| # |
Algorithm species |
Workload (program,model,library) |
Model species |
Precision |
Dataset species |
Dataset size |
Farm |
Platform species |
Platform name |
CPU name |
CPU freq (MHz) |
GPGPU name |
GPU freq (MHz) |
OS name |
SW deps and versions |
Model design |
Compiler |
Library |
Environment |
Classification time per 1 image (sec. min/max) |
Inference latency for 1 image (min, sec.) |
Inference throughput (max, images per sec.) |
Accuracy (Top1 / Top5) |
Batch size |
Model size (B) |
Memory usage (B) |
Platform peak power (W) |
Platform price ($) |
Platform price date |
Usage cost per image ($) |
Reproducibility |
| 1 |
4b8bbc192ec57f63 |
request-mxnet-inference |
d41bbf1e489ab5e0 |
fp32 |
ImageNet |
25000 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
1416 |
|
|
Ubuntu 16.04.4 LTS |
lib-mxnet: MXNet library (built, cpu) master-fb50257 (fb50257)
mxnet-model: MXNet model (net and weights) (mxnet, resnet-18) resnet-18
python: python 3.5.2
|
MXNet model (net and weights) (mxnet, resnet-18) resnet-18 |
GNU C compiler 7.2.0 |
MXNet library (built, cpu) master-fb50257 (fb50257) |
CK_BATCH_SIZE=1
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=snake.jpg
CK_GPU_FREQUENCY=max
CK_MXNET_MODEL=mobilenet
STAT_REPEAT=5
|
0.4930 .. 0.5887 |
0.4930 .. 0.5887 |
2.0 .. 1.7 |
0.613 / 0.837 |
1 |
46803089 |
None |
6.050 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 2 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
c0ad9b9800422f98 |
fp32 |
None |
None |
RPi farm: 5 devices (Avro) |
embedded |
Raspberry Pi (Raspberry Pi 3 Model B) |
BCM2709 |
1200 |
|
|
Ubuntu 16.04.2 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library 1.5
python: python 2.7.10
|
AlexNet (authors' implementation) |
None |
TensorFlow library 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
CK_TARGET_PATH=/Users/jiashenc/CK/local/machine/farm-5
|
0.1748 .. 0.1965 |
0.1748 .. 0.1965 |
5.7 .. 5.1 |
0.000 / 0.000 |
1 |
|
None |
25.000 |
200 (20170811) |
20170811 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 3 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
c0ad9b9800422f98 |
fp32 |
None |
None |
|
embedded |
Raspberry Pi (Raspberry Pi 3 Model B) |
BCM2709 |
1200 |
|
|
Ubuntu 16.04.2 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library 1.5
python: python 2.7.12
|
AlexNet (authors' implementation) |
None |
TensorFlow library 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
CK_TARGET_PATH=/home/pi/CK/local/machine/farm-5
CUDA_VISIBLE_DEVICES=
|
0.5187 .. 0.5321 |
0.5187 .. 0.5321 |
1.9 .. 1.9 |
0.000 / 0.000 |
1 |
|
None |
5.000 |
40 (20170811) |
20170811 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 4 |
4b8bbc192ec57f63 |
classification-tensorflow |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
|
2362 |
Debian GNU/Linux 9 (stretch) |
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
lib-tensorflow: TensorFlow library (from sources, cpu, xla) 1.7
library: TensorFlow library (from sources, cpu) 1.7
model-and-weights: TensorFlow python model and weights (squeezenet) ImageNet
python: python 2.7.13
weights: TensorFlow python model and weights (mobilenet-0.75-192) ImageNet
|
TensorFlow python model and weights (mobilenet-0.75-192) ImageNet |
None |
TensorFlow library (from sources, cpu) 1.7 |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=2
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_TENSORFLOW_MODEL_MOBILENET_MULTIPLIER=0.75
CK_ENV_TENSORFLOW_MODEL_MOBILENET_RESOLUTION=192
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
CK_TF_GPU_MEMORY_PERCENT=33
|
0.0468 .. 0.0797 |
0.0468 .. 0.0797 |
21.4 .. 12.5 |
0.668 / 0.871 |
1 |
10498594 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 5 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Microsoft Azure |
server |
Microsoft Corporation 7.0 (Virtual Machine) |
Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-OzTSvM.prototxt
CK_CAFFE_MODEL_FILE=tmp-OzTSvM.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0332 .. 0.0502 |
0.0332 .. 0.0502 |
30.1 .. 19.9 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
800 (20180101) |
20180101 |
1.41e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 6 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Microsoft Azure |
server |
Microsoft Corporation 7.0 (Virtual Machine) |
Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=8
CK_CAFFE_ITERATIONS=62
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-8G4yFt.prototxt
CK_CAFFE_MODEL_FILE=tmp-8G4yFt.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0266 .. 0.0285 |
|
37.6 .. 35.1 |
0.707 / 0.898 |
8 |
102462397 |
None |
105.000 |
800 (20180101) |
20180101 |
1.13e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 7 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Microsoft Azure |
server |
Microsoft Corporation 7.0 (Virtual Machine) |
Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=8
CK_CAFFE_ITERATIONS=62
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-EwUVuc.prototxt
CK_CAFFE_MODEL_FILE=tmp-EwUVuc.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0449 .. 0.0951 |
