DataCrunch Instances offer a worldclass price/performance ratio.
Let's compare our instances to some well known competitors.
DataCrunch vs. Google Cloud
We compare the DataCrunch 8V100.48M with a mid tier 8x V100 Google cloud instance in order to offer a fair price comparison. In both cases the server is dedicated.
The price of the Google instance amounts to $21.36/h for 64 vCores and 120GB RAM, considerably higher than our pricing. DataCrunch offers instances at a better price than their preemptive instances, however the DataCrunch instances are in fact dedicated.
Our bare metal instance performs better across the board and delivers a whopping 319% better value for your money.
DataCrunch vs. AWS EC2
We compare the DataCrunch 8V100.48M bare metal instance vs. AWS' p3.16xlarge. The p3.16xlarge advertises an older Xeon E5-2686 from 2016, the 8V100.48M runs 2 Xeon 4214R's from 2020. Despite the lower core count, performance is an average 30% higher.
The AWS instance has a larger amount of RAM. According to our customers, the provided memory by DataCrunch is plenty for most usecases. The DataCrunch instance has 2TB NVMe storage, which greatly outperforms Amazon EBS, both in bandwidth as IOPS.
Since the GPU's are the same, we can conclude that the 8V100.48M is much better value for all but unique circumstances.
DataCrunch vs. Paperspace
We compare the DataCrunch 1V100.6V dedicated instance with Paperspace's 1x V100 instance.
Paperspace offers slightly higher CPU performance and system memory capacity but at over 3x the cost.
DataCrunch vs. LeaderGPU
We compare the DataCrunch 2V100.10V dedicated instance with the 2x V100 system from LeaderGPU.
The system from LeaderGPU offers a greater amount of CPU and memory capacity. The SSD in the system is SATA, which highly restricts IO and bandwidth compared to our NVMe drives.
The 2V100.10V offers a 118% better price if GPU performance is what matters.