tonydiv

23 ms·

1. 151.▲
   by tonydiv 13y ago

   Here you go - http://www.eecs.berkeley.edu/~isabelle/papers/cci.pdf
2. 152.▲
   by tonydiv 13y ago

   Hey Idoh! I'm happy you saw our post, and thank you for the kind words. You're definitely right about the lesson with FB. I'm just bitter about it :)
3. 153.▲
   by tonydiv 13y ago

   We are working on that paper right now. If you're interested, feel free to provide us your email on the site.
4. 154.▲
   by tonydiv 13y ago

   We had paying users for GraphMuse, not Codentical. GraphMuse was shut down by Facebook, whereas we quit Codentical. As I stated, we had a good number of people "willing to pay for it," but they weren't paying customers yet. W
5. 155.▲
   by tonydiv 13y ago

   I'm not sure either! Either way, implementing a clustering algorithm on a GPU was thinking out of the box for us :)
6. 156.▲
   by tonydiv 13y ago

   We are not making up these numbers. Even for an IO-bound workload, we are seeing 350x speedup. The very fitting application is MapReduce, which is embarrassingly parallel.
7. 157.▲
   by tonydiv 13y ago

   We had less than 3 days to comply if I recall. Considering we had multiple customers in a handful of countries, it wasn't much time.
8. 158.▲
   by tonydiv 13y ago

   Correct. PCIe is the least of our concerns at 16GB/s. Some of our tests indicate that our GPU analyzes data at a rate of ~5GB/s. If we can feed it the data in that amount of time, we're in business. We're also paralleliz
9. 159.▲
   by tonydiv 13y ago

   "Do you actually leverage HDFS and the Hadoop scheduling infrastructure to run ParallelX jobs?" -- Yes. We could also provide what you're suggesting outside of Hadoop, no problem! "What are the limits like for input data
10. 160.▲
    by tonydiv 13y ago

    While I realize that an TOS is always subject to change, I still think it's a low blow to pull something like that. They intended it to be a one-way conversation as well. We had no way to respond. Facebook is notorious for doing this t
11. 161.▲
    by tonydiv 13y ago

    We've proven that we can also speed up IO-bound and disk-bound workloads. Memory-bound jobs are not an issue either. In fact, 12GB Nvidia and AMD GPUs were announced this week.
12. 162.▲
    by tonydiv 13y ago

    I can't thank you enough! There are a number of factors that will make this product successful: 1. The compiler needs to work well. 2. We need to teach average developers that they can build new applications/algorithms that were p
13. 163.▲
    by tonydiv 13y ago

    You're very welcome. I appreciate the kind words.
14. 164.▲
    From a Frat Social Network to a GPU Compiler for Hadoop(medium.com)

    118 points by tonydiv 13y ago|75 comments
15. 165.▲
    ParallelX compiles Hadoop Java code to OpenCL to run on GPUs(parallelx.com)

    3 points by tonydiv 13y ago|0 comments
16. 166.▲
    by tonydiv 13y ago

    Priceonomics should hire this guy. I wouldn't doubt he would dedicate himself to writing stories such at these.
17. 167.▲
    by tonydiv 13y ago

    Hmm, if there's no MR involved, I'm not sure our GPUs would be very helpful for this type of job. What do you think?
18. 168.▲
    by tonydiv 13y ago

    What types of algorithms would you implement? Would they be compute bound?
19. 169.▲
    by tonydiv 13y ago

    Our free tier will be: 26 CPU cores, 15 GB RAM, 60GB of SSD storage 1536 GPU cores, and 2GB RAM for a certain amount of time. If you'd like a reserve instance, or more power/storage/throughput, we'll charge more accordin
20. 170.▲
    by tonydiv 13y ago

    Could you provide me your preferred email? Is it the one in your HN profile?
21. 171.▲
    by tonydiv 13y ago

    What types of jobs are you trying to implement?
22. 172.▲
    by tonydiv 13y ago

    We've developed a parallel architecture using GPUs thereby reducing the cost by a few hundred factors. So, yes, we will have a freemium plan, and an affordable pricing model. Email me at tony@parallelx.com if you're interested.
23. 173.▲
    Ask HN: If a Hadoop cluster were free, what would you do?

    8 points by tonydiv 13y ago|14 comments
24. 174.▲
    by tonydiv 13y ago

    IF ONLY he had said, "One coin... to rule them all" at the very end! What a product, I love it, and I love this guy's quirkiness. Love. Love. Love.
25. 175.▲
    by tonydiv 13y ago

    5 views from the embeddable YT player for us.
26. 176.▲
    by tonydiv 13y ago

    What about the same 'trial' concept, except limit it to possibly 2 days, like a Friday and a Saturday. Assuming the company is small, it might even be possible to do Saturday/Sunday. I'd also let the interviewee choose w
27. 177.▲
    by tonydiv 13y ago

    For those referring to the social utility of Facebook, I recommend hiding your news feed. The newsfeed is not a social utility, whereas messages, profiles, looking up mutual friends, etc. IS a social utility. Here's how to hide your ne
28. 178.▲
    by tonydiv 13y ago

    Better performance might help this sell, but the second I saw the "Surface Music Kit," I got excited. As a DJ, I need to carry around an expensive laptop to perform. The iPad does not have any impressive apps to DJ (at least not p
29. 179.▲
    by tonydiv 13y ago

    This looks neat, I will definitely keep it in mind. I would be weary of using the Github Octocat mascot though. I believe Octocat is protected under copyright.
30. 180.▲
    by tonydiv 13y ago

    BIG thanks for not lying to me after I didn't perform any better reading the colored text. I would definitely consider showing the colored text first for some others, and second for others. Once I knew that you were going to ask questi