-
Notifications
You must be signed in to change notification settings - Fork 0
/
cn.html
361 lines (337 loc) · 19.7 KB
/
cn.html
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
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>TACC | iSING Lab HKUST | Scalable AI Infrastructure Designed for Evolving Machine Learning Research</title>
<link rel="stylesheet" href="./assets/bulma.min.css">
<link href="./assets/bulma-divider.min.css" rel="stylesheet">
<script src="./assets/jquery-3.6.0.min.js"></script>
<link href="./assets/style.css" rel="stylesheet">
<script src="./assets/chart.js"></script>
<script src="./assets/chartjs-adapter-date-fns.bundle.min.js"></script>
<script src="./assets/papaparse.min.js"></script>
<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=IBM+Plex+Sans:ital,wght@0,100;0,200;0,300;0,400;0,500;0,600;0,700;1,100;1,200;1,300;1,400;1,500;1,600;1,700&display=swap" rel="stylesheet">
</head>
<body>
<div class="menubar">
<div class="container">
<div class="columns is-multiline is-desktop is-align-items-center is-justify-content-center">
<div class="column is-narrow logo has-text-centered"><img src="./assets/tacc-logo.png"></div>
<div class="column has-text-centered">
<div class="columns is-mobile is-align-items-center items is-multiline">
<div class="column is-narrow item"><a href="#overview">概述</a></div>
<div class="column is-narrow item"><a href="#research">研究</a></div>
<div class="column is-narrow item"><a href="#hkust">@香港科技大学</a></div>
<div class="column is-narrow item"><a href="#collab">合作</a></div>
</div>
</div>
</div>
</div>
</div>
<div class="section banner">
<div class="container">
<div class="columns is-desktop">
<div class="column banner-text">
<p><span class="text-colored">可扩展的人工智能基础设施</span> <br> 为不断发展的 <br>机器学习研究而设计</p>
<p class="banner-subtitle">利用图灵人工智能计算云(TACC)的力量</p>
<p>
<a class="tacc-button" href="#hkust">申请访问权限</a>
<a class="tacc-button button-secondary" href="https://github.com/turingaicloud/quickstart" target="_blank">在GitHub上关注我们</a>
</p>
</div>
<div class="column banner-image"></div>
</div>
</div>
</div>
<div class="section overview" id="overview">
<div class="container">
<h2 class="section-title">TACC 是一个<span class="text-underlined text-colored">高效的集群管理解决方案</span>,针对<span class="spacing"><br> </span>大规模GPU集群中的机器学习应用进行了优化</h2>
<div class="columns">
<p class="column"></p>
<p class="column mt-4">TACC 支持广泛的机器学习应用研究,具有高性能和可扩展的基础架构,同时在软件和硬件层面上提供支持。TACC 可根据机器学习应用的工作流程进行定制,为您提供更高效的计算集群管理、部署和扩展计算密集型机器学习任务的流程。</p>
</div>
</div>
</div>
<div class="section research" id="research">
<div class="container">
<div class="columns is-multiline is-mobile">
<div class="column is-3-tablet is-6-mobile">
<img src="assets/tech-1.svg">
<p class="item-title"">统一界面</p>
<p>统一界面处理机器学习集群的用户工作配置文件,并提供任务监视、输出检索和日志流的工具。它可以从本地环境提交任务,并在调度程序提交之前包含环境配置脚本。</p>
</div>
<div class="column is-3-tablet is-6-mobile">
<img src="assets/tech-2.svg">
<p class="item-title"">ML 系统</p>
<p>ML 框架使用最先进的并行化和分布式训练技术优化模型开发。这些策略改善了数据处理和算法执行,实现了对复杂计算和大型数据集的高效处理。</p>
</div>
<div class="column is-3-tablet is-6-mobile">
<img src="assets/tech-3.svg">
<p class="item-title"">集群调度器</p>
<p>集群调度器通过分析任务完成时间、资源利用率和效率等因素来优化机器学习集群中的资源分配。这个关键组件提高了集群的效率和吞吐量。</p>
</div>
<div class="column is-3-tablet is-6-mobile">
<img src="assets/tech-4.svg">
<p class="item-title"">以人工智能为中心的网络</p>
<p>我们的研究通过有效管理大型模型的传输并利用智能网卡进行计算卸载来增强AI工作负载。这种设置优化了数据流并减少了延迟,确保了AI应用的高性能和可扩展性。</p>
</div>
</div>
</div>
</div>
<div class="section publications">
<div class="container">
<div class="section-title">学术研究论文</div>
<div class="publication-entry">
<div class="item-title">TACC:用于机器学习任务的全栈云计算基础设施</div>
<div class="item-author">Kaiqiang Xu, Xinchen Wan, Hao Wang, Zhenghang Ren, Xudong Liao, Decang Sun, Chaoliang Zeng, Kai Chen</div>
</div>
<div class="publication-entry">
<div class="item-title">面向分布式DNN训练的领域特定网络传输(NSDI '24)</div>
<div class="item-author">Hao Wang, Han Tian, Jingrong Chen, Xinchen Wan, Jiacheng Xia, Gaoxiong Zeng, Wei Bai, Junchen Jiang, Yong Wang, Kai Chen</div>