|
22.3 .. 10.5 |
0.754 / 0.923 |
8 |
95533753 |
None |
105.000 |
800 (20180101) |
20180101 |
1.91e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 8 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Microsoft Azure |
server |
Microsoft Corporation 7.0 (Virtual Machine) |
Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-ycetVU.prototxt
CK_CAFFE_MODEL_FILE=tmp-ycetVU.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/fggwork/ck-tools/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0502 .. 0.0889 |
0.0502 .. 0.0889 |
19.9 .. 11.3 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
800 (20180101) |
20180101 |
2.13e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 9 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-KdcKOP.prototxt
CK_CAFFE_MODEL_FILE=tmp-KdcKOP.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0382 .. 0.0612 |
0.0382 .. 0.0612 |
26.2 .. 16.3 |
0.753 / 0.923 |
1 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
2.02e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 10 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-WTFFut.prototxt
CK_CAFFE_MODEL_FILE=tmp-WTFFut.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0518 .. 0.0529 |
|
19.3 .. 18.9 |
0.753 / 0.923 |
64 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
2.74e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 11 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=40
CK_CAFFE_ITERATIONS=12
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-xrP7RZ.prototxt
CK_CAFFE_MODEL_FILE=tmp-xrP7RZ.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=20
|
0.0945 .. 0.0949 |
|
10.6 .. 10.5 |
0.753 / 0.923 |
40 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 12 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-iXtTQw.prototxt
CK_CAFFE_MODEL_FILE=tmp-iXtTQw.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a02d0432db5cdc64/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=10
|
0.1295 .. 0.1932 |
0.1163 .. 0.1170 |
7.7 .. 5.2 |
0.753 / 0.923 |
1 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 13 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
None |
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
ARM Mali-G71 |
807 |
Debian GNU/Linux 9 (stretch) |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c13 (d8f69c13)
opencl: OpenCL library r6p0-instr1-cl2
weights: MobileNet weights 1_0.25_128
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_0.25_128 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c13 (d8f69c13) |
CK_BATCHES_DIR=batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=batches.txt
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_MOBILENET_RESOLUTION=128
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=0.25
CK_GPU_FREQUENCY=max
CK_IMG_LIST=images.txt
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.0099 .. 0.0350 |
0.0099 .. 0.0350 |
100.7 .. 28.6 |
0.410 / 0.672 |
1 |
1990786 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 14 |
4b8bbc192ec57f63 |
request-tvm-vta-pynq |
d41bbf1e489ab5e0 |
int8 |
ImageNet |
2000 |
|
fpga |
Xilinx PYNQ-Z1 FPGA (ZYNQ XC7Z020-1CLG400C) |
Programmable logic equivalent to Artix-7 FPGA |
|
|
100 |
Ubuntu 15.10 |
lib-nnvm: NNVM library master-541cf56 (541cf56)
lib-tvm: TVM library master-e4c2af9 (e4c2af9)
lib-vta: VTA python library master-9ad9277 (9ad9277)
model: VTA model (net and weights) (resnet 18, int8, 20180404) 20180404
|
VTA model (net and weights) (resnet 18, int8, 20180404) 20180404 |
None |
None |
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=cat.jpg
CK_GPU_FREQUENCY=max
CK_MACHINE_HOST=cknowledge.ddns.net
CK_MACHINE_PORT=9091
STAT_REPEAT=5
|
0.4293 .. 0.4321 |
0.4293 .. 0.4321 |
2.3 .. 2.3 |
0.370 / 0.630 |
1 |
129770000 |
None |
2.500 |
229 (20180404) |
20180404 |
|
Repositories
ACM badges
functional, replicated
|
| 15 |
4b8bbc192ec57f63 |
classification-tensorflow |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
|
2362 |
Debian GNU/Linux 9 (stretch) |
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
lib-tensorflow: TensorFlow library (from sources, cpu, xla) 1.7
library: TensorFlow library (from sources, cpu) 1.7
model-and-weights: TensorFlow python model and weights (squeezenet) ImageNet
python: python 2.7.13
weights: TensorFlow python model and weights (mobilenet-1.0-224) ImageNet
|
TensorFlow python model and weights (mobilenet-1.0-224) ImageNet |
None |
TensorFlow library (from sources, cpu) 1.7 |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=2
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_TENSORFLOW_MODEL_MOBILENET_MULTIPLIER=1.0
CK_ENV_TENSORFLOW_MODEL_MOBILENET_RESOLUTION=224
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
CK_TF_GPU_MEMORY_PERCENT=33
|
0.0829 .. 0.1465 |
0.0829 .. 0.1465 |
12.1 .. 6.8 |
0.705 / 0.894 |
1 |
17106694 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 16 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-F7GXbg.prototxt
CK_CAFFE_MODEL_FILE=tmp-F7GXbg.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0099 .. 0.0102 |
