I'm the product manager for E2 VMs. Appreciate your blog post on this topic, however, we're having trouble replicating your benchmark results (E2 VMs show similar performance to N1 in our tests with your benchmarks), I'm wondering if there is something different about your environment that might be causing the discrepancy.
Do you have an email address where we can ask you follow-up questions? Please send it to gce-e2-feedback@google.com, to be able to discuss in detail.
Hi Eugene,
I'm the product manager for E2 VMs. Appreciate your blog post on this topic, however, we're having trouble replicating your benchmark results (E2 VMs show similar performance to N1 in our tests with your benchmarks), I'm wondering if there is something different about your environment that might be causing the discrepancy.
Do you have an email address where we can ask you follow-up questions? Please send it to gce-e2-feedback@google.com, to be able to discuss in detail.
Thanks!
Ari Liberman, PM for GCE
I performed the same sysbench and redis benchmarks on
in the GCP region/zones europe-west-1d and europe-west-1b (Belgium).
I couldn't see any significant difference between the E2 and N1 machine types in any of the benchmarks.
Here are some results from the sysbench-cpu benchmark (total time in seconds):
sysbench --test=cpu --cpu-max-prime=50000 --num-threads=4 run | grep "total time:"
e2-standard-4: 27.3814 ; 29.0544; 28.9985; 27.3325; 29.2564; 27.2810; 29.1408
n1-standard-4: 27.7668 ; 27.6391; 27.3915; 27.3113; 27.3933; 27.6044; 27.3655
The few CPU benchmark tests are showing some slightly better values for N1.
Please be aware that you might get other results. I tested only on two days in one region.