Hugging Face's logo

RajBhope
/
gpu-runtime-predictor

Tabular Regression
code
scikit-learn
gpu-runtime-prediction
code-understanding
regression
performance-modeling
Model card Files
xet
 Community
RajBhope commited on
Commit
a9d6cc9
·
verified ·
1 Parent(s): fd495f4

Upload gpu_catalog.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. gpu_catalog.json +182 -0
gpu_catalog.json ADDED
@@ -0,0 +1,182 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "NVIDIA_T4": {
3
+ "name": "NVIDIA T4",
4
+ "cuda_cores": 2560,
5
+ "tensor_cores": 320,
6
+ "memory_gb": 16,
7
+ "memory_bandwidth_gbps": 320,
8
+ "base_clock_mhz": 585,
9
+ "boost_clock_mhz": 1590,
10
+ "sm_count": 40,
11
+ "fp32_tflops": 8.1,
12
+ "fp16_tflops": 65,
13
+ "tdp_watts": 70,
14
+ "compute_capability": 7.5,
15
+ "l2_cache_mb": 4
16
+ },
17
+ "NVIDIA_V100": {
18
+ "name": "NVIDIA V100",
19
+ "cuda_cores": 5120,
20
+ "tensor_cores": 640,
21
+ "memory_gb": 32,
22
+ "memory_bandwidth_gbps": 900,
23
+ "base_clock_mhz": 1230,
24
+ "boost_clock_mhz": 1530,
25
+ "sm_count": 80,
26
+ "fp32_tflops": 15.7,
27
+ "fp16_tflops": 125,
28
+ "tdp_watts": 300,
29
+ "compute_capability": 7.0,
30
+ "l2_cache_mb": 6
31
+ },
32
+ "NVIDIA_A10G": {
33
+ "name": "NVIDIA A10G",
34
+ "cuda_cores": 9216,
35
+ "tensor_cores": 288,
36
+ "memory_gb": 24,
37
+ "memory_bandwidth_gbps": 600,
38
+ "base_clock_mhz": 885,
39
+ "boost_clock_mhz": 1695,
40
+ "sm_count": 80,
41
+ "fp32_tflops": 31.2,
42
+ "fp16_tflops": 62.5,
43
+ "tdp_watts": 150,
44
+ "compute_capability": 8.6,
45
+ "l2_cache_mb": 6
46
+ },
47
+ "NVIDIA_A100_40GB": {
48
+ "name": "NVIDIA A100 40GB",
49
+ "cuda_cores": 6912,
50
+ "tensor_cores": 432,
51
+ "memory_gb": 40,
52
+ "memory_bandwidth_gbps": 1555,
53
+ "base_clock_mhz": 765,
54
+ "boost_clock_mhz": 1410,
55
+ "sm_count": 108,
56
+ "fp32_tflops": 19.5,
57
+ "fp16_tflops": 312,
58
+ "tdp_watts": 400,
59
+ "compute_capability": 8.0,
60
+ "l2_cache_mb": 40
61
+ },
62
+ "NVIDIA_A100_80GB": {
63
+ "name": "NVIDIA A100 80GB",
64
+ "cuda_cores": 6912,
65
+ "tensor_cores": 432,
66
+ "memory_gb": 80,
67
+ "memory_bandwidth_gbps": 2039,
68
+ "base_clock_mhz": 765,
69
+ "boost_clock_mhz": 1410,
70
+ "sm_count": 108,
71
+ "fp32_tflops": 19.5,
72
+ "fp16_tflops": 312,
73
+ "tdp_watts": 400,
74
+ "compute_capability": 8.0,
75
+ "l2_cache_mb": 40
76
+ },
77
+ "NVIDIA_L4": {
78
+ "name": "NVIDIA L4",
79
+ "cuda_cores": 7424,
80
+ "tensor_cores": 232,
81
+ "memory_gb": 24,
82
+ "memory_bandwidth_gbps": 300,
83
+ "base_clock_mhz": 795,
84
+ "boost_clock_mhz": 2040,
85