0.0099 .. 0.0102 |
100.6 .. 98.5 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
4.23e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 17 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=24
CK_CAFFE_ITERATIONS=20
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-jRkPk7.prototxt
CK_CAFFE_MODEL_FILE=tmp-jRkPk7.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0057 .. 0.0057 |
|
177.0 .. 176.3 |
0.754 / 0.923 |
24 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
2.40e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 18 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (opencl) 17.12-48bc34e (48bc34e)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (opencl) 17.12-48bc34e (48bc34e) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.2108 .. 0.2138 |
0.2108 .. 0.2138 |
4.7 .. 4.7 |
0.620 / 0.848 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 19 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (opencl) 17.12-48bc34e (48bc34e)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (opencl) 17.12-48bc34e (48bc34e) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=0
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.2185 .. 0.2218 |
0.2185 .. 0.2218 |
4.6 .. 4.5 |
0.626 / 0.850 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 20 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-_lpT1j.prototxt
CK_CAFFE_MODEL_FILE=tmp-_lpT1j.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0041 .. 0.0041 |
|
242.3 .. 241.8 |
0.753 / 0.923 |
64 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.75e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 21 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
int8 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-y13Oj2.prototxt
CK_CAFFE_MODEL_FILE=tmp-y13Oj2.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-intel-i8/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0071 .. 0.0074 |
0.0071 .. 0.0074 |
140.4 .. 135.5 |
0.753 / 0.923 |
1 |
23883438.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
3.03e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 22 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
a3fcac86d42bdbc4 |
fp32 |
None |
None |
|
embedded |
NVIDIA Jetson TX1 |
ARMv8 Processor rev 1 (v8l) |
|
|
998 |
Ubuntu 16.04.4 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library (from sources, cuda) 1.5
python: python 3.5.2
|
VGG 16 (authors' implementation) |
None |
TensorFlow library (from sources, cuda) 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
|
0.2787 .. 0.2831 |
0.2787 .. 0.2831 |
3.6 .. 3.5 |
0.000 / 0.000 |
1 |
|
None |
15.000 |
435 (20180520) |
20180520 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 23 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-k2n61l.prototxt
CK_CAFFE_MODEL_FILE=tmp-k2n61l.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=10
|
0.0103 .. 0.0104 |
|
97.2 .. 96.0 |
0.707 / 0.898 |
64 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 24 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-_pf3jy.prototxt
CK_CAFFE_MODEL_FILE=tmp-_pf3jy.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=20
|
0.0144 .. 0.0149 |
0.0109 .. 0.0112 |
69.2 .. 67.1 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 25 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
None |
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
ARM Mali-G71 |
807 |
Debian GNU/Linux 9 (stretch) |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c13 (d8f69c13)
opencl: OpenCL library r6p0-instr1-cl2
weights: MobileNet weights 1_0.50_160
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_0.50_160 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c13 (d8f69c13) |
CK_BATCHES_DIR=batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=batches.txt
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_MOBILENET_RESOLUTION=160
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=0.5
CK_GPU_FREQUENCY=max
CK_IMG_LIST=images.txt
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.0173 .. 0.0654 |
0.0173 .. 0.0654 |
57.9 .. 15.3 |
0.584 / 0.822 |
1 |
5459810 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 26 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
c0ad9b9800422f98 |
fp32 |
None |
None |
|
embedded |
NVIDIA Jetson TX2 |
ARMv8 Processor rev 3 (v8l) |
2035 |
|
|
Ubuntu 16.04.3 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library 1.5
python: python 2.7.12
|
AlexNet (authors' implementation) |
None |
TensorFlow library 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
CUDA_VISIBLE_DEVICES=
|
0.2007 .. 0.2015 |
0.2007 .. 0.2015 |
5.0 .. 5.0 |
0.000 / 0.000 |
1 |
|
None |
15.000 |
499 (20180520) |
20180520 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 27 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
a3fcac86d42bdbc4 |
fp32 |
None |
None |
RPi farm: 11 devices (Avro) |
embedded |
Raspberry Pi (Raspberry Pi 3 Model B) |
BCM2709 |
1200 |
|
|
Ubuntu 16.04.2 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library 1.5
