The Official Status Thread
-
@Applied-Mediocrity said in The Official Status Thread:
volatile
is what disables reordering/removal optimizations, yes?Yes. It basically means "reading or writing this variable may have non-obvious side effects", so it forbids most optimizations.
Nowadays it's only relevant for embedded targets, since otherwise you typically don't access hardware directly. And on modern machines,
volatile
is no longer a reliable way to implement inter-thread communication, now that processors may reorder memory accesses.
-
@Zerosquare But they don't reorder them when there's
volatile
, do they?
-
This post is deleted!
-
@pie_flavor said in The Official Status Thread:
@Zerosquare But they don't reorder them when there's
volatile
, do they?volatile
is visible to C only, the processor doesn't see yourvolatile
declaration. Its meaning is basically: contents of this address can spontaneously change, so don't try to cache reads you assume didn't change.
It does not cause memory fence instructions to be emitted, so it's not usable for synchronization (in C/C++).
-
Status: Hearing plenty of fireworks tonight and the neighborhood just lost power. Assuming with high confidence the two things are related.
-
@topspin said in The Official Status Thread:
@pie_flavor said in The Official Status Thread:
@Zerosquare But they don't reorder them when there's
volatile
, do they?volatile
is visible to C only, the processor doesn't see yourvolatile
declaration. Its meaning is basically: contents of this address can spontaneously change, so don't try to cache reads you assume didn't change.
It does not cause memory fence instructions to be emitted, so it's not usable for synchronization (in C/C++).Knowing what it actually means depends on knowing the architecture well. For memory-mapped hardware, the accesses typically happen after address translation (if you have it; lots of embedded systems don't) but before any caches or write-combining hardware. In those situations, you could theoretically use that memory for synchronization… but it's a spectacularly bad idea to do so because both reads and writes of that space probably have special behaviours attached. They may also be slower than normal memory, especially to read.
An example of the latter is when you've got some piece of fancy memory mapped math accelerator in hardware (e.g., we've got a design for a fast exponentiation and logarithm module for our next-gen chips). It doesn't have any extensions added to the ARM instruction set, so to run it, you write the argument to one memory-mapped register and read the result out of another; if result isn't ready yet (complex ops take a bit!), the read stalls until the computation completes; it's way faster than doing it in software but very much not instant so a stall is all we can do.
You really wouldn't want to use that for synchronization! (We have a different bit of memory for that that's optimised for that use case.)
-
Status: Fucking hell, for a microsecond I thought @dkf had joined the BABSCon server...
-
@dkf said in The Official Status Thread:
Code works fine on my machine, but breaks (with unexpected timeouts, and no useful information about why they happened) when on our integration test infrastructure.
It was the NAT on the docker container. Once I fixed the UDP socket reprogramming so that it uses the information that the other side sees, it all works. (I thought I'd already done that, but the remote OS has quite a lot of “eccentric” interpretations of the instructions it is sent that mean that I had to not just stand on one leg while doing the reprogramming but also pat my head and rub my stomach.)
-
Status: I'm supposed to study for exam tomorrow but instead I'm reading transcripts from parliamentary session.
-
@Gąska said in The Official Status Thread:
I'm supposed to study for exam tomorrow but instead I'm reading transcripts from parliamentary session.
Procrastinate Now!
-
@dkf said in The Official Status Thread:
@Gąska said in The Official Status Thread:
I'm supposed to study for exam tomorrow but instead I'm reading transcripts from parliamentary session.
Procrastinate Now!
Procrastinate 2020
-
@Applied-Mediocrity said in The Official Status Thread:
@dkf said in The Official Status Thread:
@Gąska said in The Official Status Thread:
I'm supposed to study for exam tomorrow but instead I'm reading transcripts from parliamentary session.
Procrastinate Now!
Procrastinate 2020
Procrastinate until 2020!
-
@Gąska said in The Official Status Thread:
Procrastinate until 2020!
Okay, 2020 x 2019 x ... x 1 might be overdoing it a bit.
-
@Applied-Mediocrity said in The Official Status Thread:
@Gąska said in The Official Status Thread:
Procrastinate until 2020!
Okay, 2020 x 2019 x ... x 1 might be overdoing it a bit.
-
@Applied-Mediocrity said in The Official Status Thread:
@Gąska said in The Official Status Thread:
Procrastinate until 2020!
Okay, 2020 x 2019 x ... x 1 might be overdoing it a bit.
The best estimate I can find for the heat death of the universe is around 1e+106 years. 2020! is approximately 3.861e+5801. So yes, overdoing it just a bit, perhaps. Although Carroll and Chen hypothesize that a new universe may form through random quantum fluctuations or quantum tunneling in roughly 1e+5600 years, so, um, maybe you can procrastinate until the heat death of the next universe.
-
Status: On laptop because house sitting.
Being back on 1080p is.... Hmm, well, I miss my 4k already...
Edit: OwO But the OwO extension is making my spirits enliven!
-
Status: Sleep-deprived. Watching my son sleep his midday nap. Little bugger just learned how to climb over his baby-bed's wall last night, and started crying when he was unceremoniously heaved right back in. 21months, 12-something kg, and way too tall for his own safety.
Now I need to buy a new toddler-bed. Either one with higher walls, or one lower to the floor. Can't decide. He tosses and turns so much in his sleep that I'm afraid he'd fall down from anything without walls all around.
-
Status: It's good to log in and find we can by and large rely on @HardwareGeek for some informative -ry.
