5 ms·
I don't think the tuning you refer to is sufficient to bring the platform noise to levels required for some workloads. You will still see a lot of system call i
by bitcharmer 6y ago
I don't think the tuning you refer to is sufficient to bring the platform noise to levels required for some workloads.
You will still see a lot of system call interrupts, TLB shootdowns, timer events, etc.
Funnily enough yesterday I published an article tangential to this problem field.
http://bitcharmer.blogspot.com/2020/05/t_84.html http://bitcharmer.blogspot.com/2020/05/t_84.html
I'm not an expert on unikernels but my assumption is that you will see none of that OS jitter.
- xfs 6y agoGood writeup, but is there any mitigation for TLS shootdowns?
- ncmncm 6y agoTLB shootdowns are a product of multithreading, and of unmapping memory. Avoid one, the other, or both, and TLB shootdowns fade out of the picture. You still have cores to isolate, busybody kernel threads to suppress, and hardware interrupts to direct elsewhere, but TLB shootdown paranoia is largely a product of the current fashion favoring multi-threading over running separate processes with carefully chosen sharing.
- ncmncm 6y agoIf you have only one thread running in the process, memory usage is static, any mapped pages are writable only by the one process and are not backed by external storage, you have turned off timer interrupts and rcu reaps, and do not perform system calls, where are these TLB shootdowns and timer interrupts coming from?