python: python 2.7.10
|
VGG 16 (authors' implementation) |
None |
TensorFlow library 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
CK_TARGET_PATH=/Users/jiashenc/CK/local/machine/farm-11
|
0.3397 .. 0.5254 |
0.3397 .. 0.5254 |
2.9 .. 1.9 |
0.000 / 0.000 |
1 |
|
None |
55.000 |
440 (20170811) |
20170811 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 28 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
|
|
GeForce GTX 1080 |
1600 |
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-VaiUzH.prototxt
CK_CAFFE_MODEL_FILE=tmp-VaiUzH.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.0232 .. 0.0238 |
0.0695 .. 0.0703 |
43.1 .. 42.1 |
0.754 / 0.923 |
1 |
95533753 |
None |
180.000 |
700 (20180101) |
20180101 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 29 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
|
|
GeForce GTX 1080 |
1600 |
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=24
CK_CAFFE_ITERATIONS=20
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-GdnhgZ.prototxt
CK_CAFFE_MODEL_FILE=tmp-GdnhgZ.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.0046 .. 0.0047 |
|
217.2 .. 212.1 |
0.754 / 0.923 |
24 |
95533753 |
None |
180.000 |
700 (20180101) |
20180101 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 30 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=32
CK_CAFFE_ITERATIONS=15
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-ICKaAX.prototxt
CK_CAFFE_MODEL_FILE=tmp-ICKaAX.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0622 .. 0.0632 |
|
16.1 .. 15.8 |
0.754 / 0.923 |
32 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
3.28e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 31 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-oBhiJj.prototxt
CK_CAFFE_MODEL_FILE=tmp-oBhiJj.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0515 .. 0.0796 |
0.0515 .. 0.0796 |
19.4 .. 12.6 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
2.72e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 32 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c1 (d8f69c1)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c1 (d8f69c1) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.2369 .. 0.2417 |
0.2369 .. 0.2417 |
4.2 .. 4.1 |
0.722 / 0.896 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 33 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c1 (d8f69c1)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c1 (d8f69c1) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=0
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.1465 .. 0.1484 |
0.1465 .. 0.1484 |
6.8 .. 6.7 |
0.722 / 0.896 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 34 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
|
|
GeForce GTX 1080 |
1600 |
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-5_2l8I.prototxt
CK_CAFFE_MODEL_FILE=tmp-5_2l8I.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.0134 .. 0.0137 |
0.0045 .. 0.0045 |
74.9 .. 73.2 |
0.707 / 0.898 |
1 |
102462397 |
None |
180.000 |
700 (20180101) |
20180101 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 35 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
|
|
GeForce GTX 1080 |
1600 |
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cudnn) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=24
CK_CAFFE_ITERATIONS=20
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-mmZrQF.prototxt
CK_CAFFE_MODEL_FILE=tmp-mmZrQF.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.0045 .. 0.0045 |
|
222.6 .. 222.3 |
0.707 / 0.898 |
24 |
102462397 |
None |
180.000 |
700 (20180101) |
20180101 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 36 |
4b8bbc192ec57f63 |
classification-tensorflow |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
|
2362 |
Debian GNU/Linux 9 (stretch) |
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
lib-tensorflow: TensorFlow library (from sources, cpu, xla) 1.7
library: TensorFlow library (from sources, cpu) 1.7
model-and-weights: TensorFlow python model and weights (squeezenet) ImageNet
python: python 2.7.13
weights: TensorFlow python model and weights (mobilenet-0.75-160) ImageNet
|
TensorFlow python model and weights (mobilenet-0.75-160) ImageNet |
None |
TensorFlow library (from sources, cpu) 1.7 |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=2
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_TENSORFLOW_MODEL_MOBILENET_MULTIPLIER=0.75
CK_ENV_TENSORFLOW_MODEL_MOBILENET_RESOLUTION=160
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
CK_TF_GPU_MEMORY_PERCENT=33
|
0.0367 .. 0.0676 |
0.0367 .. 0.0676 |
27.3 .. 14.8 |
0.644 / 0.856 |
1 |
10498594 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 37 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-JAbQVt.prototxt
CK_CAFFE_MODEL_FILE=tmp-JAbQVt.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0328 .. 0.0488 |
0.0328 .. 0.0488 |
30.5 .. 20.5 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.73e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 38 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=24
