Cloud Computing > (주)코코링크 CoCoLink corp.

Cloud Computing
Cloud Computing Service
We're currently providing a small service.
It provides cloud computing based on HPC hardware and software that it produces, and further provides data processing services.
Delivers a range of cloud services based on high-performance, in-house development servers.

Cloud Computing Supply
We are leasing computing resources for scientific engineering calculations, 3D rendering, etc.
We have over 80 nodes of 3,000-core CPU-based computing systems, and we continue to scale with over 200 GPUs.
It offers applications such as Luxol OLED, Luxol OGX, and Luxol VFX, as well as a variety of commercial programs.
It also provides system deployment advice for economic growth.

Retention system and purpose
CPU HPC Cluster : Ideal for simulation with scientific and engineering applications.
GPU HPCs : Ideal for tasks that require high efficiency, such as artificial intelligence (AI) and rendering.
GPU HPC Cluster : It is suitable for industries that require extensive data processing and accurate output, such as oil exploration.

HPC Cluster Service
Build ~10PF systems based on the Klimax PetaCube to enhance various supercomputing infrastructures.
Migrate to a variety of GPUs, if necessary, to provide the highest cost efficiency.

Code Migration/Optimization
The in-house code or open source code, written with the existing Fortran, Python, Matlab, is transplanted into C language and is transferred to CUDA or HIP to speed it up.
In this case, performance gains of between 100 and 2,000 times over the CPU core are possible, allowing a large number of solvers to be processed on a single node and achieving very high computational efficiency.

CPUCluster (Based on 2018.08)
XeonServer Dual Xeon Processor 36
Xeon PHI Server Xeon Phi Processor(68 core, 2.6GHz) 44
Interconnect Infiniband EDR+ Gbe
Storage Locally 1TB
Operating System MPI &Compilers Linux OpenMPI,gcc, PGI Compilers, Intel Compilers
AI Cloud
This service server platform for deep learning based artificial intelligence research provides various frameworks such as Caffe and SensorFlow.
Optimized systems with 20 GPUs can help maximize utilization.

Various GPUs Support
Nvidia's Tesla GPU as well as a range of cost-effective GPUs can be applied to reduce user cost burden. This maximizes the stability of the business by securing a competitive advantage.

AICluster (Based on 2018.08)
Compute Server Klimax-210 4
GPU GTX 1080Ti 64
Interconnect Infiniband EDR+ Gbe
Storage Locally 1TB SSD + 10TB HDD
Operating System MPI &Compilers Linux OpenMPI,gcc, CUDA
Rendering Cloud
Delivers rendering cloud services that support a variety of 3D rendering environments, including Maya, Cinema4D, Redshift, and more.
The Klimax-R10 with 10 GPUs is the platform to provide competitive services.

Maximize cost efficiency with low-cost hardware
Using Gaming GPU, hardware costs are minimized.
High-efficiency, low-cost systems based on more economical Radeon 7 GPU, etc.

Render-farm (Based on 2018.08)
Compute Server Klimax-R10 8
CPU GTX 1080Ti 80
Interconnect Gigabit Ethernet
Storage Locally 1TB  SSD + 5TB HDD
Operating System MS Windows Redshift, Octane, etc.