×

Announcing: Slashdot Deals - Explore geek apps, games, gadgets and more. (what is this?)

Thank you!

We are sorry to see you leave - Beta is different and we value the time you took to try it out. Before you decide to go, please take a look at some value-adds for Beta and learn more about it. Thank you for reading Slashdot, and for making the site better!

Comments

top

Supercomputers' Growing Resilience Problems

fintler Re:Old problem (112 comments)

You can't checkpoint jobs at this scale. It will take longer to checkpoint a job then to compute an answer. This is further compounded when the job takes several months to run. A 1000 node cluster is very tiny compared to the scale they're talking about.

about 2 years ago
top

Supercomputers' Growing Resilience Problems

fintler Re:ummm, no. (112 comments)

Google is having the same problems that this article describes -- they haven't fixed it either.

If your problem domain can always be broken down into map-reduce, you can easily solve it with a hadoop-like environment to get fault tolerance. If your application falls outside of map-reduce (the applications this article is referring to), you need to start duplicating state (very expensive on systems of this scale) to recover from failures.

more than 2 years ago
top

Supercomputers' Growing Resilience Problems

fintler Re:Old problem (112 comments)

How do you give the work to another node when the failed node contains the only copy of its state (like in an MPI job)? Duplicating the state on multiple nodes is way too expensive.

more than 2 years ago
top

Supercomputers' Growing Resilience Problems

fintler Re:"and they halt operations when they do so" (112 comments)

Checkpoints will probably stick around for quite some time, but the model will need to change. Rather than serializing everything all the way down to a parallel filesystem, the data could potentially be checkpointed to a burst buffer (assuming a per-node design) or a nearby node (experimental SCR design). Of course, it's correct that even this won't scale to larger systems.

I think we'll probably have problems with getting data out to the nodes of the cluster before we start running into problems with checkpointing. The typical NFS home directory isn't going to scale. We'll need to switch over to something like udsl projections or another IO forwarding layer in the near future.

more than 2 years ago
top

Cray Unveils Its First GPU Supercomputer

fintler Re:Into the Realm? (76 comments)

ANL's Mira is going to be roughly half as fast as LLNL's Sequoia.

more than 3 years ago
top

"Mythical Man-Month" Supposedly Busted By MIT Startup

fintler Re:10 years ago (231 comments)

If the developer is schizophrenic, then it might be n^2, otherwise it's probably (n(n-1))/2.

more than 4 years ago
top

What Computer Science Can Teach Economics

fintler So, this economist and... (421 comments)

So, this economist and a computer scientist are sitting at a bar.... and these 5 girls walk in....

more than 5 years ago
top

Windows Is Dead – Long Live Midori?

fintler Re:TLA conflict (695 comments)

There's a reason for that edit button on top of the Wikipedia article.

more than 6 years ago

Submissions

top

New Mexico driver's license soon will not be valid for air travel

fintler fintler writes  |  about 2 years ago

fintler (140604) writes "Due to the expected enforcement of the REAL ID Act on Jan. 15, 2013, New Mexico travelers will need to be prepared to have TSA approved identification to be allowed to board a commercial aircraft. New Mexico driver's licenses are not valid valid identification after the law is in effect."
Link to Original Source
top

Keccak is the winner of NIST's SHA-3 competition

fintler fintler writes  |  more than 2 years ago

fintler writes "The National Institute of Standards and Technology (NIST) is pleased to announce the selection of Keccak as the winner of the SHA-3 Cryptographic Hash Algorithm Competition and the new SHA-3 hash algorithm. Keccak was designed by a team of cryptographers from Belgium and Italy, they are:

* Guido Bertoni (Italy) of STMicroelectronics,
* Joan Daemen (Belgium) of STMicroelectronics,
* Michaël Peeters (Belgium) of NXP Semiconductors, and
* Gilles Van Assche (Belgium) of STMicroelectronics.

NIST formally announced the SHA-3 competition in 2007 with an open call for the submission of candidate hash algorithms, and received 64 submissions from cryptographers around the world. In an ongoing review process, including two open conferences, the cryptographic community provided an enormous amount of expert feedback, andNIST winnowed the original 64 candidates down to the five finalist candidates – BLAKE, Grøstl, JH, Keccak and Skein. These finalists were further reviewed in a third public conference in March 2012.

