HN user

eerpini

160 karma

http://blog.satisheerpini.net

Posts34
Comments18
View on HN
www.awayfromshor.com 12y ago

Backpacking across the Middle East, Eastern Europe and South East Asia

eerpini
1pts0
www.khoslaventures.com 12y ago

Open Letter to 60 Minutes and CBS

eerpini
1pts0
blog.allotrop.com 13y ago

Shoot down the stars

eerpini
8pts2
tginnovations.wordpress.com 13y ago

Sleep deprivation and desynchronised brain activity

eerpini
1pts0
seattletimes.com 13y ago

Amazon buys campus for $1 Billion

eerpini
2pts0
blog.satisheerpini.net 14y ago

Session(login) logging in linux explained.

eerpini
1pts0
lwn.net 14y ago

LLVM and Linux

eerpini
2pts0
sayspy.blogspot.com 14y ago

Playing with the Ninja build system

eerpini
2pts0
blog.satisheerpini.net 14y ago

Memory at each node on a NUMA machine using libnuma (linux only)

eerpini
1pts0
blog.satisheerpini.net 14y ago

Quick everyday scripting in bash

eerpini
2pts0
krishnakanthmallikc.blogspot.com 14y ago

Using SIMD for hardware acceleration

eerpini
9pts2
03240c9.netsolhost.com 14y ago

Top 5 Reasons Do-It-Yourself (DIY) Analytics for Hadoop Doesn’t Make Sense

eerpini
1pts0
blog.computationalcomplexity.org 14y ago

Is 99.8% secure secure?

eerpini
15pts2
codeanand.com 14y ago

Good code and an ideal couple.

eerpini
2pts0
www.reghardware.com 14y ago

Court defies Apple demand to ban Samsung tablet

eerpini
1pts0
www.hpcinthecloud.com 14y ago

Cloud computing for Spies?

eerpini
1pts0
www.cs.purdue.edu 14y ago

The identity crisis in Computer Science

eerpini
5pts2
dustin.github.com 14y ago

Scaling memcached with vbuckets

eerpini
1pts0
blog.satisheerpini.net 14y ago

Visualizing directories and files as a graph

eerpini
1pts0
timemachine.gigapan.org 15y ago

The closest thing to a time machine

eerpini
2pts0
www.tomshardware.com 15y ago

IBM to build 10 PFlops supercomputer

eerpini
3pts0
www.sciencemag.org 15y ago

Screen Savers of the world unite

eerpini
1pts0
blog.dispatched.ch 15y ago

Vim as a python IDE : some neat plugins and TIPS

eerpini
4pts1
www.ibm.com 15y ago

Customizing the vi and Vim editors

eerpini
3pts0
lwn.net 15y ago

TTY-based group scheduling (the 200+ line linux patch that does wonders)

eerpini
17pts3
kotaku.com 15y ago

Kinect Hacked to Track Breasts, Or At Least Man-Boobs

eerpini
1pts0
www.ibm.com 15y ago

Developing custom plugins for the Vim editor

eerpini
1pts0
www.blackducksoftware.com 15y ago

Analysis of Open Source Projects for the Cloud (very informative)

eerpini
1pts0
0pointer.de 15y ago

Systemd for administrators : killing services

eerpini
2pts0
www.pcworld.com 15y ago

4 reasons to give Julia(mint Linux 10) a try

eerpini
1pts0

There is an animation further down that shows the magnetic field generated by the sun when it is a dipole. Apparently the 3-d wave like pattern better shields from cosmic rays originating outside the solar system.

I am not sure if DoorDash's IPO price is more than it should be, but value creation (which is "real" in my view) need not always result in immediate profits and the stock market tends to be forward looking in general. How exactly are Apple, Google, Tesla and for that matter even Uber vaporware companies ? Some of them might be overvalued (by a large margin), but they still have real value in this world.

Yes memory access patterns seem to be the most common bottleneck for most parallel code. I was implementing a parallel version of quick sort recently and I have similar stories to tell. Optimize the code to avoid cache misses frequently and you end up getting a near optimal speedup.

Well my friend, here are a few things you should know: 1. If you are interested in research, you should approach professors/find out what they are doing etc, even at the worst of schools, there is always someone doing something worthwhile. 2. As it looks like you are already following hacker news, keeping up to the posts on the front page itself should give you a lot of ideas and things to tinker around with. 3. Believe me, undergrad education is useless, everyone does it to get a degree. What really matters is the environment around. The people around you will help you learn more at school than the course itself does. 4. Ask yourself these simple questions, if the answer is a "no", try working towards making it a "yes" : Do you have a GITHUB Account ? Have you tried getting your hands dirty with any OSS project ? (not necessarily code, configuration ? testing ?) Do you follow journals in your area of Interest ? Which courses do you love at school ? Have you supplemented the course with some extra reading ?

These are just a few questions that can get you started, for example, the course for algorithms/data structures at most schools in india sucks big time, why not follow it up with a reading of AOCP or the MIT algorithms book. And yes as someone already mentioned, concentrating on work rather than cribbing about the system will take you better places. BTW, I did my undergrad from a so called "reputed private university" in India too.

The performance is measured with LinPack which mostly does FP operations and is optimized to run on the architecture, when normal programs are run, it is tough to run them at full utilization, this is because, a lot of time might be spent in data transfer or I/O apart from the computation ... there would be various other factors !

mostly the problems with utilizing the system completely are the same ones a common programmer faces when trying to use GPUs for general purpose programming, the problem of transferring the data to be handled from the main memory to the GPU's memory. Though Tianhe-1A has a very high peak performance, the sustained performance seems to be comparatively low !

well that is easier said than done, all the machines on the amazon cluster are machines with common architecture and traditional interconnects, most of these supercomputers have a custom architecture( here I mean just the type of processors and the inter processor connect on the same node) and interconnect, the cost of running programs on these machines would outweigh their usability for the cloud.

yes, they seem to be using the Power 7 architecture by IBM , which is quite energy efficient compared to most other architectures, and also they are planning to use UPC(Unified Parallel C) or similar languages for applications running on the system, that a good step towards better usability. But that does not matter as a super computer is always(mostly) used by a small group of highly qualified scientists.