Add an im2col based benchmark. (#800)

* Add an im2col based benchmark.

* Reshape the final result.
This commit is contained in:
Laurent Mazare
2023-09-10 16:56:28 +01:00
committed by GitHub
parent 3dd5804299
commit 559944146f

View File

@ -6,7 +6,7 @@ extern crate intel_mkl_src;
extern crate accelerate_src; extern crate accelerate_src;
use candle::quantized::GgmlType; use candle::quantized::GgmlType;
use candle::{Device, Result, Tensor, D}; use candle::{CpuStorage, Device, Layout, Result, Shape, Tensor, D};
use clap::{Parser, Subcommand}; use clap::{Parser, Subcommand};
trait Benchmark { trait Benchmark {
@ -19,6 +19,48 @@ trait Benchmark {
const ITERS: usize; const ITERS: usize;
} }
struct Im2Col(usize, usize);
impl candle::CustomOp1 for Im2Col {
fn name(&self) -> &'static str {
"im2col"
}
fn cpu_fwd(&self, storage: &CpuStorage, layout: &Layout) -> Result<(CpuStorage, Shape)> {
let &Self(h_k, w_k) = self;
let (b, c, h, w) = layout.shape().dims4()?;
let (h_out, w_out) = (h - h_k + 1, w - w_k + 1);
let slice = storage.as_slice::<f32>()?;
let src = match layout.contiguous_offsets() {
None => candle::bail!("input has to be contiguous"),
Some((o1, o2)) => &slice[o1..o2],
};
let mut dst = vec![0f32; b * h_out * w_out * c * h_k * w_k];
let (s_b, s_c, s_h) = (c * h * w, h * w, w);
for b_idx in 0..b {
let src_idx = b_idx * s_b;
let dst_idx = b_idx * h_out * w_out * c * h_k * w_k;
for h_idx in 0..h_out {
let dst_idx = dst_idx + h_idx * w_out * c * h_k * w_k;
for w_idx in 0..w_out {
let dst_idx = dst_idx + w_idx * c * h_k * w_k;
for c_idx in 0..c {
let dst_idx = dst_idx + c_idx * h_k * w_k;
let src_idx = c_idx * s_c + src_idx;
for h_k_idx in 0..h_k {
let src_idx = src_idx + (h_idx + h_k_idx) * s_h + w_idx;
let dst_idx = dst_idx + h_k_idx * w_k;
dst[dst_idx..dst_idx + w_k]
.copy_from_slice(&src[src_idx..src_idx + w_k])
}
}
}
}
}
let storage = candle::WithDType::to_cpu_storage_owned(dst);
Ok((storage, (b * h_out * w_out, c * h_k * w_k).into()))
}
}
// Conv1d example as used in whisper. // Conv1d example as used in whisper.
struct Conv1d; struct Conv1d;
impl Benchmark for Conv1d { impl Benchmark for Conv1d {
@ -53,7 +95,32 @@ impl Benchmark for Conv2d {
d.0.conv2d(&d.1, 0, 1, 1, 1) d.0.conv2d(&d.1, 0, 1, 1, 1)
} }
const ITERS: usize = 1; const ITERS: usize = 5;
}
// Conv2d example as used in stable-diffusion, im2col implementation.
struct Conv2dIm2Col;
impl Benchmark for Conv2dIm2Col {
type PreProcessData = (Tensor, Tensor);
type RunResult = Tensor;
fn preprocess() -> Result<Self::PreProcessData> {
let inp = Tensor::randn(0f32, 1., (2, 320, 96, 96), &Device::Cpu)?;
let w = Tensor::randn(0f32, 1., (320, 320, 3, 3), &Device::Cpu)?;
Ok((inp, w))
}
fn run_one(d: &Self::PreProcessData) -> Result<Self::RunResult> {
// d.0.conv2d(&d.1, 0, 1, 1, 1)
let (b, _, h, w) = d.0.dims4()?;
let (h_k, w_k) = (3, 3);
let (h_out, w_out) = (h - h_k + 1, w - w_k + 1);
let col = d.0.apply_op1_no_bwd(&Im2Col(h_k, w_k))?;
let res = col.matmul(&d.1.flatten_from(1)?.t()?)?;
res.reshape((b, (), h_out, w_out))
}
const ITERS: usize = 5;
} }
struct Matmul; struct Matmul;
@ -145,6 +212,7 @@ fn run<B: Benchmark>(iters: Option<usize>) -> Result<()> {
enum Task { enum Task {
Conv1d, Conv1d,
Conv2d, Conv2d,
Conv2dIm2Col,
Matmul, Matmul,
Qmatmul, Qmatmul,
Softmax, Softmax,
@ -167,6 +235,7 @@ fn main() -> Result<()> {
match args.task { match args.task {
Task::Conv1d => run::<Conv1d>(args.iters)?, Task::Conv1d => run::<Conv1d>(args.iters)?,
Task::Conv2d => run::<Conv2d>(args.iters)?, Task::Conv2d => run::<Conv2d>(args.iters)?,
Task::Conv2dIm2Col => run::<Conv2dIm2Col>(args.iters)?,
Task::Matmul => run::<Matmul>(args.iters)?, Task::Matmul => run::<Matmul>(args.iters)?,
Task::Softmax => run::<Softmax>(args.iters)?, Task::Softmax => run::<Softmax>(args.iters)?,
Task::SoftmaxLastDim => run::<SoftmaxLastDim>(args.iters)?, Task::SoftmaxLastDim => run::<SoftmaxLastDim>(args.iters)?,