CK_CAFFE_ITERATIONS=20
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-I3amQ8.prototxt
CK_CAFFE_MODEL_FILE=tmp-I3amQ8.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=2
|
0.0370 .. 0.0404 |
|
27.0 .. 24.8 |
0.707 / 0.898 |
24 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.95e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 39 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-ZTsvrJ.prototxt
CK_CAFFE_MODEL_FILE=tmp-ZTsvrJ.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0021 .. 0.0022 |
|
465.3 .. 458.8 |
0.704 / 0.897 |
64 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
9.13e-07 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 40 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-u2KPZ_.prototxt
CK_CAFFE_MODEL_FILE=tmp-u2KPZ_.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0041 .. 0.0041 |
0.0041 .. 0.0041 |
245.1 .. 241.9 |
0.704 / 0.897 |
1 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.73e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 41 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-gGv0uB.prototxt
CK_CAFFE_MODEL_FILE=tmp-gGv0uB.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.2356 .. 0.2397 |
0.2254 .. 0.2260 |
4.2 .. 4.2 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 42 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-CZ5Cmo.prototxt
CK_CAFFE_MODEL_FILE=tmp-CZ5Cmo.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.2229 .. 0.2241 |
|
4.5 .. 4.5 |
0.754 / 0.923 |
64 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 43 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 5.4.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-voPrDv.prototxt
CK_CAFFE_MODEL_FILE=tmp-voPrDv.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.2240 .. 0.2262 |
0.2240 .. 0.2262 |
4.5 .. 4.4 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
9.52e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 44 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
GNU C compiler 5.4.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=40
CK_CAFFE_ITERATIONS=12
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-y2uIiU.prototxt
CK_CAFFE_MODEL_FILE=tmp-y2uIiU.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-inception-v3-fp32/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
|
0.2108 .. 0.2130 |
|
4.7 .. 4.7 |
0.754 / 0.923 |
40 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
8.96e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 45 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
c0ad9b9800422f98 |
fp32 |
None |
None |
|
embedded |
NVIDIA Jetson TX1 |
ARMv8 Processor rev 1 (v8l) |
1734 |
|
|
Ubuntu 16.04.4 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library (from sources, cuda) 1.5
python: python 3.5.2
|
AlexNet (authors' implementation) |
None |
TensorFlow library (from sources, cuda) 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
CUDA_VISIBLE_DEVICES=
|
0.1842 .. 0.1847 |
0.1842 .. 0.1847 |
5.4 .. 5.4 |
0.000 / 0.000 |
1 |
|
None |
15.000 |
435 (20180520) |
20180520 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 46 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
None |
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
ARM Mali-G71 |
807 |
Debian GNU/Linux 9 (stretch) |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c13 (d8f69c13)
opencl: OpenCL library r6p0-instr1-cl2
weights: MobileNet weights 1_1.0_192
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_192 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c13 (d8f69c13) |
CK_BATCHES_DIR=batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=batches.txt
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_MOBILENET_RESOLUTION=192
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMG_LIST=images.txt
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.0397 .. 0.1689 |
0.0397 .. 0.1689 |
25.2 .. 5.9 |
0.704 / 0.882 |
1 |
17106694 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 47 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=40
CK_CAFFE_ITERATIONS=12
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-AJJSv2.prototxt
CK_CAFFE_MODEL_FILE=tmp-AJJSv2.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=20
|
0.0174 .. 0.0177 |
|
57.3 .. 56.6 |
0.754 / 0.923 |
40 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 48 |
4b8bbc192ec57f63 |
caffe |
1b339ddb13408f8f |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (inception-v3, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (inception-v3, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-pHjHEG.prototxt
CK_CAFFE_MODEL_FILE=tmp-pHjHEG.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/f9db041999a38218/inceptionv3.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=10
|
0.0216 .. 0.0362 |
0.0176 .. 0.0179 |
46.3 .. 27.6 |
0.754 / 0.923 |
1 |
95533753 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 49 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
None |
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
ARM Mali-G71 |
807 |
