: usize, lo: f32, step: f32, m: usize) {
|
|
let d = lo + step * (((r + b) % m) as f32);
|
|
let bytes = f16::from_f32(d).to_le_bytes();
|
|
block[off] = bytes[0];
|
|
block[off + 1] = bytes[1];
|
|
}
|
|
|
|
fn reference_row<T: GgmlType>(
|
|
wq: &[u8],
|
|
nn: usize,
|
|
nblk: usize,
|
|
blk: usize,
|
|
tsz: usize,
|
|
x: &[f32],
|
|
) -> f32 {
|
|
let mut acc = 0f32;
|
|
let mut deq = vec![0f32; blk];
|
|
for b in 0..nblk {
|
|
let base = (nn * nblk + b) * tsz;
|
|
let blocks = as_blocks::<T>(&wq[base..base + tsz]);
|
|
T::to_float(&blocks, &mut deq);
|
|
let xb = &x[b * blk..b * blk + blk];
|
|
for j in 0..blk {
|
|
acc += deq[j] * xb[j];
|
|
}
|
|
}
|
|
acc
|
|
}
|
|
|
|
#[allow(clippy::too_many_arguments)]
|
|
fn check<T: GgmlType, F: Fn(&mut [u8], usize, usize)>(
|
|
dev: &RocmDevice,
|
|
log: &mut String,
|
|
name: &str,
|
|
qt: RocmQuantType,
|
|
n: usize,
|
|
k: usize,
|
|
blk: usize,
|
|
tsz: usize,
|
|
fill: F,
|
|
) {
|
|
assert_eq!(k % blk, 0, "{name}: k must be a multiple of {blk}");
|
|
let nblk = k / blk;
|
|
let xf: Vec<f32> = (0..k).map(val).collect();
|
|
let xh: Vec<f16> = xf.iter().map(|&v| f16::from_f32(v)).collect();
|
|
let xbf: Vec<bf16> = xf.iter().map(|&v| bf16::from_f32(v)).collect();
|
|
let xf_h: Vec<f32> = xh.iter().map(|v| v.to_f32()).collect();
|
|
let xf_b: Vec<f32> = xbf.iter().map(|v| v.to_f32()).collect();
|
|
let xst_h = dev.storage_from_slice(&xh).expect("upload x f16");
|
|
let xst_b = dev.storage_from_slice(&xbf).expect("upload x bf16");
|
|
|
|
let wq_bytes = build_bytes(n, k, blk, tsz, fill);
|
|
let wst = dev.storage_from_slice(&wq_bytes).expect("upload w");
|
|
|
|
let y_h = dev
|
|
.matvec_quant(qt, &wst, &xst_h, n, k)
|
|
.expect("matvec f16");
|
|
let y_b = dev
|
|
.matvec_quant(qt, &wst, &xst_b, n, k)
|
|
.expect("matvec bf16");
|
|
assert_eq!(
|
|
y_h.dtype(),
|
|
hanzo_ml::DType::F16,
|
|
"{name}: f16 matvec keeps f16"
|
|
);
|
|
assert_eq!(
|
|
y_b.dtype(),
|
|
hanzo_ml::DType::BF16,
|
|
"{name}: bf16 matvec keeps bf16"
|
|
);
|
|
let yh = to_f32(&y_h);
|
|
let yb = to_f32(&y_b);
|
|
|
|
let mut ref_h = vec![0f32; n];
|
|
let mut ref_b = vec![0f32; n];
|
|
for nn in 0..n {
|
|
ref_h[nn] = reference_row::<T>(&wq_bytes, nn, nblk, blk, tsz, &xf_h);
|
|
ref_b[nn] = reference_row::<T>(&wq_bytes, nn, nblk, blk, tsz, &xf_b);
|
|
}
|
|
let scale = ref_b
|
|
.iter()
|
|
.chain(ref_h.iter())
|
|
.fold(0f32, |m, &v| m.max(v.abs()))
|
|
.max(1.0);
|
|
let tol = 0.01 * scale;
|
|
let tol_b = 0.02 * scale;
|
|
let mut max_err_h = 0f32;
|
|
let mut max_err_b = 0f32;
|
|
let mut nbad = 0usize;
|
|
for i in 0..n {
|
|
let eh = (yh[i] - ref_h[i]).abs();
|
|
let eb = (yb[i] - ref_b[i]).abs();
|
|
max_err_h = max_err_h.max(eh);
|
|
max_err_b = max_err_b.max(eb);
|
|
if eh > tol || eb > tol_b {
|
|
nbad += 1;
|
|
}
|
|
}
|
|
log.push_str(&format!(
|
|
"{name:8} n={n} k={k} nbad={nbad}/{n} max_err_f16={max_err_h:.5} max_err_bf16={max_err_b:.5} (tol {tol:.5}) scale={scale:.3}\n"
|
|
));
|
|
assert!(nbad == 0, "{name} IQ4_NL decode mismatch: nbad={nbad} max_err_f16={max_err_h} max_err_bf16={max_err_b} tol={tol}");
|
|
}
|
|
|
|
#[allow(clippy::too_many_arguments)]
|
|
fn moe_check<T: GgmlType, F: Fn(&mut [u8], usize, usize)>(
|
|
dev: &RocmDevice,
|
|
log: &mut String,
|
|
name: &str,
|
|
qt: RocmQuantType,
|
|
e_cnt: usize,
|
|
nrows: usize,
|
|
n: usize,
|
|
k: usize,
|
|
blk: usize,
|
|
tsz: usize,
|
|
fill: F,
|
|
) {
|
|
let nblk = k / blk;
|
|
let expert_bytes = n * nblk * tsz;
