First, some background on what I am doing. I am currently running the latest release of Truenas Scale (22.02.0.1) on my home NAS/VM box. I am running Flux node software within a Ubuntu Server 20.04.4 VM that requires a minimum system spec to operate. The current spec requirements for the node are a 4 core 8 thread CPU, 32GB of ram, and 440GB SSD storage. The node software uses Sysbench to measure single thread CPU performance to make sure you meet the minimum spec of 640 events per second. I currently have 4 cores and 8 threads of the E5-2690v2 CPU passed through to the VM along with the 32GB RAM and 445GB of storage on a NVME drive.
Running Sysbench within the VM gives some really bad single thread CPU benchmarks. My E5-2690v2 at 3GHz is scoring just 360 EPS. I have the same software running on a Raspberry Pi 4 and it scores 625 EPS with it's measly 1.5GHz CPU. Double the speed... I have another machine running Proxmox with a Ryzen 5700g CPU and it scores 3500 EPS (obviously a much faster CPU).
I can't imagine a Raspberry Pi 4 CPU is twice as fast as a E5-2690v2, even with the Xeon being 10 years old. I have tried different options for the CPU mode and model in the VM setup such as Custom/Ivybridge, Custom/kvm64, Host Model, and Host Passthrough. I also tried setting up the CPU as 8 cores with 1 thread each as well as 4 cores with 2 threads each. All of the settings changes show no difference in performance. Is there something else that I could be missing here or is the Pi really that much faster? Any help would be appreciated. Thanks!
Running Sysbench within the VM gives some really bad single thread CPU benchmarks. My E5-2690v2 at 3GHz is scoring just 360 EPS. I have the same software running on a Raspberry Pi 4 and it scores 625 EPS with it's measly 1.5GHz CPU. Double the speed... I have another machine running Proxmox with a Ryzen 5700g CPU and it scores 3500 EPS (obviously a much faster CPU).
I can't imagine a Raspberry Pi 4 CPU is twice as fast as a E5-2690v2, even with the Xeon being 10 years old. I have tried different options for the CPU mode and model in the VM setup such as Custom/Ivybridge, Custom/kvm64, Host Model, and Host Passthrough. I also tried setting up the CPU as 8 cores with 1 thread each as well as 4 cores with 2 threads each. All of the settings changes show no difference in performance. Is there something else that I could be missing here or is the Pi really that much faster? Any help would be appreciated. Thanks!
