forked from h2oai/h2o-2
-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.txt
45 lines (34 loc) · 2.15 KB
/
README.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
This directory contains scripts to run benchmarks on the latest H2O build.
Use the `runBench.sh' script to perform the benchmarks:
USAGE: runBench.sh [options]
This script obtains the latest h2o jar from S3 and runs the benchmarks for PCA, KMeans, GLM, and BigKMeans.
OPTIONS:
-h Show this message
-t Run task:
Choices are:
all -- Runs PCA, KMeans, GLM, and BigKMeans
pca -- Runs PCA on Airlines/AllBedrooms/Covtype data
kmeans -- Runs KMeans on Airlines/AllBedrooms/Covtype data
glm -- Runs logistic/linear regression on Airlines/AllBedrooms
glm2 -- Runs logistic/linear regression on Airlines/AllBedrooms
gbm -- Runs GBM on Airlines/AllBedrooms/Covtype data
bigkmeans -- Runs KMeans on 180 GB & 1TB of synthetic data
-j JSON config:
Choices are:
161 -- Runs benchmark(s) on single machine on 161 (100GB)
162 -- Runs benchmark(s) on single machine on 162 (100GB)
163 -- Runs benchmark(s) on single machine on 163 (100GB)
164 -- Runs benchmark(s) on single machine on 164 (100GB)
161_163 -- Runs benchmark(s) on four machines 161-163 (150GB Each)
161_164 -- Runs benchmark(s) on four machines 161-164 (100GB Each)
-b Build number:
Choices are limited to the available builds you have locally. Valid arguments are integral.
-d Debug mode:
Run benchmark tasks in debug mode (smaller datasets).
-L Logging:
Use this flag to enable logging on all machines passed in with the -j option.
Use the `startloggers.sh' script to capture logs. This script has other features and is used internally by `runBench.sh':
A. Start big & little loggers (see BMLogs/README.txt)
B. Stop loggers
C. Clear caches on all systems passed in with the -j option to `runBench/sh'. Depends on root-owned compiled `flush.c' (w/ setuid bit set).
D. Gather ICED logs