3090 ai-benchmark.com [run under NGC 20.10-tf1]

gistfile1.txt

>>   AI-Benchmark-v.0.1.2   
>>   Let the AI Games begin..

2020-11-11 16:50:17.642293: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 2599990000 Hz
2020-11-11 16:50:17.643424: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x560d110 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-11-11 16:50:17.643440: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
2020-11-11 16:50:17.645168: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcuda.so.1
2020-11-11 16:50:17.767055: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x560f580 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-11-11 16:50:17.767084: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): GeForce RTX 3090, Compute Capability 8.6
2020-11-11 16:50:17.768002: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:17.768029: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:17.770245: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:17.771047: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:17.771265: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:17.773617: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:17.774173: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:17.774286: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:17.775851: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:17.775877: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:18.038214: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:50:18.038250: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:50:18.038256: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:50:18.039962: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
2020-11-11 16:50:18.041529: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:18.041559: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:18.041579: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:18.041588: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:18.041596: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:18.041604: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:18.041613: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:18.041621: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:18.043161: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:18.043179: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:50:18.043185: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:50:18.043191: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:50:18.044716: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
2020-11-11 16:50:18.045795: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:18.045816: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:18.045831: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:18.045840: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:18.045848: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:18.045856: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:18.045864: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:18.045873: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:18.047381: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:18.047395: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:50:18.047402: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:50:18.047407: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:50:18.048917: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
*  TF Version: 1.15.4
*  Platform: Linux-5.4.0-52-generic-x86_64-with-Ubuntu-18.04-bionic
*  CPU: N/A
*  CPU RAM: 63 GB
*  GPU/0: GeForce RTX 3090
*  GPU RAM: 21.8 GB
*  CUDA Version: 11.1
*  CUDA Build: V11.1.74

The benchmark is running...
The tests might take up to 20 minutes
Please don't interrupt the script

1/19. MobileNet-V2

2020-11-11 16:50:25.051044: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:25.051077: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:25.051096: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:25.051105: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:25.051113: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:25.051123: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:25.051133: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:25.051142: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:25.052687: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:25.053943: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:25.053967: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:25.053985: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:25.053995: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:25.054005: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:25.054015: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:25.054025: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:25.054035: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:25.055571: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:25.055597: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:50:25.055603: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:50:25.055609: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:50:25.057169: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
2020-11-11 16:50:26.542664: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:27.996046: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
1.1 - inference | batch=50, size=224x224: 57.4 ± 133.4 ms
1.2 - training  | batch=50, size=224x224: 209 ± 184 ms

2/19. Inception-V3

2020-11-11 16:50:49.900087: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:50:49.900125: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:50:49.900148: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:50:49.900158: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:50:49.900168: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:50:49.900177: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:50:49.900186: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:50:49.900195: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:50:49.901040: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:50:49.901066: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:50:49.901072: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:50:49.901078: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:50:49.901948: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
2.1 - inference | batch=20, size=346x346: 40.7 ± 0.6 ms
2.2 - training  | batch=20, size=346x346: 117.5 ± 0.7 ms

3/19. Inception-V4

2020-11-11 16:51:05.145630: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:51:05.145666: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:51:05.145686: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:51:05.145695: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:51:05.145704: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:51:05.145713: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:51:05.145722: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:51:05.145734: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:51:05.146592: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:51:05.146617: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:51:05.146623: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:51:05.146629: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:51:05.147486: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
3.1 - inference | batch=10, size=346x346: 35.7 ± 1.0 ms
3.2 - training  | batch=10, size=346x346: 149.0 ± 0.4 ms

4/19. Inception-ResNet-V2

2020-11-11 16:51:23.419787: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:51:23.419824: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:51:23.419845: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:51:23.419853: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:51:23.419862: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:51:23.419871: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:51:23.419879: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:51:23.419888: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:51:23.420736: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:51:23.420759: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:51:23.420766: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:51:23.420772: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:51:23.421633: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
4.1 - inference | batch=10, size=346x346: 49.0 ± 0.7 ms
4.2 - training  | batch=8, size=346x346: 157.4 ± 0.6 ms

5/19. ResNet-V2-50

2020-11-11 16:51:51.334653: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:51:51.334689: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:51:51.334709: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:51:51.334718: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:51:51.334728: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:51:51.334736: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:51:51.334743: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:51:51.334751: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:51:51.335604: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:51:51.335630: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:51:51.335636: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:51:51.335642: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:51:51.336505: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
5.1 - inference | batch=10, size=346x346: 25.3 ± 0.8 ms
5.2 - training  | batch=10, size=346x346: 71.9 ± 0.5 ms

