File size: 2,250 Bytes
f98af2f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
<!doctype html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width,initial-scale=1">
  <title>PyC Kernel Suite — Latest Results</title>
  <link rel="stylesheet" href="./assets/css/site.css">
</head>
<body>
  <main class="container">
    <header class="site-header"><nav class="site-nav" aria-label="Primary">
      <a href="./index.html">Main</a><a href="./results-insights.html">Results Insights</a><a class="active" href="./kernel-results.html">Latest Kernel Results</a><a href="./inference/index.html">Inference Portal</a>
    </nav></header>
    <section class="hero"><p class="hero-kicker">Published April 21, 2026</p><h1>Latest kernel-suite results</h1>
      <p class="hero-copy">Standalone Hopper GEMM/kernel measurements from the latest campaign. These are not inference-serving results and must not be compared directly with the inference portal.</p>
      <p class="section-text"><strong>Comparison rule:</strong> compare rows only within the same shape, precision, architecture, and measurement lane. TFLOP/s is the kernel-reported best timing, not end-to-end application throughput.</p>
      <p><a href="./results/kernel-suite-latest.json">Machine-readable manifest</a></p>
    </section>
    <section class="flow"><h2>Measured kernels</h2><div class="table-wrap"><table><thead><tr><th>Kernel</th><th>Arch</th><th>Shape</th><th>Precision</th><th>Best ms</th><th>TFLOP/s</th></tr></thead><tbody id="rows"><tr><td colspan="6">Loading…</td></tr></tbody></table></div></section>
    <section class="flow"><h2>Latest profiles</h2><div class="chart-grid" id="plots"></div></section>
  </main>
  <script>
    fetch('./results/kernel-suite-latest.json',{cache:'no-store'}).then(r=>r.json()).then(data=>{
      document.getElementById('rows').innerHTML=data.rows.map(row=>`<tr><td>${row.kernel}</td><td>${row.arch||'—'}</td><td>${row.shape||'—'}</td><td>${row.precision||'—'}</td><td>${row.best_ms??'—'}</td><td>${row.tflops??'—'}</td></tr>`).join('');
      document.getElementById('plots').innerHTML=(data.plots||[]).map(p=>`<figure><img src="${p.path}" alt="${p.label}" loading="lazy"><figcaption>${p.label}</figcaption></figure>`).join('');
    });
  </script>
</body>
</html>