</div>
<div class="publication-entry">
<p class="item-button"><a href="https://www.cse.ust.hk/~kaichen/publication.html" target="_blank" class="tacc-button button-secondary">查看更多研究论文</a></p>
</div>
</div>
</div>
<div class="section cluster" id="hkust">
<div class="container">
<h2 class="section-title"><span class="is-hidden-desktop"><br></span>香港科技大学的 TACC 集群 <img class="hkust-logo" src="assets/UST.svg"> <span id="stats-date" class="is-size-6 "></span></h2>
<div class="columns is-align-items-center">
<div class="column is-half">
<div class="columns is-multiline is-mobile">
<div class="column is-6">
<p class="text-large">2</p>
<p class="text-small"">TACC管理的集群</p>
</div>
<div class="column is-6">
<div class="text-large" id="stats-user">397</div>
<p class="text-small"">自2021/05以来活跃的TACC用户</p>
</div>
<div class="column is-6">
<p class="text-large" id="stats-task" >24,303</p>
<p class="text-small"">TACC集群上的任务处理</p>
</div>
<div class="column is-6">
<p class="text-large" id="stats-gpu">235,739</p>
<p class="text-small"">用于ML任务的GPU小时数</p>
</div>
</div>
</div>
<div class="column text-center px-6">
<canvas id="userChart" width="400" height="200"></canvas>
</div>
</div>
<div class="columns is-align-items-center">
<div class="column">
<div class="columns is-multiline is-tablet">
<div class="column is-12">
<div class="cards cards-main">
<p class="item-title">招募先锋</p>
<p class="item-subtitle">早期采用者申请现已开放!立即加入,推动您的AI研究</p>
<p class="item-button"><a class="tacc-button" href="https://forms.gle/eHzg8mqMHmEziU7r8" target="_blank">申请免费访问TACC@HKUST</a></p>
</div>
</div>
<div class="column is-6">
<a href="https://github.com/turingaicloud/quickstart" target="_blank">
<div class="cards">
<p class="item-title">
文档和教程
<svg style="vertical-align:middle" focusable="false" preserveAspectRatio="xMidYMid meet" fill="currentColor" width="16" height="16" viewBox="0 0 32 32" role="img" xmlns="http://www.w3.org/2000/svg"><path d="M10 6L10 8 22.59 8 6 24.59 7.41 26 24 9.41 24 22 26 22 26 6 10 6z"></path></svg>
</p>
<p class="item-subtitle">提供全面的文档和示例,帮助您入门</p>
</div>
</a>
</div>
<div class="column is-6">
<a href="mailto:tacc-contact@lists.ust.hk">
<div class="cards">
<p class="item-title">
联系TACC团队
<svg style="vertical-align:middle" focusable="false" preserveAspectRatio="xMidYMid meet" fill="currentColor" width="16" height="16" viewBox="0 0 32 32" role="img" xmlns="http://www.w3.org/2000/svg"><path d="M10 6L10 8 22.59 8 6 24.59 7.41 26 24 9.41 24 22 26 22 26 6 10 6z"></path></svg>
</p>
<p class="item-subtitle">与香港科技大学的TACC运营团队联系,报告问题或提供反馈</p>
</div>
</a>
</div>
</div>
</div>
<div class="column p-6 description">
<p><strong>香港科技大学的TACC</strong></p>
<p>在香港科技大学,TACC管理着超过160张GPU卡的集群,用于机器学习的研究和教育,并向研究社区开放。 </p>
<p>与2023年初相比,2023年的TACC活跃用户增加了84%,达到397人,处理的ML任务增加了115%,达到24,303个。</p>
<p>TACC支持超过40个研究项目,并在SIGMOD、KDD、CVPR和UbiComp等顶级会议上共收到了22次引用。</p>
</div>
</div>
</div>
</div>
<div class="section join" id="collab">
<div class="container">
<div class="columns is-align-items-center">
<div class="column is-half">
<p class="section-title">来自研究者,为了研究者</p>
<p class="section-subtitle">
在您的集群中部署TACC,一种先进的集群管理和任务处理方案,提高效率和可靠性。
我们基于研究成果的解决方案包括全面的硬件监控、简便的维护以及高效的任务调度和执行。
</p>
</div>
<div class="column is-half picture"></div>
</div>
<div class="columns is-multiline is-desktop">
<div class="column">
<div class="cards cards-main">
<p class="item-title">高度 <span class="text-colored">稳定性</span></p>
<p class="item-subtitle">TACC通过提供持续的24/7硬件监控、强大的容器化选项以及公平和高效的任务调度机制,提高了操作可靠性。</p>
</div>
</div>
<div class="column">
<div class="cards cards-main">
<p class="item-title">增强的 <span class="text-colored">可用性</span></p>
<p class="item-subtitle">使用TACC,任务的提供和管理变得轻松。用户可以通过命令行、Web UI或API提交和监视任务,使系统易于访问。</p>
</div>
</div>
<div class="column">
<div class="cards cards-main">
<p class="item-title">最大化的 <span class="text-colored">性能</span></p>
<p class="item-subtitle">TACC集成了来自学术界的最先进的ML系统和网络技术,旨在最大化AI应用的性能和可扩展性。</p>
</div>
</div>
</div>
<div class="columns is-multiline is-desktop">
