FP16_perf_check

import torch
import time

torch.backends.cudnn.benchmark = True


# 1a)
I, J, K = 64, 1024, 1024
A = torch.randn(I, J, device='cuda', dtype=torch.half)
B = torch.randn(J, K, device='cuda', dtype=torch.half)

# warumup
for _ in range(50):
    C = torch.matmul(A, B)
torch.cuda.synchronize()

nb_iters = 1000
torch.cuda.synchronize()
t0 = time.time()
for _ in range(nb_iters):
    C = torch.matmul(A, B)
torch.cuda.synchronize()
t1 = time.time()
print('1a) {:.3f}us per iteration)'.format((t1 - t0) / nb_iters * 1e6))

# 1b)
I, J, K = 1, 1024, 1024
A = torch.randn(I, J, device='cuda', dtype=torch.half)
B = torch.randn(J, K, device='cuda', dtype=torch.half)

# warumup
for _ in range(50):
    C = torch.matmul(A, B)
torch.cuda.synchronize()

nb_iters = 1000
torch.cuda.synchronize()
t0 = time.time()
for _ in range(nb_iters):
    C = torch.matmul(A, B)
torch.cuda.synchronize()
t1 = time.time()
print('1b) {:.3f}us per iteration'.format((t1 - t0) / nb_iters * 1e6))



# 2a)
I, J, K = 63, 1023, 1023
A = torch.randn(I, J, device='cuda', dtype=torch.half)
B = torch.randn(J, K, device='cuda', dtype=torch.half)

# warumup
for _ in range(50):
    C = torch.matmul(A, B)
torch.cuda.synchronize()

nb_iters = 1000
torch.cuda.synchronize()
t0 = time.time()
for _ in range(nb_iters):
    C = torch.matmul(A, B)
torch.cuda.synchronize()
t1 = time.time()
print('2a) {:.3f}us per iteration'.format((t1 - t0) / nb_iters * 1e6))

    
# 2b)
I, J, K = 1, 1023, 1023
A = torch.randn(I, J, device='cuda', dtype=torch.half)
B = torch.randn(J, K, device='cuda', dtype=torch.half)

# warumup
for _ in range(50):
    C = torch.matmul(A, B)
torch.cuda.synchronize()

nb_iters = 1000
torch.cuda.synchronize()
t0 = time.time()
for _ in range(nb_iters):
    C = torch.matmul(A, B)
torch.cuda.synchronize()
t1 = time.time()
print('2b) {:.3f}us per iteration'.format((t1 - t0) / nb_iters * 1e6))