Debian GNU/Linux 9 (stretch) |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (opencl) 18.03-e40997bb (e40997bb)
opencl: OpenCL library r6p0-instr1-cl2
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (opencl) 18.03-e40997bb (e40997bb) |
CK_BATCHES_DIR=batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=batches.txt
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=0
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMG_LIST=images.txt
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.0517 .. 0.2039 |
0.0517 .. 0.2039 |
19.3 .. 4.9 |
0.722 / 0.904 |
1 |
17106694 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 50 |
4b8bbc192ec57f63 |
request-tvm-nnvm-inference |
07d4e7aa3750ddc6 |
fp16 |
ImageNet |
50000 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
|
800 |
Ubuntu 16.04.4 LTS |
lib-mxnet: MXNet library (built, cpu) master-fb50257 (fb50257)
lib-nnvm-tvm: NNVM library master-69c5ebb (69c5ebb)
mxnet-model: MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0
python: python 3.5.2
|
MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0 |
GNU C compiler 7.2.0 |
MXNet library (built, cpu) master-fb50257 (fb50257) |
CK_BATCH_SIZE=1
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=snake.jpg
CK_GPU_FREQUENCY=max
CK_TVM_DTYPE=float16
CK_TVM_MODEL=mobilenet
STAT_REPEAT=5
|
0.0558 .. 0.0591 |
0.0558 .. 0.0591 |
17.9 .. 16.9 |
0.667 / 0.877 |
1 |
8512054.5 |
None |
6.050 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 51 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-2QMOjx.prototxt
CK_CAFFE_MODEL_FILE=tmp-2QMOjx.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=36
|
0.0065 .. 0.0066 |
0.0065 .. 0.0066 |
154.5 .. 151.9 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
2.75e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 52 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=48
CK_CAFFE_ITERATIONS=10
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-7xf1pa.prototxt
CK_CAFFE_MODEL_FILE=tmp-7xf1pa.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=18
|
0.0033 .. 0.0060 |
|
304.1 .. 166.5 |
0.707 / 0.898 |
48 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.40e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 53 |
4b8bbc192ec57f63 |
classification-tensorflow |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
|
2362 |
Debian GNU/Linux 9 (stretch) |
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
lib-tensorflow: TensorFlow library (from sources, cpu, xla) 1.7
library: TensorFlow library (from sources, cpu) 1.7
model-and-weights: TensorFlow python model and weights (squeezenet) ImageNet
python: python 2.7.13
weights: TensorFlow python model and weights (mobilenet-0.25-128) ImageNet
|
TensorFlow python model and weights (mobilenet-0.25-128) ImageNet |
None |
TensorFlow library (from sources, cpu) 1.7 |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=2
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_TENSORFLOW_MODEL_MOBILENET_MULTIPLIER=0.25
CK_ENV_TENSORFLOW_MODEL_MOBILENET_RESOLUTION=128
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
CK_TF_GPU_MEMORY_PERCENT=33
|
0.0124 .. 0.0331 |
0.0124 .. 0.0331 |
80.5 .. 30.2 |
0.407 / 0.657 |
1 |
1990786 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 54 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
None |
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
ARM Mali-G71 |
807 |
Debian GNU/Linux 9 (stretch) |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (request) request-d8f69c13 (d8f69c13)
opencl: OpenCL library r6p0-instr1-cl2
weights: MobileNet weights 1_0.50_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_0.50_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (request) request-d8f69c13 (d8f69c13) |
CK_BATCHES_DIR=batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=batches.txt
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=0.5
CK_GPU_FREQUENCY=max
CK_IMG_LIST=images.txt
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.0243 .. 0.0949 |
0.0243 .. 0.0949 |
41.1 .. 10.5 |
0.648 / 0.852 |
1 |
5459810 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 55 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (opencl) 18.03-e40997b (e40997b)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (opencl) 18.03-e40997b (e40997b) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.2372 .. 0.2420 |
0.2372 .. 0.2420 |
4.2 .. 4.1 |
0.722 / 0.896 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 56 |
4b8bbc192ec57f63 |
mobilenets-armcl-opencl |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
ARM Mali-T860 |
800 |
Ubuntu 16.04.4 LTS |
compiler: GNU C compiler 7.2.0
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
library: ARM Compute Library (opencl) 18.03-e40997b (e40997b)
opencl: OpenCL library
weights: MobileNet weights 1_1.0_224
xopenme: xOpenME run-time library 0.3
|
MobileNet weights 1_1.0_224 |
GNU C compiler 7.2.0 |
ARM Compute Library (opencl) 18.03-e40997b (e40997b) |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=1
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_COMPUTE_DEVICE_ID=0