-
Status: Registered at Kaggle (under a different username than here), and look what my default avatar is:
-
Status: Cleaning up hard drive. Found a strange hidden folder in C: root named "GrandeDevice". After a short investigation, I've concluded it must've been created by a driver installer for my big fat Xerox photocopier. Whoever made the driver, chose a very fitting name.
-
@Gąska said in The Official Status Thread:
Status: Registered at Kaggle (under a different username than here), and look what my default avatar is:
Now waiting for @Polygeekery to challenge you to a fist fight...
-
@Tsaukpaetra said in The Official Status Thread:
Edit: OwO But the OwO extension is making my spirits enliven!
Surely you mean “But the OwO extension is making my spiwits enwiven!”?
-
@dkf that, or his vodka is haunted.
-
Status: fuck, Skylake is not 8th gen. Dammit, should have checked before redoing the whole fucking computer (motherboard swap).
-
Status: Git 1, my foot 0.
-
Status: literally thousands of stray vertices. Oops.
But all cleared up now. Weirdly, the UV unwrap is a lot cleaner.Also: pressure problems in my ear again – I hope it's not already blocked with wax…
-
Status: house-sitting for mother. A bitch and three pussies. One pussy seems enamored with my underwear. What the fuck.
-
status:
https://i.imgur.com/0cA5L0u.png
What a day.
-
@Tsaukpaetra said in The Official Status Thread:
pussies
Status: Perhaps it was a bad idea to test my cat allergy matrix for a whole week...
-
Status: Epic.
Not bad for a day's work, considering this was my first time for almost everything.
-
-
Status: Sewer outflow pipe is (at least partially) blocked AGAIN. Call in to the answering service of the drain people, waiting on a callback. At least it hasn't hit "overflowing out of the busted basement toilet onto the floor" level yet.
-
@Tsaukpaetra said in The Official Status Thread:
A bitch and three pussies.
I'll take care of the pussies.
Wait, what?!
-
@topspin said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
A bitch and three pussies.
I'll take care of the pussies.
Wait, what?!
Good, I might need some reprieve, my throat is sore....
-
Status: Microwaved a pizza that says not to microwave. I don't see the problem...
-
-
Status: Back at work (Fitbit says I slept 5 hours tonight, feels even less ), going through half a million emails. There's 3 projects that want a status report or phone conference now.
I don't know WTF the status is, I was on holiday. Get off my back!
Also, reading WTDWTF on Mondays causes Firefox to crash. I assume it's IT updating stuff again which gets triggered.
-
@Tsaukpaetra said in The Official Status Thread:
@Tsaukpaetra said in The Official Status Thread:
pussies
Status: Perhaps it was a bad idea to test my cat allergy matrix for a whole week...
Status: well, whatever. Type 4 nasal irritation well underway, cascading into information, then into sinus blockage, then into pressure headache and fun throat issues as drainage occurs.
I might visit the pharmacy for antihistamines...
-
@Tsaukpaetra
It's clear you can't handle all that pussy
-
Status: Major breakage with github turned out to be due to my phone resting on the Enter key of the second keyboard.
-
@Luhmann said in The Official Status Thread:
@Tsaukpaetra
It's clear you can't handle all that pussyI am indeed over my head.
-
@topspin said in The Official Status Thread:
reading WTDWTF on Mondays causes Firefox to crash.
That sounds like a fun bug report.
-
Status: Trying to understand Microsoft's licencing for enterprise products, heck after talking directly with them I'm not convinced they understand them.
Also, they lied to me about a migration path, I don't need an AD trust to migrate NTFS shares.
-
@Cursorkeys said in The Official Status Thread:
after talking directly with them I'm not convinced they understand them.
You may very well be right.
-
@Cursorkeys said in The Official Status Thread:
Status: Trying to understand Microsoft's licencing for enterprise products, heck after talking directly with them I'm not convinced they understand them.
I attended an IT meeting last month (as a substitute, thankfully I don't have to do that normally) and IT explained to us that MS is changing all our volume licensing terms for Visual Studio. The details were complicated, but can be best summed up as "we really don't know yet", so I didn't even try to understand it.
-
@Cursorkeys said in The Official Status Thread:
Also, they lied to me about a migration path, I don't need an AD trust to migrate NTFS shares.
Though, if you want ADMT to perform an automatic translation of the permissions on the share, you do need a two-way trust.
-
@acrow said in The Official Status Thread:
I'm afraid he'd fall down from anything without walls all around.
Make him sleep on the floor
-
@izzion said in The Official Status Thread:
@Cursorkeys said in The Official Status Thread:
Also, they lied to me about a migration path, I don't need an AD trust to migrate NTFS shares.
Though, if you want ADMT to perform an automatic translation of the permissions on the share, you do need a two-way trust.
The problem I have is that I need to change the domain name from
.local
. Any method of doing that still seems to be a really bad idea. So, my current thinking is to re-create the users and security groups on the new domain and then useSubInACL
to do the needful.The advice on ditching
.local
seems to be mixed, but it's already caused problems with not being able to get a replacement SAN certificate since they changed the rules a few years back.
-
@Cursorkeys
ADMT is an ideal tool for that sort of change, and is the recommended practice. The short of it is basically:- Create a new domain with the target new name
- Set up a two-way trust between the
.local
domain and the target domain - ADMT does magic
- ???
- New domain!
Especially for a straight rename like this, ADMT will work very well with default settings, and will handle security translations for file servers, user workstations, etc. It doesn't really do security translations for SQL, but if you do the necessary setup so that it can import SID History, your SQL servers should be able to handle "translating" new user accounts to the equivalent old account's ACLs if you miss a translation (though I would still recommend updating the mappings in the SQL servers ASAP).
-
CurrentVPN status: It's doing this again