<div class="column">
<a href="mailto:tacc-contact@lists.ust.hk?subject=Deploying TACC in my own cluster">
<div class="tacc-button button-large">
在您自己的集群中部署TACC
<svg focusable="false" preserveAspectRatio="xMidYMid meet" fill="currentColor" width="16" height="16" viewBox="0 0 16 16" role="img" class="cds--btn__icon" xmlns="http://www.w3.org/2000/svg"><path d="M9.3 3.7L13.1 7.5 1 7.5 1 8.5 13.1 8.5 9.3 12.3 10 13 15 8 10 3z"></path></svg>
</div>
</a>
</div>
<div class="column">
<a href="https://arxiv.org/abs/2110.01556">
<div class="tacc-button button-large">
阅读我们的经验论文
<svg focusable="false" preserveAspectRatio="xMidYMid meet" fill="currentColor" width="16" height="16" viewBox="0 0 16 16" role="img" class="cds--btn__icon" xmlns="http://www.w3.org/2000/svg"><path d="M9.3 3.7L13.1 7.5 1 7.5 1 8.5 13.1 8.5 9.3 12.3 10 13 15 8 10 3z"></path></svg>
</div>
</a>
</div>
</div>
</div>
</div>
<div class="section footer">
<div class="container">
<div class="columns is-align-items-center">
<div class="logo column is-narrow"><img src="assets/tacc-logo-white.png" /></div>
<div class="column item is-narrow"><a href="mailto:tacc-contact@lists.ust.hk">联系我们</a></div>
<div class="column item is-narrow"><a href="https://ising.cse.ust.hk" target="_blank">关于iSING实验室</a></div>
<div class="column item is-narrow"><a href="https://tacc.ust.hk/v1/trs.html" target="_blank">由RGC主题研究资助</a></div>
<div class="column item is-narrow"><a href="https://tacc.ust.hk/v1/tos.html" target="_blank">服务条款(TACC@HKUST)</a></div>
</div>
</div>
</div>
<script>
function createOption() {
return {
type: 'line',
options: {
plugins: {
legend: {
display: false
},
title: {
display: true,
text: 'Chart Title', // Replace with your title
font: {
size: 14,
family: 'IBM Plex Sans'
}
}
},
scales: {
x: {
type: 'time',
time: {
unit: 'month',
tooltipFormat: 'yyyy-MM',
displayFormats: {
month: 'yy-MM'
}
},
grid: {
display: true,
drawBorder: true,
drawOnChartArea: true,
color: function(context) {
if (context.tick && context.tick.label.endsWith('-01')) {
return '#999';
} else {
return '#fff'; // Light gray for other grid lines
}
}
}
},
B: {
type: 'linear',
position: 'right',
grid: {display: false},
}
},
interaction: {
mode: 'nearest',
intersect: false,
axis: 'x'
}
}
};
}
function setupChart(data) {
var dates = data.map(item => item.Date);
var userNumbers = data.map(item => item.Users);
var taskNumbers = data.map(item => item.Tasks);
var Color1 = '#452d69'; // Nvidia Green
var Color2 = '#76b900'; // Nvidia Dark
var ctxUser = document.getElementById('userChart').getContext('2d');
var userOptions = createOption();
userOptions.data = {
labels: dates,
datasets: [{
label: 'User Growth',
yAxisID: 'A',
data: userNumbers,
backgroundColor: Color1,
borderColor: Color1,
borderWidth: 2,
pointRadius: 0 // Removes the dots
}, {
label: 'Task Growth',
yAxisID: 'B',
data: taskNumbers,
backgroundColor: Color2,
borderColor: Color2,
borderWidth: 2,
pointRadius: 0 // Removes the dots
}]
};
userOptions.options.plugins.title.text = 'User and Task Growth Over Time';
new Chart(ctxUser, userOptions);
//update the number figure
var lastEntry = data[data.length - 1];
var user = parseInt(lastEntry.Users);
var task = parseInt(lastEntry.Tasks);
var gpu = Math.ceil(task * 9.7);
const lastDate = new Date(lastEntry.Date);
const formattedDate = lastDate.toLocaleString('en-US', { month: 'long', year: 'numeric'});
document.getElementById("stats-date").innerHTML = "Data as of "+formattedDate;
document.getElementById("stats-task").innerHTML = task.toLocaleString('en-US', {minimumFractionDigits: 0})
document.getElementById("stats-user").innerHTML = user.toLocaleString('en-US', {minimumFractionDigits: 0})
document.getElementById("stats-gpu").innerHTML = gpu.toLocaleString('en-US', {minimumFractionDigits: 0})
}
fetch('https://tacc.ust.hk/stats.csv')
.then(response => response.text())
.then(csvText => Papa.parse(csvText, { header: true, skipEmptyLines:true }))
.then(data => setupChart(data.data));
</script>
</body>
</html>