CK_COMPUTE_PLATFORM_ID=0
CK_CONVOLUTION_METHOD_HINT=0
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/fursin/CK/local/env/6b7b8b5d320f99fa/
CK_ENV_MOBILENET_RESOLUTION=224
CK_ENV_MOBILENET_WIDTH_MULTIPLIER=1.0
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_PREPARE_ALWAYS=NO
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
|
0.1462 .. 0.1490 |
0.1462 .. 0.1490 |
6.8 .. 6.7 |
0.722 / 0.896 |
1 |
17106694 |
None |
5.000 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 57 |
4b8bbc192ec57f63 |
classification-tensorflow |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
500 |
|
embedded |
HiKey960 (HiKey960 ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd09-1 |
|
|
2362 |
Debian GNU/Linux 9 (stretch) |
imagenet-aux: ImageNet dataset (aux) 2012
imagenet-val: ImageNet dataset (validation) 2012
lib-tensorflow: TensorFlow library (from sources, cpu, xla) 1.7
library: TensorFlow library (from sources, cpu) 1.7
model-and-weights: TensorFlow python model and weights (squeezenet) ImageNet
python: python 2.7.13
weights: TensorFlow python model and weights (mobilenet-0.50-160) ImageNet
|
TensorFlow python model and weights (mobilenet-0.50-160) ImageNet |
None |
TensorFlow library (from sources, cpu) 1.7 |
CK_BATCHES_DIR=../batches
CK_BATCH_COUNT=2
CK_BATCH_LIST=../batches
CK_BATCH_SIZE=1
CK_CONVOLUTION_METHOD_HINT=1
CK_CPU_FREQUENCY=max
CK_ENV_DATASET_IMAGENET_VAL=/home/anton/CK_TOOLS/dataset-imagenet-ilsvrc2012-val-min/
CK_ENV_TENSORFLOW_MODEL_MOBILENET_MULTIPLIER=0.5
CK_ENV_TENSORFLOW_MODEL_MOBILENET_RESOLUTION=160
CK_GPU_FREQUENCY=max
CK_IMAGE_FILE=
CK_IMAGE_LIST=../images
CK_RESULTS_DIR=predictions
CK_SKIP_IMAGES=0
CK_TF_GPU_MEMORY_PERCENT=33
|
0.0245 .. 0.0624 |
0.0245 .. 0.0624 |
40.8 .. 16.0 |
0.594 / 0.820 |
1 |
5459810 |
None |
4.500 |
239 (20170425) |
20170425 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 58 |
f4f75a0a6b65e4cd |
request-iot-benchmark |
c0ad9b9800422f98 |
fp32 |
None |
None |
|
embedded |
NVIDIA Jetson TX1 |
ARMv8 Processor rev 1 (v8l) |
|
|
998 |
Ubuntu 16.04.4 LTS |
lib-keras: Keras API, TensorFlow based (prebuilt, ReQuEST@ASPLOS'18 artifact) 2.1.3
lib-tensorflow: TensorFlow library (from sources, cuda) 1.5
python: python 3.5.2
|
AlexNet (authors' implementation) |
None |
TensorFlow library (from sources, cuda) 1.5 |
CK_CPU_FREQUENCY=max
CK_GPU_FREQUENCY=max
|
0.0200 .. 0.0203 |
0.0200 .. 0.0203 |
50.0 .. 49.2 |
0.000 / 0.000 |
1 |
|
None |
15.000 |
435 (20180520) |
20180520 |
|
Repositories
ACM badges
available, reusable, replicated
|
| 59 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-obvDV3.prototxt
CK_CAFFE_MODEL_FILE=tmp-obvDV3.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0271 .. 0.0276 |
|
36.9 .. 36.3 |
0.704 / 0.897 |
64 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.43e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 60 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
Google Cloud |
server |
Google (Google Compute Engine) |
Intel(R) Xeon(R) CPU |
2000 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/gfursin/CK/ck-caffe/program/caffe/tmp/tmp-bTCeEh.prototxt
CK_CAFFE_MODEL_FILE=tmp-bTCeEh.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/gfursin/CK-TOOLS/caffemodel-resnet50-intel-i8/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=4
|
0.0203 .. 0.0371 |
0.0203 .. 0.0371 |
49.3 .. 27.0 |
0.704 / 0.897 |
1 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
1.07e-06 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 61 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 5.4.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=64
CK_CAFFE_ITERATIONS=7
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-wcCc4K.prototxt
CK_CAFFE_MODEL_FILE=tmp-wcCc4K.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.1538 .. 0.1554 |
|
6.5 .. 6.4 |
0.707 / 0.898 |
64 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
6.54e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 62 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
Amazon |
server |
Amazon EC2 (c5.9xlarge) |
Intel(R) Xeon(R) Platinum 8124M CPU @ 3.00GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 5.4.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/ubuntu/CK/ck-caffe/program/caffe/tmp/tmp-Fha9hx.prototxt
CK_CAFFE_MODEL_FILE=tmp-Fha9hx.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/ubuntu/CK-TOOLS/caffemodel-resnet50-fp32/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.1716 .. 0.1753 |
0.1716 .. 0.1753 |
5.8 .. 5.7 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
7.29e-05 |
Repositories
ACM badges
available, reusable, replicated
Review
|
| 63 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=8
CK_CAFFE_ITERATIONS=62
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-8zBTXt.prototxt
CK_CAFFE_MODEL_FILE=tmp-8zBTXt.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.1567 .. 0.1572 |
|
6.4 .. 6.4 |
0.707 / 0.898 |
8 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 64 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
