8x NVIDIA V100 Tensor Core GPUs
40,000 CUDA and 5,000 Tensor cores
896 TFLOPS of tensor performance
96 Intel Xeon Scalable Skylake CPUs
768 GiB CPU and 256GiB GPU RAM
100Gbps network bandwidth
1.76TiB NVMe SSD storage, 700K IOPSLearn More
2x Intel Xeon Scalable 8275CL (Cascade Lake)
96 parallel processing CPUs
4 TFLOPS of dual precision performance
768GiB DDR4 262GiB/s RAM
48MiB L2 and 66MiB L3 caches
100Gbps network bandwidth
3.5TiB NVMe SSD storage
1.4M IOPS and 5.3GiBps IO bandwidth
Kogence brings scientific supercomputing to a web browser of any personal device. Just select the supercompute infrastructure you desire and click run.It's really that easy! Scale-out by clustering these nodes together to create and run your own on-demand high-performance, low-latency compute grid within seconds. Create and pay for high-end infrastructure only when needed.
Our infrastructure is truly on-demand. There is no queueing or job scheduling. We auto-provision additional machines for the jobs as and when you need them. Add the power of parallel computing to your workflow. We offer private hosted solutions to our enterprise customers. Entire infrastructure operates from behind the corporate firewall.
Kogence platform is designed grounds up to be the grand central for all your organization's high performance computing needs. Kogence platform consists of a collaboration engine, a stack builder and an orchestration engine.
Stack builder helps you create your own arbitrary container stack and your hardware cluster. Orchestration engine consists of kComposition, kScaling and kMigration technology. These technologies automatically composes your multi-containers stack, launch and then automatically scale your cluster up and down as per the needs of your application.
On Kogence all software, simulators and solvers are deployed in independent individual docker containers. Platform is extensible and user can deploy their own containers. Users can build complex simulation workflows combining multiple software. Kogence' kComposition technology automatically links the selected containers such that solver binaries in one of the containers can be called from any of the containers selected in the workflow. There is no need to write composition scripts. At the same time, inorder to maintain secure computing environment, logged in user remains a non-privildged user both in the host machine as well as in any of the running containers.Learn More
Kogence clusters start with a single master node and zero worker nodes. Once you submit your workload, cluster scales up and down automatically as needed.
Cloud-native Kogence clusters are single user clusters. This greatly simplifies the management and avoid needing a dedicated IT team to manage your organization's multi-user multi-application clusters. Self scaling allows to achieve extremely high utilization compareable to onprem multi-user clusters while simplifying the management.
Kogence compute nodes are failure and preemption tolerant. All Kogence nodes are cost-effective infrastructure procured through real-time bidding on the spot market. In general such infrastructure can be preempted if another market participant is willing to pay a higher price. Kogence infrastructure if preempt tolerant and kMigration technology saves the state of you simulation, inform the scheduler about the pending move, moves your workload to new node and your simulation restarts from exact same state as before.Learn More
Kogence brings 100's of pre-installed software and a library of 1000's of functional simulation models that can be forked and executed within seconds. With Kogence you will use the least amount of time and resources to determine the preferred approach and simulation software for the problem at hand.
Got a new research challenge? Explore our Model Library before you make up your mind about the path forward. Protect your design team against very expensive detours and change of course at later stages of projects. Kogence Model Library is built by a team of PhDs from premier research labs and institutes.
When ideas come together, a unique clarity emerges leading to the development of better solutions in less time. Bringing unity to the team through success. Kogence allows global teams to work seamlessly with pooled resources and zero setup. No need to set up separate infrastructure or software for each member of your geographically distributed team.
Each team member can fork a simulation and iterate on it independently. The team admin controls access, sharing and resource permissions. Share a project wiki links with team members and managers for quick status updates.
At Kogence we take security very seriously. We offer a private hosted solution to our Enterprise customers. Enterprise workloads run on dedicated private infrastructure. Entire compute infrastructure operates inside a virtual private cloud (VPC) that can be placed behind the customer's corporate firewall and is only accessible from customer's corporate locations.
All data remains encrypted with keys that only customers retain and all data communication happens via encrypted tunnels of industry standard security protocols of SSH and SSL. Additional layers of security, administrative controls and authentication are in place and can be custom configured to meet your organization's security requirements.
All Kogence projects are version controlled in a git-LFS repository. Kogence projects come with a built-in WYSIWYG wiki that is automatically populated with latest results from simulations. Kogence platform also has a built-in code editor and latency free remote interactive desktop connection. Project admins have granular levels of permission controls.
Kogence brings 100's of pre-installed software and a library of 1000's of functional simulation models that can be forked and executed within seconds. With Kogence, you will use the least amount of time and resources to determine the right approach and the right simulation software for the problem. Insulate your design team from the high cost of directional changes at later stages of any project. With Kogence, the infrastructure is a snap and can save months in setup effort.
Think of the cost of your resources, the team members, the IT costs and most importantly the time it takes to acquire the technology stack they need to succeed. Time for additional software tooling, hardware vendor selection, licensing and acquisition for each product and eventually months of precious time wasted.