+ "sm_count": 58,
86
+ "fp32_tflops": 30.3,
87
+ "fp16_tflops": 121,
88
+ "tdp_watts": 72,
89
+ "compute_capability": 8.9,
90
+ "l2_cache_mb": 48
91
+ },
92
+ "NVIDIA_L40S": {
93
+ "name": "NVIDIA L40S",
94
+ "cuda_cores": 18176,
95
+ "tensor_cores": 568,
96
+ "memory_gb": 48,
97
+ "memory_bandwidth_gbps": 864,
98
+ "base_clock_mhz": 1110,
99
+ "boost_clock_mhz": 2520,
100
+ "sm_count": 142,
101
+ "fp32_tflops": 91.6,
102
+ "fp16_tflops": 183.2,
103
+ "tdp_watts": 350,
104
+ "compute_capability": 8.9,
105
+ "l2_cache_mb": 96
106
+ },
107
+ "NVIDIA_RTX_3090": {
108
+ "name": "NVIDIA RTX 3090",
109
+ "cuda_cores": 10496,
110
+ "tensor_cores": 328,
111
+ "memory_gb": 24,
112
+ "memory_bandwidth_gbps": 936,
113
+ "base_clock_mhz": 1395,
114
+ "boost_clock_mhz": 1695,
115
+ "sm_count": 82,
116
+ "fp32_tflops": 35.6,
117
+ "fp16_tflops": 71,
118
+ "tdp_watts": 350,
119
+ "compute_capability": 8.6,
120
+ "l2_cache_mb": 6
121
+ },
122
+ "NVIDIA_RTX_4090": {
123
+ "name": "NVIDIA RTX 4090",
124
+ "cuda_cores": 16384,
125
+ "tensor_cores": 512,
126
+ "memory_gb": 24,
127
+ "memory_bandwidth_gbps": 1008,
128
+ "base_clock_mhz": 2235,
129
+ "boost_clock_mhz": 2520,
130
+ "sm_count": 128,
131
+ "fp32_tflops": 82.6,
132
+ "fp16_tflops": 165.2,
133
+ "tdp_watts": 450,
134
+ "compute_capability": 8.9,
135
+ "l2_cache_mb": 72
136
+ },
137
+ "NVIDIA_H100_SXM": {
138
+ "name": "NVIDIA H100 SXM",
139
+ "cuda_cores": 16896,
140
+ "tensor_cores": 528,
141
+ "memory_gb": 80,
142
+ "memory_bandwidth_gbps": 3350,
143
+ "base_clock_mhz": 1095,
144
+ "boost_clock_mhz": 1830,
145
+ "sm_count": 132,
146
+ "fp32_tflops": 67,
147
+ "fp16_tflops": 989,
148
+ "tdp_watts": 700,
149
+ "compute_capability": 9.0,
150
+ "l2_cache_mb": 50
151
+ },
152
+ "NVIDIA_H100_PCIe": {
153
+ "name": "NVIDIA H100 PCIe",
154
+ "cuda_cores": 14592,
155
+ "tensor_cores": 456,
156
+ "memory_gb": 80,
157
+ "memory_bandwidth_gbps": 2039,
158
+ "base_clock_mhz": 1095,
159
+ "boost_clock_mhz": 1620,
160
+ "sm_count": 114,
161
+ "fp32_tflops": 48,
162
+ "fp16_tflops": 756,
163
+ "tdp_watts": 350,
164
+ "compute_capability": 9.0,
165
+ "l2_cache_mb": 50
166
+ },
167
+ "NVIDIA_RTX_A6000": {
168
+ "name": "NVIDIA RTX A6000",
169
+ "cuda_cores": 10752,
170
+ "tensor_cores": 336,
171
+ "memory_gb": 48,
172
+ "memory_bandwidth_gbps": 768,
173
+ "base_clock_mhz": 1410,
174
+ "boost_clock_mhz": 1860,
175
+ "sm_count": 84,
176
+ "fp32_tflops": 38.7,
177
+ "fp16_tflops": 77.4,
178
+ "tdp_watts": 300,
179
+ "compute_capability": 8.6,
180
+ "l2_cache_mb": 6
181
+ }
182
+ }