Each compute element consists of 2 vCPUs and 4GB of memory.
Jobs can be cofigured with anywhere from 1 to 24 compute elements.
for Nivida K40 or K80 GPUs
Jobs can be configured with up to 2 K40s or 16 K80s.
example configuration pricing
|1||2||4||—||$0.01||90% cheaper than an AWS c4.large instance*|
|8||16||32||—||$0.08||90% cheaper than an AWS c4.4xlarge instance*|
|16||32||64||1 K80||$0.26||71% cheaper than an AWS p2.xlarge instance*|
|*All pricing comparisions as of 2/24/2021|
Quality-of-Service (QoS) discounts
Not in a rush? Get a discount on your hourly rate by allowing us to pause and reschedule your job at times to balance the workload in our compute grid.
Your job will be running at least 90% of the time during any particular period.
Your job will be running approximately 50% of the time during any particular period.
Your job will be running approximately 25% of the time during any particular period.
Your job will be running only when there are compute resources available that can't be allocated to jobs at a higher QoS.
Currently all Lancium Compute jobs run at High QOS regardless of the level chosen.
Advance purchase discounts
Planning on running a lot of jobs? Lancium Compute offers advance purchase options that provide a discount off the on-demand hourly rates for long-term volume commitments. Contact our sales team for more information.