Aloha...yes we realize that this meeting is following a three day weekend for come lucky folks...Consider joining us either in person or virtually:…Continue
Started by Bonnie B. Hilory Feb 16, 2013.
Aloha...Bring you ideas for discussion...Lots of ways to rsvpMeet up:http://www.meetup.com/Salesforce-Hawaii-User-Group-HUG/or…Continue
Started by Bonnie B. Hilory Mar 18, 2012.
Here is the link to sign up for the salesforce user group.…Continue
Started by Bonnie B. Hilory. Last reply by Bonnie B. Hilory Aug 13, 2011.
Been using this easy on online office suit the usual business applications online for about 3 years and still havent tred all the features. But I have been using the spread sheet and word processing…Continue
Tags: apps, cloud, applications, online
Started by John Sydney Yamane. Last reply by Daniel Nishimura Jun 25, 2011.
Comment
March 2013 Salesforce User Group...come join us...
Free...including pastries and coffee!
Understood. If you do frequent requests and have a low latency requirement I can see why HPC on AWS would be suboptimal. If you do relatively infrequent requests for large jobs that require maximum computational power, its a great option. The AWS HPC 17,024 core 240.09 TeraFLOPS cc2.8xlarge cluster I mentioned was, by itself, the 42nd fastest machine on the planet.
Most of my researchers need low latency systems with LOTS of computing power, so for the most part HPC on AWS isn't a good fit. We're currently using LOTS of HPC clusters and are looking at consolidating them into one big cluster. However, VDI and traditional server virtualization is something we also need to do. So we're also looking at management systems that would allow us to flip back and forth between clouds and HPC.
Perhaps we are just using different terminology (or I'm being slow, which is entirely possible :-) AWS is Amazon's cloud. HPC on AWS gives you the ability to access banks of GPUs via a cloud model. Is this not what you're after?
We run HPC in a big way here....but I'm talking Clouds not HPC...VDI is definitely a direction we want to move in and RemoteFX with GPU in the server can solve a massive number of problems for us. I've got an engineering sample of the ThinLinx TLX1000 RemoteFX/HDX/RDP thin client. I've also got the software version on my mac and iPad.
Hi Brian - Have you considered HPC on AWS? Its cheap, monstrously fast, scalable and comes with nice management tools. You can choose the underlying processor architecture (including NVidia Tesla GPUs) in your instance configuration. They've achieved 240.09 TeraFLOPS on a cluster of cc2.8xlarge instances.
This would most likely be far cheaper than building and managing your own GPU cloud, although possibly not as fun :-)
Is anyone looking at building a private cloud? I'm pushing hard to bring up a private cloud at UHM SOEST primarily to support RemoteFX VDI for matlab and other scientific visualization applications where I need GPU resources.
Hi Bruce - I look forward to hearing more about your Azure project. We do almost all of our apps on EC2 and Heroku but we did one project on Azure for a finance customer. Its an interesting platform.
My company, Inovaware, is preparing to deploy a SaaS solution on the Microsoft Azure platform and I just spoke with the Microsoft Azure evangelist for Hawaii and he's interested in getting connected with other companies in Hawaii looking to deploy on Azure. If you or someone you know is working on Azure let me know and I can hook you up.
© 2024 Created by Daniel Leuck. Powered by
You need to be a member of Cloud Computing to add comments!