Wikidata on Allegrograph


Wolfgang Fahl

Import

Import
state  ✅
url  https://wiki.bitplan.com/index.php/Wikidata_on_Allegrograph
target  Allegrograph
start  2022-04-15
end  
days  0.3
os  CentOS Linux release 7.9.2009
cpu  AMD EPYC 7302, 3.0GHz
ram  256
triples  16.8
comment  

Freitext

Craig Norvell wrote on 2022-07-25:

We loaded

https://dumps.wikimedia.org/wikidatawiki/entities/latest-all.nt.gz  (Probably from mid-April 2022 time frame)

Machine - 32core (2x16 AMD EPYC 7302, 3.0GHz), 256GB RAM, CentOS Linux release 7.9.2009 with SSDs.  Load using only 16 cores took <6 hours. 

Triple count - 16,882,554,798

Size on Disk - 1,486GB

For faster queries we optimized the indices (Oscore to 1.0) and that took <6 hours.

74% of the queries took <100 milliseconds.

98.5% finished in <1sec

There was a single query that took 7.19 seconds and we might have a look to see if adding an index might change that result.

We have not investigated different hardware configs (<RAM, Spinning Disks, <Cores, etc).  

Import[edit]

Import
state  ✅
url  https://wiki.bitplan.com/index.php/Wikidata_on_Allegrograph
target  Allegrograph
start  2022-04-15
end  
days  0.3
os  CentOS Linux release 7.9.2009
cpu  AMD EPYC 7302, 3.0GHz
ram  256
triples  16.8
comment  

Freitext[edit]

Craig Norvell wrote on 2022-07-25:

We loaded

https://dumps.wikimedia.org/wikidatawiki/entities/latest-all.nt.gz  (Probably from mid-April 2022 time frame)

Machine - 32core (2x16 AMD EPYC 7302, 3.0GHz), 256GB RAM, CentOS Linux release 7.9.2009 with SSDs.  Load using only 16 cores took <6 hours. 

Triple count - 16,882,554,798

Size on Disk - 1,486GB

For faster queries we optimized the indices (Oscore to 1.0) and that took <6 hours.

74% of the queries took <100 milliseconds.

98.5% finished in <1sec

There was a single query that took 7.19 seconds and we might have a look to see if adding an index might change that result.

We have not investigated different hardware configs (<RAM, Spinning Disks, <Cores, etc).  
🖨 🚪