-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathresearch.html
More file actions
100 lines (96 loc) · 5.69 KB
/
Copy pathresearch.html
File metadata and controls
100 lines (96 loc) · 5.69 KB
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
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>Research - Tanish Samir Desai</title>
<meta name="description" content="Research work and publications" />
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;600;700&display=swap" rel="stylesheet">
<link rel="stylesheet" href="styles.css" />
<link rel="icon" href="assets/favicon.ico" />
</head>
<body>
<main class="container">
<a href="index.html" class="back-button">Back to Home</a>
<header class="header">
<h1 class="name">Tanish Samir Desai</h1>
<p class="contact">
<a href="mailto:tanishdesai37@gmail.com">tanishdesai37@gmail.com</a>
<span class="sep">|</span>
Vadodara, India
</p>
<p class="links">
<a id="link-linkedin" href="#" target="_blank" rel="noopener">LinkedIn</a>
<span class="sep">|</span>
<a id="link-github" href="#" target="_blank" rel="noopener">GitHub</a>
<span class="sep">|</span>
<a id="link-scholar" href="#" target="_blank" rel="noopener">Google Scholar</a>
<span class="sep">|</span>
<a id="link-resume" href="#" target="_blank" rel="noopener">Resume</a>
</p>
</header>
<section>
<h2>Research</h2>
<div class="item">
<div class="item-head">
<div>
<span class="item-title">GPU Power Cap Optimization (HPDC '25 – Best Poster Award)</span>
<a href="https://dl.acm.org/doi/abs/10.1145/3731545.3735119" class="publication-button" target="_blank" rel="noopener">View Publication</a>
<span class="item-sub">ACM HPDC '25 Symposium, Notre Dame, USA</span>
</div>
<div class="item-meta">2025</div>
</div>
<ul class="bullets">
<li>Developed a machine learning–driven framework that dynamically adjusts GPU power caps to maximize performance-per-watt within constrained energy budgets.</li>
<li>Conducted extensive evaluation across 20+ diverse workloads from Rodinia and Parboil benchmark suites, achieving up to 18% higher computational throughput compared to NVIDIA's baseline power management strategies, while consuming equivalent average power.</li>
<li>Demonstrated scalability benefits in multi-GPU cluster environments, where our approach yielded 15–20% energy savings per compute node, translating to substantial cost reductions at enterprise HPC data center scales.</li>
<li>Recognized with the Best Poster Award at HPDC '25, underscoring the critical intersection of artificial intelligence and sustainable computing practices in high-performance environments.</li>
</ul>
</div>
<div class="item">
<div class="item-head">
<div>
<span class="item-title">GPU Power Consumption Prediction (ICTAI '24)</span>
<a href="https://ieeexplore.ieee.org/abstract/document/10849425/" class="publication-button" target="_blank" rel="noopener">View Publication</a>
<span class="item-sub">36th IEEE ICTAI</span>
</div>
<div class="item-meta">2024</div>
</div>
<ul class="bullets">
<li>Created a novel static analysis–based prediction system that estimates GPU power consumption without requiring expensive hardware instrumentation or runtime profiling.</li>
<li>Conducted comprehensive comparative analysis of multiple machine learning paradigms—including gradient boosting, deep neural networks, and transfer learning techniques—across six distinct CUDA benchmark collections.</li>
<li>Enabled rapid power profiling by transforming traditionally time-intensive measurement workflows (hours) into near-instantaneous static analysis (seconds), empowering both kernel optimization researchers and infrastructure teams managing energy efficiency metrics.</li>
</ul>
</div>
<div class="item">
<div class="item-head">
<div>
<span class="item-title">GPU Temperature Prediction (HiPC '24)</span>
<a href="https://ieeexplore.ieee.org/abstract/document/10898724/" class="publication-button" target="_blank" rel="noopener">View Publication</a>
<span class="item-sub">31st IEEE HiPC</span>
</div>
<div class="item-meta">2024</div>
</div>
<ul class="bullets">
<li>Engineered a sophisticated LSTM-based neural architecture capable of accurately forecasting GPU thermal states by modeling intricate temporal dependencies and workload evolution patterns.</li>
<li>Leveraged recurrent learning capabilities to capture dynamic runtime characteristics, outperforming traditional static prediction approaches through adaptive, context-aware thermal forecasting.</li>
<li>Delivered practical value in operational settings by enabling preemptive thermal intervention strategies, which mitigate performance throttling events, enhance system reliability margins, and prolong hardware service lifetimes in large-scale data center and HPC deployment environments.</li>
</ul>
</div>
<p class="links">
<a id="link-scholar-inline" href="#" target="_blank" rel="noopener">View on Google Scholar</a>
</p>
</section>
<footer class="footer">
<p>
<a id="link-resume-inline" href="#" target="_blank" rel="noopener">Download my resume</a>
<span class="sep">|</span>
© <span id="year"></span> Tanish Samir Desai
</p>
</footer>
</main>
<script src="script.js"></script>
</body>
</html>