fix sytax

.github/workflows/short-tests.yml

@@ -40,7 +40,7 @@ jobs:
       - uses: actions/checkout@v4
       - name: Run Simulation
         run: echo "skipped SASS-Simulation. Will perform in merge queue"
-  SASS-Simulation:
+  mail-test:
     runs-on: tgrogers-raid
     defaults:
       run:

runner.py

@@ -0,0 +1,60 @@
+#!/usr/bin/env python3
+from jwt import JWT, jwk_from_pem
+import time
+import requests
+
+pem = "aalp-self-hosted-runner.2024-06-21.private-key.pem"
+client_id = "Iv23lixzRYO8eKFZQCxt"
+installation_id = "52081584"  # Replace with your actual installation ID
+
+# Open PEM
+with open(pem, 'rb') as pem_file:
+    signing_key = jwk_from_pem(pem_file.read())
+
+payload = {
+    # Issued at time
+    'iat': int(time.time()),
+    # JWT expiration time (in sec, 10 minutes maximum)
+    'exp': int(time.time()) + 60,
+    # GitHub App's client ID
+    'iss': client_id
+}
+
+# Create JWT
+jwt_instance = JWT()
+jwt = jwt_instance.encode(payload, signing_key, alg='RS256')
+
+# Set the URL for the request
+url = f"https://api.github.com/app/installations/{installation_id}/access_tokens"
+
+# Set the headers
+headers = {
+    "Accept": "application/vnd.github+json",
+    "Authorization": f"Bearer {jwt}",
+    "X-GitHub-Api-Version": "2022-11-28"
+}
+
+# Send the POST request
+response = requests.post(url, headers=headers)
+
+# Get the "token" field from the response
+token = response.json().get("token")
+
+# get token for creating the runners
+org = "accel-sim"  # Replace with your organization name
+
+# Set the URL for the request
+url = f"https://api.github.com/orgs/{org}/actions/runners/registration-token"
+
+# Set the headers
+headers = {
+    "Accept": "application/vnd.github+json",
+    "Authorization": f"Bearer {token}",
+    "X-GitHub-Api-Version": "2022-11-28"
+}
+
+# Send the POST request
+response = requests.post(url, headers=headers)
+
+token = response.json().get("token")
+print(token)

test.cu

@@ -0,0 +1,52 @@
+#include <iostream>
+#include <cuda_runtime.h>
+
+__global__ void testDependency(int *start, int *results, int n)
+{
+  int index = threadIdx.x + blockIdx.x * blockDim.x;
+
+  int dependency_var = 0;
+
+  // # pragma unroll
+  for (int i = index; i < n; ++i)
+  {
+    // if (start[i] < n && i == start[i])
+    // {
+      dependency_var = start[i];
+    // }
+  }
+  results[index] = dependency_var;
+}
+
+int main()
+{
+  unsigned n = 1024*1024;
+  int results[n];
+
+  int *d_results;
+  cudaMalloc(&d_results, n * sizeof(int));
+
+  int *start, *d_start;
+  cudaMalloc(&d_start, n * sizeof(int));
+  cudaMallocHost(&start, n * sizeof(int));
+
+  for (int i = 0; i < n; ++i)
+  {
+    start[i] = i;
+  }
+
+  cudaMemcpy(d_start, start, n * sizeof(unsigned), cudaMemcpyHostToDevice);
+
+  testDependency<<<1, 1>>>(d_start, d_results, n);
+
+  cudaMemcpy(results, d_results, n * sizeof(int), cudaMemcpyDeviceToHost);
+
+  cudaFree(d_results);
+
+  for (int i = 0; i < n / 1024; ++i)
+  {
+    std::cout << "Result[" << i << "]: " << results[i] << std::endl;
+  }
+
+  return 0;
+}