6/19. ResNet-V2-152

2020-11-11 16:52:01.151372: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:52:01.151410: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:52:01.151436: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:52:01.151448: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:52:01.151458: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:52:01.151469: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:52:01.151478: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:52:01.151489: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:52:01.152338: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:52:01.152361: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:52:01.152368: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:52:01.152374: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:52:01.153236: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
6.1 - inference | batch=10, size=256x256: 31.9 ± 0.5 ms
6.2 - training  | batch=10, size=256x256: 131.2 ± 0.5 ms

7/19. VGG-16

2020-11-11 16:52:22.632797: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:52:22.632835: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:52:22.632855: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:52:22.632863: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:52:22.632872: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:52:22.632881: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:52:22.632889: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:52:22.632906: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:52:22.633761: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:52:22.633785: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:52:22.633793: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:52:22.633799: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:52:22.634679: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
7.1 - inference | batch=20, size=224x224: 41.1 ± 0.7 ms
7.2 - training  | batch=2, size=224x224: 53.5 ± 0.5 ms

8/19. SRCNN 9-5-5

2020-11-11 16:52:31.432670: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:52:31.432703: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:52:31.432730: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:52:31.432741: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:52:31.432752: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:52:31.432762: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:52:31.432771: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:52:31.432781: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:52:31.433626: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:52:31.433646: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:52:31.433652: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:52:31.433658: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:52:31.434526: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
8.1 - inference | batch=10, size=512x512: 66.0 ± 0.8 ms
8.2 - inference | batch=1, size=1536x1536: 56.1 ± 5.6 ms
8.3 - training  | batch=10, size=512x512: 123 ± 5 ms

9/19. VGG-19 Super-Res

2020-11-11 16:52:53.839184: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:52:53.839219: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:52:53.839240: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:52:53.839247: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:52:53.839257: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:52:53.839265: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:52:53.839274: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:52:53.839283: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:52:53.840131: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:52:53.840149: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:52:53.840156: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:52:53.840162: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:52:53.841028: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
9.1 - inference | batch=10, size=256x256: 52.6 ± 0.5 ms
9.2 - inference | batch=1, size=1024x1024: 92.2 ± 1.6 ms
9.3 - training  | batch=10, size=224x224: 146 ± 1 ms

10/19. ResNet-SRGAN

2020-11-11 16:53:15.269768: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:53:15.269801: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:53:15.269821: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:53:15.269829: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:53:15.269837: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:53:15.269845: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:53:15.269855: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:53:15.269863: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:53:15.270717: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:53:15.270737: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:53:15.270743: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:53:15.270749: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:53:15.271606: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
10.1 - inference | batch=10, size=512x512: 63.4 ± 0.8 ms
10.2 - inference | batch=1, size=1536x1536: 59.3 ± 0.5 ms
10.3 - training  | batch=5, size=512x512: 91.8 ± 2.2 ms

11/19. ResNet-DPED

2020-11-11 16:53:34.147341: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:53:34.147376: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:53:34.147398: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:53:34.147406: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:53:34.147416: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:53:34.147425: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:53:34.147433: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:53:34.147446: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:53:34.148300: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:53:34.148323: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:53:34.148329: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:53:34.148335: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:53:34.149193: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
11.1 - inference | batch=10, size=256x256: 58.4 ± 0.5 ms
11.2 - inference | batch=1, size=1024x1024: 111.8 ± 0.7 ms
11.3 - training  | batch=15, size=128x128: 89.8 ± 0.6 ms

12/19. U-Net

2020-11-11 16:54:00.417269: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:54:00.417302: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:54:00.417322: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:54:00.417330: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:54:00.417341: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:54:00.417350: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:54:00.417358: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:54:00.417368: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:54:00.418226: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:54:00.418247: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:54:00.418254: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:54:00.418260: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:54:00.419124: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
12.1 - inference | batch=4, size=512x512: 105.9 ± 0.6 ms
12.2 - inference | batch=1, size=1024x1024: 129.1 ± 0.6 ms
12.3 - training  | batch=4, size=256x256: 113.5 ± 0.5 ms