|
|
let mut bank: Vec<u8> = Vec::with_capacity(e_cnt * expert_bytes);
|
|
for e in 0..e_cnt {
|
|
bank.extend_from_slice(&build_bytes(n, k, blk, tsz, |b, r, bb| {
|
|
fill(b, r * (e + 1) + e, bb)
|
|
}));
|
|
}
|
|
let wbank = dev.storage_from_slice(&bank).expect("upload bank");
|
|
let ids: Vec<u32> = (0..nrows).map(|s| ((s * 3 + 1) % e_cnt) as u32).collect();
|
|
let ids_dev = dev.storage_from_slice(&ids).expect("upload ids");
|
|
let xf: Vec<f32> = (0..nrows * k).map(val).collect();
|
|
let xh: Vec<f16> = xf.iter().map(|&v| f16::from_f32(v)).collect();
|
|
let xbf: Vec<bf16> = xf.iter().map(|&v| bf16::from_f32(v)).collect();
|
|
let xf_h: Vec<f32> = xh.iter().map(|v| v.to_f32()).collect();
|
|
let xf_b: Vec<f32> = xbf.iter().map(|v| v.to_f32()).collect();
|
|
let xst_h = dev.storage_from_slice(&xh).expect("upload x f16");
|
|
let xst_b = dev.storage_from_slice(&xbf).expect("upload x bf16");
|
|
|
|
let y_h = dev
|
|
.moe_matvec_quant(qt, &wbank, &xst_h, &ids_dev, nrows, n, k)
|
|
.expect("moe f16");
|
|
let y_b = dev
|
|
.moe_matvec_quant(qt, &wbank, &xst_b, &ids_dev, nrows, n, k)
|
|
.expect("moe bf16");
|
|
let yh = to_f32(&y_h);
|
|
let yb = to_f32(&y_b);
|
|
|
|
let mut nbad = 0usize;
|
|
let mut max_err_h = 0f32;
|
|
let mut max_err_b = 0f32;
|
|
let mut scale = 1f32;
|
|
for (s, &eid) in ids.iter().enumerate() {
|
|
let eb = &bank[eid as usize * expert_bytes..(eid as usize + 1) * expert_bytes];
|
|
let xrow_h = &xf_h[s * k..s * k + k];
|
|
let xrow_b = &xf_b[s * k..s * k + k];
|
|
for r in 0..n {
|
|
let rh = reference_row::<T>(eb, r, nblk, blk, tsz, xrow_h);
|
|
let rb = reference_row::<T>(eb, r, nblk, blk, tsz, xrow_b);
|
|
scale = scale.max(rh.abs()).max(rb.abs());
|
|
let eh = (yh[s * n + r] - rh).abs();
|
|
let ebb = (yb[s * n + r] - rb).abs();
|
|
max_err_h = max_err_h.max(eh);
|
|
max_err_b = max_err_b.max(ebb);
|
|
let tol = 0.02 * scale.max(1.0);
|
|
if eh > tol || ebb > tol {
|
|
nbad += 1;
|
|
}
|
|
}
|
|
}
|
|
log.push_str(&format!(
|
|
"{name:8} MoE E={e_cnt} nrows={nrows} n={n} k={k} nbad={nbad}/{} max_err_f16={max_err_h:.5} max_err_bf16={max_err_b:.5} scale={scale:.3}\n",
|
|
nrows * n
|
|
));
|
|
assert!(
|
|
nbad == 0,
|
|
"{name} IQ4_NL MoE mismatch: nbad={nbad} max_err_f16={max_err_h} max_err_bf16={max_err_b}"
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn qmatvec_iq4nl_numeric() {
|
|
let mut log = String::new();
|
|
let dev = RocmDevice::new(0).expect("rocm device");
|
|
// Single block per row, production decode, wide-k FFN, partial last warp-row (17), vocab-sized out.
|
|
let shapes: &[(usize, usize)] = &[
|
|
(64, 32),
|
|
(4096, 4096),
|
|
(1024, 3072),
|
|
(17, 4096),
|
|
(4096, 256),
|
|
];
|
|
// IQ4_NL: 18 B, 32 elems, d (f16) at byte 0. Small exact-f16 d keeps the +-127 codebook outputs
|
|
// inside f16 range (same reasoning as the IQ4_XS gate -- shared KVALUES_IQ4NL LUT).
|
|
for &(n, k) in shapes {
|
|
check::<BlockIQ4nl, _>(
|
|
&dev,
|
|
&mut log,
|
|
"IQ4_NL",
|
|
RocmQuantType::IQ4_NL,
|
|
n,
|
|
k,
|
|
32,
|
|
18,
|
|
|blk, r, b| {
|
|
put_d(blk, 0, r, b, 0.0078125, 0.00390625, 5);
|
|
},
|
|
);
|
|
}
|
|
// MoE: 8 experts, 16 routed slots, n=128 out, k=512 in (16 blocks of 32).
|
|
moe_check::<BlockIQ4nl, _>(
|
|
&dev,
|
|
&mut log,
|
|
"IQ4_NL",
|
|
RocmQuantType::IQ4_NL,
|
|
8,
|
|
16,
|
|
128,
|
|
512,
|
|
32,
|
|
18,
|
|
|blk, r, b| {
|
|
put_d(blk, 0, r, b, 0.0078125, 0.00390625, 5);
|
|
},
|
|
);
|
|
eprintln!("{log}");
|
|
}
|