This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Starting Application: ..........................Error 306: Error retrieving file 'logs/startup.log' | |
and | |
[1][TT-MBP01][26] % vmc crashlogs retwisj-tt /Users/joubinhoushyar 22:18:17 11-04-15 | |
====> logs/stderr.log <==== | |
Apr 16, 2011 2:14:45 AM org.apache.coyote.http11.Http11Protocol init | |
INFO: Initializing Coyote HTTP/1.1 on http-50162 | |
Apr 16, 2011 2:14:45 AM org.apache.catalina.startup.Catalina load |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
os x 10.6.7 / intel core i5 (mbp 15) | |
from .zshrc: | |
export GOROOT=/Users/alphazero/Code/lang/go | |
export GOOS=darwin | |
export GOARCH=amd64 | |
export GOBIN=$GOROOT/bin | |
export PATH=$PATH:$GOBIN |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Ippolito, Bob: Drop ACID and think about Data. March 2009. – Talk at Pycon on 2009-03-28. http://blip.tv/file/1949416/ |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
diff --git a/src/pkg/redis/reflect.go b/src/pkg/redis/reflect.go | |
index b9f2985..25a43d2 100644 | |
--- a/src/pkg/redis/reflect.go | |
+++ b/src/pkg/redis/reflect.go | |
@@ -16,7 +16,7 @@ | |
package redis | |
import ( | |
- "reflect"; | |
+ "reflect" |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# logged in as admin account | |
# xcode 4 already installed | |
# java already installed | |
# standard place to install on OS X | |
cd /usr/share | |
# I like to keep sources around when building | |
# /etc/share/src doesn't exists, so just create it | |
sudo mkdir src |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# read $EHOME as the path to where eclipse is installed | |
# edit eclipse.ini | |
cd $EHOME/eclipse/Eclipse.app/Contents/MacOS/ | |
vim eclipse.ini | |
# add the following key/value to end of the ini file | |
# value can be anything including embedded spaces. | |
-Duser.name=Your Name <[email protected]> |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
old MBP (2 core) running Redis 2.2.6, the benchmark (in Eclipse) and The Clash on iTunes ;) | |
The bench loops and async INCRs a key 100,000 times and then blocks on the last INCR future to get the value, and emits the time for the 100000 INCRs. | |
Single threaded. The chunk pipeline actually saturates more fully with more threads. | |
Code in (chunk-pipeline) branch. This is the first cut so hope to push the performance a bit more, and definitely want to squeeze that jitter (must be the GC). | |
This is a grab of both runs from the same point in their respective runs. | |
--- new JRedisChunkedPipeline --- |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Enhanced ChunkedPipelineConnection. | |
Same micro-benchmark as (v.1 https://gist.github.com/1638321) -- (TalkingHeads playing, not The Clash) | |
Important to note that on a 2 core machine this pipeline does NOT utilize 100% CPU, and, it is mostly user-space. (Cores toggle (since JVM threads are unfortunately not fixed to a specific core) at ~80%/60% with Redis o/c serving the same on the same box.) | |
[commit: https://github.com/alphazero/jredis/commit/c86b68190d8a86ba779615cbf2d2ce5eed6d7149] | |
01327372586763 Mon Jan 23 21:36:26 EST 2012 JRedis[tid:10] <INFO>: counter: 24100000 msec:314 ops/msecs:318.000000 [q-delta:215428000] [delta:314742000] [sample latency:99314000] | |
01327372587081 Mon Jan 23 21:36:27 EST 2012 JRedis[tid:10] <INFO>: counter: 24200000 msec:317 ops/msecs:315.000000 [q-delta:224593000] [delta:317202000] [sample latency:92609000] |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Bench records metrics every every 10Mbits of data transfer (qitems are 8 byte byte[]) | |
Noteworthy: almost like clockwork, after 600 steps (e.g. 6Gbits) GC kicks in and the LinkedBlockingQueue starts to jitter: | |
-- begin -- | |
[00001] bytes:1250000 - delta: 88 msec - bps:113160574 - wps: 1768133 | |
[00002] bytes:1250000 - delta: 55 msec - bps:179937022 - wps: 2811515 | |
[00003] bytes:1250000 - delta: 60 msec - bps:166428119 - wps: 2600439 | |
[00004] bytes:1250000 - delta: 62 msec - bps:159948816 - wps: 2499200 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Same micro benchmark as the one used to test BlockingLinkedQueue, but using my modified implementation of Michael/Scott's 2Lock concurrent queue. | |
https://github.com/alphazero/jacomecha | |
Bench records metrics every every 10Mbits of data transfer (qitems are 8 byte byte[]) | |
Notes: | |
(1) Performance appears to be measurably better (e.g. ~ 292,921 8b word transfer/sec more) | |
(2) Doesn't exhibit the same jitter as BlockingLinkedQueue. LBQ drops to as low as | |
[00707] bytes:1250000 - delta: 347 msec - bps: 28802921 - wps: 450045 |