13/19. Nvidia-SPADE

2020-11-11 16:54:20.189088: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:54:20.189127: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:54:20.189152: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:54:20.189163: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:54:20.189174: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:54:20.189184: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:54:20.189193: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:54:20.189204: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:54:20.190065: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:54:20.190088: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:54:20.190096: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:54:20.190101: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:54:20.191105: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
13.1 - inference | batch=5, size=128x128: 42.5 ± 0.5 ms
13.2 - training  | batch=1, size=128x128: 78.6 ± 0.6 ms

14/19. ICNet

2020-11-11 16:54:35.955618: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:54:35.955655: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:54:35.955677: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:54:35.955685: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:54:35.955695: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:54:35.955704: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:54:35.955712: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:54:35.955723: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:54:35.956584: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:54:35.956609: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:54:35.956616: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:54:35.956622: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:54:35.957491: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
14.1 - inference | batch=5, size=1024x1536: 141.7 ± 0.5 ms
14.2 - training  | batch=10, size=1024x1536: 440.0 ± 0.7 ms

15/19. PSPNet

2020-11-11 16:55:04.770049: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:55:04.770084: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:55:04.770104: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:55:04.770112: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:55:04.770119: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:55:04.770128: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:55:04.770137: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:55:04.770147: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:55:04.770993: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:55:04.771014: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:55:04.771021: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:55:04.771027: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:55:04.771893: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
15.1 - inference | batch=5, size=720x720: 184 ± 3 ms
15.2 - training  | batch=1, size=512x512: 82.0 ± 0.6 ms

16/19. DeepLab

2020-11-11 16:55:20.985651: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:55:20.985685: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:55:20.985707: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:55:20.985715: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:55:20.985725: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:55:20.985734: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:55:20.985742: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:55:20.985752: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:55:20.986625: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:55:20.986650: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:55:20.986656: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:55:20.986662: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:55:20.987529: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
16.1 - inference | batch=2, size=512x512: 61.8 ± 0.5 ms
16.2 - training  | batch=1, size=384x384: 72.9 ± 0.9 ms

17/19. Pixel-RNN

2020-11-11 16:55:36.642363: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:55:36.642397: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:55:36.642418: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:55:36.642426: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:55:36.642437: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:55:36.642449: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:55:36.642458: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:55:36.642468: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:55:36.643341: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:55:36.643368: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:55:36.643375: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:55:36.643382: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:55:36.644263: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
17.1 - inference | batch=50, size=64x64: 291 ± 4 ms
17.2 - training  | batch=10, size=64x64: 1013 ± 18 ms

18/19. LSTM-Sentiment

2020-11-11 16:57:07.547479: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:57:07.547523: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:57:07.547548: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:57:07.547558: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:57:07.547566: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:57:07.547575: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:57:07.547591: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:57:07.547600: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:57:07.550535: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:57:07.550569: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:57:07.550576: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:57:07.550582: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:57:07.551963: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
18.1 - inference | batch=100, size=1024x300: 329 ± 3 ms
18.2 - training  | batch=10, size=1024x300: 416 ± 12 ms

19/19. GNMT-Translation

2020-11-11 16:57:30.014893: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1665] Found device 0 with properties: 
name: GeForce RTX 3090 major: 8 minor: 6 memoryClockRate(GHz): 1.725
pciBusID: 0000:67:00.0
2020-11-11 16:57:30.014930: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
2020-11-11 16:57:30.014956: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcublas.so.11
2020-11-11 16:57:30.014966: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcufft.so.10
2020-11-11 16:57:30.014976: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcurand.so.10
2020-11-11 16:57:30.014986: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusolver.so.11
2020-11-11 16:57:30.014997: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcusparse.so.11
2020-11-11 16:57:30.015008: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudnn.so.8
2020-11-11 16:57:30.015860: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1793] Adding visible gpu devices: 0
2020-11-11 16:57:30.015883: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1206] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-11-11 16:57:30.015889: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212]      0 
2020-11-11 16:57:30.015895: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1225] 0:   N 
2020-11-11 16:57:30.016756: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1351] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 22356 MB memory) -> physical GPU (device: 0, name: GeForce RTX 3090, pci bus id: 0000:67:00.0, compute capability: 8.6)
19.1 - inference | batch=1, size=1x20: 88.5 ± 0.7 ms

Device Inference Score: 19350
Device Training Score: 21227
Device AI Score: 40577