NIST chose Keccak over the four other excellent finalists for its elegant design, large security margin, good general performance, excellent efficiency in hardware implementations, and for its flexibility. Keccak uses a new “sponge construction” chaining mode, based on a fixed permutation, that can readily be adjusted to trade generic security strength for throughput, and can generate larger or smaller hash outputs as required. The Keccak designers have also defined a modified chaining mode for Keccak that provides authenticated encryption. Additionally, Keccak complements the existing SHA-2 family of hash algorithms well. NIST remains confident in the security of SHA-2 which is now widely implemented, and the SHA-2 hash algorithms will continue to be used for the foreseeable future, as indicated in the NIST hash policy statement. One benefit that Keccak offers as the SHA-3 winner is its difference in design and implementation properties from that of SHA-2. It seems very unlikely that a single new cryptanalytic attack or approach could threaten both algorithms. Similarly, the very different implementation properties of the two algorithms will allow future application and protocol designers greater flexibility infinding one of the two hash algorithms that fits well with their requirements. NIST thanks the many people in companies, universities, laboratories and organizations around the world that participated in and contributed to the SHA-3 competition, especially the submitters of all the candidate algorithms, and the many others who contributed expert cryptanalysis, and performance studies. NIST could not have done the competition without them.

A detailed report of the final round of the competition will be published in the near future. Information about the SHA-3 competition is available at: http://www.nist.gov/hash-competition"

Link to Original Source
top

Radio Station Prank to Datamine Email Addresses

fintler fintler writes  |  more than 5 years ago

fintler writes "A Philadelphia radio station, The Beat (100.3), ran a contest on April fools day and managed to increase their e-mail database by around 2000 people. At the end of the day, the contest turned out to be a fake. According to FCC regulations, "[It is required] that a licensee that broadcasts or advertises information about a contest that it conducts shall fully and accurately disclose the material terms of the contest, and shall conduct the contest substantially as announced or advertised. No contest description shall be false, misleading or deceptive with respect to any material term". So, what should happen now?"
Link to Original Source
top

Work Related Websites Blocked At Work

fintler fintler writes  |  more than 4 years ago

fintler writes "I work for a rather large fortune 50 company that does software development and general scientific research worldwide. My job is as a software engineer in a CMMI Lvl 5 facility. This company has a policy of blocking websites. Of course, there are rules applied to block websites that are obviously NSFW. I have no problem with that. It's their network, they can block whatever they think is morally objectionable. The problem is, there's no process to have a website reviewed for being unblocked. There are many sites blocked that are directly relevant to work. For example, some documentation on http://www.w3.org/ is blocked, even though we have a web-tier to maintain and develop. All support tickets placed with the company-wide help desk go unanswered and there doesn't seem to be any process setup to have websites unblocked. Even my managers are frustrated by this. My current solution is to just put off working on things until I can go home and look up something. Some here have tried questionable methods like ssh tunneling, but considering some of the work we do, the network is highly monitored and persistent connections to suspicious IP addresses are monitored (one of my coworkers had some security people interview him for doing exactly this). As someone in my position, how would you handle looking up topics that you need to research for software development?"
top

Shuttleworth Supports OVC

fintler fintler writes  |  more than 6 years ago

fintler writes "The Open Voting Consortium, is a not-for-profit organization dedicated to the development, maintenance, and delivery of trustful and open voting systems for use in public elections. Today, they have received a donation from Mark Shuttleworth for $5,000 as well as permission to use his name for endorsement purposes."
Link to Original Source
top

VCS for a University CS Department

fintler fintler writes  |  more than 7 years ago

fintler writes "I've been slated to create a version control system for my University as an independent study next semester. The system would cater to only Computer Science and Information Science students (about 100-200 people). Can anyone give some advice on what kind of services should be provided and how they should be implemented? I'm currently thinking along the line of having something like a gForge and svn setup. However, I've also been thinking about doing a system using trac that has a bit of a custom "project" creation website. Any advice would be appreciated."
Link to Original Source

Journals

fintler has no journal entries.

Slashdot Login

Need an Account?

Forgot your password?