fp32 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, fp32) fp32
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (cpu) master-18b09e8 (18b09e8)
|
Caffe model (net and weights) (resnet50, fp32) fp32 |
GNU C compiler 6.3.0 |
BVLC Caffe framework (cpu) master-18b09e8 (18b09e8) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-_Gj38I.prototxt
CK_CAFFE_MODEL_FILE=tmp-_Gj38I.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/a9914423b819a458/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/a9914423b819a458/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
|
0.1664 .. 0.1668 |
0.1580 .. 0.1591 |
6.0 .. 6.0 |
0.707 / 0.898 |
1 |
102462397 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 65 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=48
CK_CAFFE_ITERATIONS=10
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-HpyAax.prototxt
CK_CAFFE_MODEL_FILE=tmp-HpyAax.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=10
|
0.0249 .. 0.0252 |
|
40.2 .. 39.7 |
0.704 / 0.897 |
48 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 66 |
4b8bbc192ec57f63 |
caffe |
d777f6335496db61 |
int8 |
ImageNet |
50000 |
|
server |
Hewlett-Packard (HP Z640 Workstation) |
Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz |
2600 |
|
|
Ubuntu 16.04.4 LTS |
caffemodel: Caffe model (net and weights) (resnet50, int8) intel-i8
dataset-imagenet-aux: ImageNet dataset (aux) 2012
dataset-imagenet-lmdb: ImageNet dataset (validation, LMDB) (resize 320) (500 images) 2012
lib-caffe: BVLC Caffe framework (intel, request) request-92110fe (92110fe)
|
Caffe model (net and weights) (resnet50, int8) intel-i8 |
Intel compiler 18.0.1 |
BVLC Caffe framework (intel, request) request-92110fe (92110fe) |
CAFFE_COMPUTE_DEVICE_ID=0
CK_CAFFE_BATCH_SIZE=1
CK_CAFFE_ITERATIONS=500
CK_CAFFE_MODEL=/home/fursin/CK/ck-caffe/program/caffe/tmp/tmp-ES2WBh.prototxt
CK_CAFFE_MODEL_FILE=tmp-ES2WBh.prototxt
CK_CAFFE_MODEL_MEAN_BIN=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet_mean.binaryproto
CK_CAFFE_MODEL_WEIGHTS=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet-50-model.caffemodel
CK_CPU_FREQUENCY=max
CK_ENV_MODEL_CAFFE_WEIGHTS=/home/fursin/CK/local/env/1a7c0379a2ca4184/ResNet-50-model.caffemodel
CK_GPU_FREQUENCY=max
OMP_NUM_THREADS=10
|
0.0305 .. 0.0477 |
0.0255 .. 0.0423 |
32.8 .. 21.0 |
0.704 / 0.897 |
1 |
25615599.25 |
None |
105.000 |
1166 (20141212) |
20141212 |
|
Repositories
ACM badges
available, reusable, replicated
Review
|
| 67 |
4b8bbc192ec57f63 |
request-tvm-nnvm-inference |
d41bbf1e489ab5e0 |
fp16 |
ImageNet |
25000 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
|
800 |
Ubuntu 16.04.4 LTS |
lib-mxnet: MXNet library (built, cpu) master-fb50257 (fb50257)
lib-nnvm-tvm: NNVM library master-69c5ebb (69c5ebb)
mxnet-model: MXNet model (net and weights) (mxnet, resnet-18) resnet-18
python: python 3.5.2
|
MXNet model (net and weights) (mxnet, resnet-18) resnet-18 |
GNU C compiler 7.2.0 |
MXNet library (built, cpu) master-fb50257 (fb50257) |
CK_BATCH_SIZE=1
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=snake.jpg
CK_GPU_FREQUENCY=max
CK_TVM_DTYPE=float16
CK_TVM_MODEL=mobilenet
STAT_REPEAT=5
|
0.1258 .. 0.1269 |
0.1258 .. 0.1269 |
7.9 .. 7.9 |
0.613 / 0.837 |
1 |
23401544.5 |
None |
6.050 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 68 |
4b8bbc192ec57f63 |
request-mxnet-inference |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
50000 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
1416 |
|
|
Ubuntu 16.04.4 LTS |
lib-mxnet: MXNet library (built, cpu) master-fb50257 (fb50257)
mxnet-model: MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0
python: python 3.5.2
|
MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0 |
GNU C compiler 7.2.0 |
MXNet library (built, cpu) master-fb50257 (fb50257) |
CK_BATCH_SIZE=1
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=snake.jpg
CK_GPU_FREQUENCY=max
CK_MXNET_MODEL=mobilenet
STAT_REPEAT=5
|
0.2483 .. 0.4360 |
0.2483 .. 0.4360 |
4.0 .. 2.3 |
0.667 / 0.877 |
1 |
17024109 |
None |
6.050 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
| 69 |
4b8bbc192ec57f63 |
request-tvm-nnvm-inference |
07d4e7aa3750ddc6 |
fp32 |
ImageNet |
50000 |
|
embedded |
Rockchip (Rockchip RK3399 Firefly Board (Linux Opensource) ) |
0x41-8-0x0-0xd03-4 ; 0x41-8-0x0-0xd08-2 |
|
|
800 |
Ubuntu 16.04.4 LTS |
lib-mxnet: MXNet library (built, cpu) master-fb50257 (fb50257)
lib-nnvm-tvm: NNVM library master-69c5ebb (69c5ebb)
mxnet-model: MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0
python: python 3.5.2
|
MXNet model (net and weights) (mxnet, mobilenet, 1.0) mobilenet1.0 |
GNU C compiler 7.2.0 |
MXNet library (built, cpu) master-fb50257 (fb50257) |
CK_BATCH_SIZE=1
CK_CPU_FREQUENCY=max
CK_DATASET_FILENAME=snake.jpg
CK_GPU_FREQUENCY=max
CK_TVM_DTYPE=float32
CK_TVM_MODEL=mobilenet
STAT_REPEAT=5
|
0.0939 .. 0.0946 |
0.0939 .. 0.0946 |
10.6 .. 10.6 |
0.667 / 0.877 |
1 |
17024109.0 |
None |
6.050 |
149 (20180416) |
20180416 |
|
Repositories
Report
DOIs
ACM badges
available, reusable, replicated
|
