From 2ef940eb2abc35ba3a71cf3710dfccb8c00bf843 Mon Sep 17 00:00:00 2001
From: Damian Szwichtenberg <damian.szwichtenberg@intel.com>
Date: Thu, 12 Oct 2023 08:02:04 +0200
Subject: [PATCH] Add softmax benchmarks

---
 benchmark/ops/softmax.py | 69 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 69 insertions(+)
 create mode 100644 benchmark/ops/softmax.py

diff --git a/benchmark/ops/softmax.py b/benchmark/ops/softmax.py
new file mode 100644
index 000000000..acd974958
--- /dev/null
+++ b/benchmark/ops/softmax.py
@@ -0,0 +1,69 @@
+import argparse
+
+import torch
+import pyg_lib
+
+from time import perf_counter as timestamp
+from torch_geometric.utils import scatter, segment
+
+
+def softmax_reference(src, ptr, dim=0):
+    dim = dim + src.dim() if dim < 0 else dim
+    size = ([1] * dim) + [-1]
+    count = ptr[1:] - ptr[:-1]
+    ptr = ptr.view(size)
+    src_max = segment(src.detach(), ptr, reduce='max')
+    src_max = src_max.repeat_interleave(count, dim=dim)
+    out = (src - src_max).exp()
+    out_sum = segment(out, ptr, reduce='sum') + 1e-16
+    out_sum = out_sum.repeat_interleave(count, dim=dim)
+
+    return out / out_sum
+
+
+def measure_perf(impl_func, ptr, out_grad, num_warmups, num_steps, backward):
+    t_fwd = t_bwd = 0
+    for i in range(num_warmups + num_steps):
+        src = torch.randn(num_rows, num_heads)
+        src.requires_grad = backward
+
+        t_start = timestamp()
+        out = impl_func(src=src, ptr=ptr)
+        if i >= num_warmups:
+            t_fwd += timestamp() - t_start
+
+        if backward:
+            t_start = timestamp()
+            out.backward(out_grad)
+            if i >= num_warmups:
+                t_bwd += timestamp() - t_start
+
+    return t_fwd, t_bwd
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--backward', action='store_true')
+    parser.add_argument('--num-heads', type=int, default=4)
+    args = parser.parse_args()
+
+    num_rows, num_heads = 50000, args.num_heads
+    num_warmups, num_steps = 100, 500
+    group_size = 100
+
+    ptr = torch.arange(0, num_rows + 1, group_size)
+    out_grad = torch.randn(num_rows, num_heads)
+
+    func_args = [ptr, out_grad, num_warmups, num_steps, args.backward]
+
+    t_fwd, t_bwd = measure_perf(softmax_reference, *func_args)
+    print(f'Vanilla forward: {t_fwd:.4f}s')
+    if args.backward:
+        print(f'Vanilla backward: {t_bwd:.4f}s')
+    print('=========================')
+
+    t_fwd, t_bwd = measure_perf(pyg_lib.ops.softmax, *func_args)
+    print(f'pyg_lib forward:  {t_fwd:.4f}s')
+    if args.backward:
+        print(f'pyg_lib backward: {t_bwd:.4f}s')
+