On-demand dedicated GPU compute instances, powered by NVIDIA
Deploy your workloads faster, without any CAPEX investment
Scalable HPC infrastructure available on flexible monthly plans
End-to-end HPC environment: Compute, Storage, Network and Security
With the constant increase in data, compute and analysis of the same to return precise results require huge computing prowess. In layman terms, an HPC Infrastructure aggregates the power of multiple high-end processors, which is boosted with a GPU to provide quick and accurate results.
The image below showcases that a single server node with V100 GPUs can replace over 60 CPU. For example, for SPECFEM3D (Geodynamics software), a single node with four V100’s will do the work of 54 dual-socket CPU nodes. With lower networking, power, and rack space overheads, accelerated nodes provide higher application throughput at substantially reduced costs.
AI/ML are used commercially across industries for efficient and faster computations. Training machines to return accurate results require a vast amount of data. AI delivers better than average accuracy on complex tasks. However, training with these large scale and complex data sets are computationally intensive. This is where HPC supercharges AI and Data analytics to deliver data faster and gives your business an edge over others.
Financial services can use the power of artificial intelligence backed by High Performance Compute and take their businesses to new heights. Risk modelling, real-time fraud detection, high-frequency trading applications, customer engagement and more are increasingly powered by AI/ML.
|On-Premise HPC||Yotta HPCaaS|
|Upfront Investment in HPC Infrastructure||Requires large capital investment in GPU servers, storage, network, security etc. which could run into tens of millions of Rupees depending on the scale of your HPC infrastructure||Zero CAPEX investment with flexible pricing. Ready to use pre-provisioned HPC infrastructure including switching routing infrastructure, internet bandwidth, firewall, load balancer and intrusion protection system. Our monthly plans start from Rs. 11,000|
|Investment in Skill Sets||You need to hire expensive resources who can manage your HPC infrastructure||Our expert team manages the infrastructure maintenance, so that you can focus on your project|
|AI/ML/DL Libraries||Additional cost||Includes support from Yotta HPC Partner Ecosystem|
|Cost of Upgrades||Additional capital expense||Yotta handles upgrades and updates for optimum performance|
|Power Infrastructure||Unreliable Power at high cost||Fail-safe power infrastructure ensures 100% uptime|
|Scale & Elasticity||Scalability is an issue on-premises||Scale up or down as per your workloads|
|NVIDIA T4 Tensor Core GPUs||NVIDIA V100 Tensor Core GPUs||NVIDIA A100 GPUs|
|Most Advanced Inference Accelerator||Universal Data Centre GPU||Unprecedented acceleration at every scale|
|BEST SUITABLE APPLICATION OR USAGE||AI TRAINING /INFERENCE, DATA SCIENCES, HPC , GAMING||AI TRAINING, DEEP LEARNING|
DATA SCIENCE, HPC
|AI, DATA ANALYTICS, AND HPC|
|SINGLE PRECISION PERFORMANCE – FP32||8.1 TFLOPS||14 TFLOPS (PCIe)|
15.7 TFLOPS (SXM2)
|HALF PRECISION PERFORMANCE- FP16||65 TFLOPS||112 TFLOPS (PCIe)|
112 TFLOPS (SXM2)
|GPU MEMORY||16GB||32/16GB HBM2||40GB HBM2|
|MEMORY BANDWIDTH||320 GB/SEC||900 GB/SEC||1,555 GB/SEC|
|SYSTEM INTERFACE / FORM FACTOR||LOW PROFILE|
PCI EXPRESS FORM FACTOR
|DUAL SLOT, FULL HEIGHT|
PCI EXPRESS FORM FACTOR SXM2 / NVLINK
|POWER||70 WATT||250 WATT (PCIe)|
300 WATT (SXM2)