|
From: | James |
Subject: | Re: state of the ’Pond for earnest tadpoles |
Date: | Sun, 3 Jan 2021 10:07:24 +0000 |
User-agent: | Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101 Thunderbird/78.6.0 |
Hello On 02/01/2021 16:58, Thomas Morley wrote:
Am Sa., 2. Jan. 2021 um 14:41 Uhr schrieb James <pkx166h@posteo.net>:On 02/01/2021 12:20, Thomas Morley wrote:A full `make doc` takes hours for me, even if invoked with `make doc -j5 CPU_COUNT=5` Thus I hardly do so, but use the CG-documented methods:Hours? Really? Perhaps 'an hour' if you were using some very, very old CPU - but even using a single CPU on an 'old' i5 Intel system a full make doc for me took less than 50 mins. That last time it took longer than an hour was when I had an old (8+ years ago) iMac running make doc in a linux VM. Jamestime make doc -j5 CPU_COUNT=5 -> real 77m53,168s user 204m51,501s sys 28m57,342s Ok, hours was not exactly correct, but significant more than one hour Cheers, Harm
I don't mean to flog this dead horse, but I was curious. I did a make doc with one cpu on my laptop (it has an i7 7700HQ CPU - circa 2017) and while I don't have the exact numbers to hand, as I am currently at work, it was something like real: 45m user:55m (cannot recall sys values), my desktop before that was from 2013 (i5 something) and I was getting sub hour make doc times even then but using 3 CPUs.
So you having those timings while using a -j5 option ... wow!I am obviously inhabiting some technological bubble that I wasn't aware I was in.
:D James
[Prev in Thread] | Current Thread | [Next in Thread] |