45 were found based on your criteria

  • Hourly – Less than 1 month – 10-30 hrs/week – Posted
    Create a simple application that can be tuned from memory bound to compute bound and that runs on both multi-core CPU and NVIDIA GPUs. Optimize for both platforms and demonstrate performance differences measuring flops and percent of peak bandwidth achieved. Experience with CUDA and C and a deep understanding of the performance impact of multi-level cache, memory placement and data alignment on both platforms are necessary.
  • Hourly – 3 to 6 months – 30+ hrs/week – Posted
    i) Build and profile image encoder/decoder software to identify bottlenecks. ii) Optimize scalar code and parallelize with threads. iii) Create linkable library from application. iv) Port to accelerator e.g. Intel Phi or NVIDIA GPU
  • Hourly – Less than 1 week – Less than 10 hrs/week – Posted
    I seek a CUDA expert that is familiar with matrix operations (preferably have experience in signal or image processing) to complete my CUDA code on C++ on a Mac platform. The remaining parts of the program to be coded on CUDA are: adding random Poisson noise on the existing matrix writing an efficient and fast 2D convolution function , writing an iterative deconvolution (which basically consists of several 2D convolutions and some matrix multiplication) finding average / variance across matrices (The attached ...
  • Hourly – 1 to 3 months – Less than 10 hrs/week – Posted
    Hello there, We are a computer vision company which is using server side CUDA processing of videos to recreate 3D scenes. We are trying to find errors in our methods and code. We are in need of an engineer / scientist who knows CUDA to do some unit testing and validation. Ideal candidate will have computer vision experience. Seeking all CUDA developers!!
  • Fixed-Price – Est. Budget: $100.00 Posted
    I require a program to be created in C++ and CUDA. The program will use the VLC-Qt library ( This library uses the Qt library, which is a cross-platform library, and the VLClib library for managing video. The program should use only unmanaged code, and be usable on both Windows and Linux operating systems. The program will synchronize many videos which play in different tiles simultaneously. That is, all videos should ...
  • Fixed-Price – Est. Budget: $75.00 Posted
    I require a program to be created in C# and CUDA. I am an experienced C# programmer, but have no experience with CUDA or GPUs. You may have to walk me through some basics, as to whether I'd need to obtain an SDK from NVIDIA for managed CUDA, etc. You may even need to help me set up my GPU. The program will synchronize many videos which play in different tiles simultaneously. That is, all videos should play evenly ...
  • Fixed-Price – Est. Budget: $5,000.00 Posted
    The Software must be able to detect and overlay in realtime alternative graphics/animations to replace preset/specified graphics, logos & animations. These will originate from an original TV HDMI source which will be captured within a PC ( Client machine ) Once a certain Logo or Animation is detected it shall be replaced with custom graphics and animations. Client Spec PC Hardware : - AMD Quad core 3.8ghz, 8gb DDR Ram , 60GB SSD drive , Avermedia dark crystal SDK capture card , Micro ATX Asus ...
  • Hourly – Less than 1 month – 10-30 hrs/week – Posted
    The role will be to work alongside existing team members in building a demo of a DNN which implements a NNLM to understand text from social media. You need a good understanding of CUDA, C, C++, DNNs and NNLMs. Due to the emerging nature of the work, some learning is expected but you need to have a solid understanding of the core technologies.
  • Fixed-Price – Est. Budget: $60.00 Posted
    Install CUDA and CUBLAS on a 64 bit Ubuntu 11.04 machine and make R package gputools work to use CUBLAS. Write down every step you took. You will be granted SSH access to the machine. Here is the URL for GPU tools: This may be a start for future Linux admin stuff for our company. So although not required for this job, EC2 sklis are a plus. Proficient in CUDA ...