Re: Approximate latency number for the pyramid of speed.

There’s also a graphical view along with a predictor of where the numbers converge to in future years: https://people.eecs.berkeley.edu/~rcs/research/interactive_latency.html Sent from my iPhone  > On 25 Sep 2017, at 14:18, Richard Warburton > wrote: > > Hi, >

Re: A simple function in Java and JITed version of it

> On 19 Nov 2017, at 16:32, John Hening wrote: > > I am analyzing the following Java code: > > class Main { > > private static class C{ > private C ref; > } > > > private static void foo(C c, C c2){ >c.ref = c2; > } > } >

Re: JITed very simple method.

> On 3 Feb 2018, at 23:31, John Hening wrote: > > public static int dontOptimize; > > > public static int simple(){ > return dontOptimize; > } > > And JITed version of that: > > > 0x7fdc75112aa0: mov%eax,-0x14000(%rsp) > 0x7fdc75112aa7: push %rbp >

Re: How Thread.currentThread() in java work behind the scene?

Every native operating system has some form of thread identifier which can be acquired by calling e.g. pthread_self(). The thread data structure stored in the thread’s local area has more information (e.g. the limits of the current tlab) and the associated Java thread object. For efficiency in

Re: Why are page faults used for safe point polling instead of other signals?

There’s a couple of overlapping questions here. I hope I can answer them, if not necessarily the right order. Reads are used rather than writes because reads don’t incur cross-CPU cache invalidations. If threads were writing to a page, then cache invalidation traffic would be sent between

Re: Why are page faults used for safe point polling instead of other signals?

> On 29 May 2020, at 17:46, Steven Stewart-Gallus > wrote: > > Okay I have an idea. > I can't shake the idea you could do fun tricks with thread local executable > pages. > > The theoretically fastest way of safepoint polling is inserting a trap > instruction. Under what basis are you