JakSprats

# Intel(R) Core(TM)2 Duo CPU     T9600  @ 2.80GHz
# Ubuntu 9.10 64 bit

git clone https://github.com/luvit/luvit.git
(cd luvit
  make
  make install
)

git clone git://github.com/JakSprats/Alchemy-Database.git

JakSprats

#include <iostream>
#include <fstream>
#include <string>
#include <cstring>
#include <ctime>
#include <cstdlib>

#include <db_cxx.h>

using namespace std;

JakSprats

# download
git clone git://github.com/JakSprats/Alchemy-Database.git

# make the package
cd Alchemy-Database; make;

# change to directory w/ server, client, & benchmark-tool
cd redis/src 
  

JakSprats

#!/bin/bash

T=$(tempfile)
time sudo find /usr -type f| sort -R | head -n 4000 | sed -e "s/\ /\\\ /g" > $T
rm -f rand.files

I=0
FILES=""
cat $T | while read file; do
  if [ $I -lt 25 ]; then

JakSprats

. ./bash_functions.sh

$CLI DROP TABLE item
$CLI CREATE TABLE item "(id INT, name TEXT)"

$CLI DROP TABLE word2item
$CLI CREATE TABLE word2item "(id INT, word_id INT, type INT, seller_id INT, pub_type INT, tag_id INT, item_id INT)"

$CLI CREATE INDEX ind_w2i_w ON word2item "(word_id)"
$CLI CREATE INDEX ind_w2i_i ON word2item "(item_id)"

JakSprats

CLI=./redisql-cli

$CLI CONFIG ADD LUA test/alchemy.lua
$CLI LUA "function cap_per_fk(max, tbl, fkname, fkval, pkname)
              wc    = fkname .. '=' .. fkval;
              cnt_s = select('COUNT(*)', tbl, wc);
              cnt   = tonumber(string.sub(cnt_s, 2));
              if (cnt > max) then
                  dnum = cnt - max;
                  wcob = wc .. ' ORDER BY ' .. pkname .. ' LIMIT ' .. dnum;

JakSprats

BENCH=./gen-benchmark
1.) use PK (no sorting)
taskset -c 1 $BENCH -q -c $C -n $REQ -s -A MULTI -Q SELECT \* FROM obycol WHERE "id BETWEEN 1 AND 10 ORDER BY id LIMIT 4 OFFSET 3"
38207.73 requests per second
2.) use unindexed column (sort) and sort to one column
taskset -c 1 $BENCH -q -c $C -n $REQ -s -A MULTI -Q SELECT \* FROM obycol WHERE "id BETWEEN 1 AND 10 ORDER BY m LIMIT 4 OFFSET 3"
28815.95 requests per second
3.) use unindexed column (sort) and sort to two columns
taskset -c 1 $BENCH -q -c $C -n $REQ -s -A MULTI -Q SELECT \* FROM obycol WHERE "id BETWEEN 1 AND 10 ORDER BY j,m LIMIT 4 OFFSET 3"
28535.61 requests per second

JakSprats

import datetime
import errno
import socket
import threading
import time
import warnings
from itertools import chain, imap
from redis.exceptions import ConnectionError, ResponseError, InvalidResponse, WatchError
from redis.exceptions import RedisError, AuthenticationError

JakSprats

This test is not that conclusive, but its a start to test how much of a "tax" running AlchemyDB as a VirtualMachine must be paid.

All tests were run from one machine to another (both w/ phenom X4 @ 3.0 GHz CPUs)

The bare-metal was run on a 64bit Ubuntu 9.10
The Virtual machine was run on the same 64 bit Ubuntu 9.10 and the Virtual Machine was a Debian 5 (32 bit), the Hypervisor was VMware workstation 7.0.

A little bit stupid that the Virtual Machine was 32bit, but this is just a quick nothing-special first test to get a ballpark idea ... No special configuration was done in any of this, this is basically right out of the box.

JakSprats

nrows: TOTAL: (WHERE fk=1)1000020

set count=4 WHERE fk = 1
elapsed time: 4375.12(ms)

nrows: (WHERE fk=1) count = 4: 10000196
nrows: (WHERE fk=1) count = 99: 0

BATCH UPDATE: set count=99 WHERE fk = 1
elapsed time: 5210.04(ms)