Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
ivom2gi
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
91.
▲
GC effect on throughput and latency explained
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
92.
▲
Java Collection overhead: stats from 500+ applications
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
93.
▲
Robinhood (backed by a16z) started offering free equity trading platform
(robinhood.io)
3 points
by
ivom2gi
13y ago
|
1 comments
94.
▲
Performance tuning legacy applications: picking low hanging fruits first
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
95.
▲
Throughput vs latency - explaining the optimization process
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
96.
▲
Scott Adams challenged over Elbonian/Estonian jokes
(blog.grabcad.com)
6 points
by
ivom2gi
13y ago
|
11 comments
97.
▲
Benchmarking Java Garbage Collection algorithms with JIRA load tests
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
98.
▲
Java stacktraces: friend or foe?
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
99.
▲
Stats from ~85K runtimes analyzing different Java Garbage Collector usage
(plumbr.eu)
2 points
by
ivom2gi
13y ago
|
0 comments
100.
▲
by
ivom2gi
13y ago
As covered in the company blog: http://plumbr.eu/blog/happy-to-announce-a-1m-seed-round
101.
▲
Side effects on HashMap implementation change
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
102.
▲
Beautiful design concept explained via Java ThreadLocal source
(plumbr.eu)
2 points
by
ivom2gi
13y ago
|
0 comments
103.
▲
Want to drive your engineering mad? Give them a dying app without a stacktrace.
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
104.
▲
Debugging a crashing app without the stacktrace
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
105.
▲
Demystifying ThreadLocal variables
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
106.
▲
Increasing throughput and reducing latency: Asynchronous Servlets
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
107.
▲
Benchmarking your Java agent against real world cases
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
108.
▲
Tackling memory leaks with machine learning
(plumbr.eu)
42 points
by
ivom2gi
13y ago
|
1 comments
109.
▲
Six Java APIs to avoid
(plumbr.eu)
7 points
by
ivom2gi
13y ago
|
0 comments
110.
▲
Why we dropped Eclipse in favour of IntelliJ
(plumbr.eu)
4 points
by
ivom2gi
13y ago
|
1 comments
111.
▲
Java Exceptions without stacktrace - find out why
(plumbr.eu)
3 points
by
ivom2gi
13y ago
|
0 comments
112.
▲
Finding two memory leaks in a single case study
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
113.
▲
Exception handling - a way for unmainatainable or slow code?
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
114.
▲
Why is your software aging
(plumbr.eu)
3 points
by
ivom2gi
13y ago
|
1 comments
115.
▲
Different Java OutOfMemoryErrors explained
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
116.
▲
Measuring and explaining the overhead of Java Collections
(plumbr.eu)
3 points
by
ivom2gi
13y ago
|
0 comments
117.
▲
Managing the complexity of the source code
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
118.
▲
Finding out who is calling my code in Java dynamically
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
119.
▲
Reducing memory consumption in a Java application via String.intern()
(plumbr.eu)
1 points
by
ivom2gi
13y ago
|
0 comments
120.
▲
Why does Java consume more memory than you allocate?
(plumbr.eu)
2 points
by
ivom2gi
13y ago
|
0 comments
More ›