amd-identity

Deadline

19 days 20 hours (2025-06-08 00:00 UTC)

Language

Python

GPU Type

MI300

Description

This task is purely for testing the submission system. There will be *no* points. > Input: (input_tensor, output_tensor) > - input_tensor: Input data > - output_tensor: Pre-allocated empty tensor of the same shape as `input_tensor` > Output: Should return `output_tensor` after it has been filled with the values from `input_tensor`.`

Reference Implementation

import torch
from task import input_t, output_t
from utils import make_match_reference


def generate_input(size: int, seed: int) -> input_t:
    gen = torch.Generator(device='cuda')
    gen.manual_seed(seed)
    data = torch.empty(size, device='cuda', dtype=torch.float16)
    data.uniform_(0, 1, generator=gen)
    return data, torch.empty_like(data)


def ref_kernel(data: input_t) -> output_t:
    input, output = data
    output[...] = input
    return output


check_implementation = make_match_reference(ref_kernel)

Rankings

MI300

__seal 🥇 5.502μs submission.py
tendazeal 🥈 6.791μs   +1.289μs submission.py
gau.nernst 🥉 6.807μs   +0.016μs submission.py
hatoo 7.708μs   +0.901μs amd-identity.py
D++ 18.728μs   +11.020μs solution.py
chess 19.694μs   +0.966μs identity.py
Erik S. 19.755μs   +0.061μs submission.py
mdda123 20.181μs   +0.426μs submission-hip.py
beetle0315 20.234μs   +0.053μs submission.py
_hui_xu 20.415μs   +0.181μs submission.py
siro 20.856μs   +0.441μs submission.py
Jerry Chiu 21.866μs   +1.010μs submission.py
zhubenzhu 21.871μs   +0.006μs submission.py
blurbird 21.939μs   +0.068μs amd-identity.py
phileasfogg2197 22.241μs   +0.302μs submission.py
gbsvf 22.282μs   +0.041μs amd-identity.py
mooglevich 22.370μs   +0.089μs submission.py
bobmarleybiceps 22.541μs   +0.170μs submission.py
Hoyoun Jung 22.589μs   +0.048μs submission.py
Kareem 22.591μs   +0.003μs submission.py
Shivam 22.659μs   +0.067μs identity_template.py
cudawarped 22.971μs   +0.312μs submission.py
jayce0098 23.142μs   +0.172μs submission.py
Rik - OCI 23.706μs   +0.564μs submission.py
Austin Liu 23.948μs   +0.242μs submission.py
syzyzygy 24.207μs   +0.259μs amd_identity.py