Browse categories
Explore
Fiverr Pro
English
$
USD
I will do GPU performance acceleration for you with opencl or cuda
Turgut B
About this gig
I will accelerate your code using OpenCL and CUDA. I can increase the performance of algorithms and operations including, but not limited to, AI and ML programmes. I will write kernels for algorithms that you can use anywhere in your programs to increase performance by utilizing the GPU's performance.
The kernels are cross-platform and cross-vendor so any OS and any GPU will work with the kernels. I will also help with setting up and integrating OpenCL or CUDA on your system to get it working. I will also analyze your code to make sure the GPU actually increases performance and doesn't cause a bottleneck (sometimes it's slower to use the GPU for certain algorithms).
Included with the service
- Code analysis for performance (Standard package and up)
- Memory optimization
- OpenCL and CUDA kernels for C/C++
- Profiling
- For the Premium packet I will create a scalable GPU system tailored for your application and integrate into your project for future use
Get to know Turgut B
Turgut B
Quick developer and translator for all needs
- FromTurkey
- Member sinceDec 2020
- Avg. response time1 hour
Languages
English, Turkish, Spanish
With 5 years of professional coding experience and a strong background in software design, development, and deployment. I also ran my own software company for a year, delivering end-to-end solutions for clients across various industries. Skilled in building efficient, reliable, and scalable systems tailored to client needs. I value clear communication, clean code, and timely delivery :)
Languages I know:
- English (Native)
- Turkish (Native)
- Spanish (Basic)
FAQ
Will everything become faster with GPU?
No. Not every code or algorithm gets faster with GPU right away. Algorithms that can benefit from parallel cores get faster.
What is a kernel and how can I use it in my code?
It is basically a function that is ran on the GPU and runs on parallel inside the GPU cores. You can call it in your code base when an algorithm needs to get faster.
How will this improve my AI / ML code?
Neural networks, the underlying algorithm behind AI and ML, can benefit heavily from parallel computing. Also many statistical analysis algorithms benefit heavily from GPU acceleration.
Can you do device specific code?
Only for NVIDIA GPU's with CUDA. OpenCL allows for cross-device coding but I cannot do device specific nuanced code for other vendors.
