#off-topic-tech
1 messages ยท Page 57 of 1
Bc I found one I can get for reaaally cheap
(And it works with my business name)
how much is the renewal
40 a year
yes but the .com redirects to some IT/AV company I've never heard of
So likely no real conflict since I'd be on very small local scale and primarily doing repair
I need help with spawning an actor in Unreal if anyone has advanced experience in that area.
(I am specifying advanced because I've gone through about 20 people I know -including my instructor- and none of them can figure this out).
in 2 probably in like 3-4 ehh they will probably be trying to start pushing AM6 or something
wheres the a770?
I say we have 4 years at least am5
If AMD designed the socket well, there is no need for change before either DDR6 support or more RAM channels is needed.
With AM4 they would probably have wanted to raise the socket power limit in the last CPU generation.
well what i mean by that is they will probably start sniffing trails of AM6 in like 3-4 years not drop AM5
I'm so close to 16k
But I'm hitting thermal limitations of my cooler
The jet engine approaches @pure karma
hard to beat my 1660 super whit the shitiest bearing on earth that spins at 3600RPM
well its held together by hopes and dreams
Lol
so what a shocker
it also runs off of hopes and dreams so guess i see some sort of corelation
Aaaaaa.
Why am I so close
And yet
30 ish pts is all between me and 16k pts on cb
Shut Windows services etc. off and disconnect internet?
I did
I seem cb to realtime as well
No improvement
That is my highest score to date
Secondary and tertiary RAM timing tuning timE?
Doesn't effect cbr23 as much. But at a decent state. (63ns latency 47Gb/s)
Yes, not much, but might have been enough to get the final points.
I don't think the ram will take anything tighter
But that is pretty close already to what can be done.
The bandwidth might be little low.
Its already pretty tight
It is a massive improvement from my micron rev F kit
10gb/s faster or so
my new pc sounds like a boring wind tunnel and that's really about it
Mine sounds like a jet plane
What settings effect bandwith the most
As the latency is decent on these hynix dimms
Depends, for this I would probably first try to tune the tRFC to edge of stability.
That is usually included in XMP at very loose setting.
And as on AM4 you cannot tune tREFI, that matters much.
So that 312 as timing is below that.
Anything else I may wanna tune
For the full refresh, so 1.
Then the others are half refresh and quarter refresh, but at least AM5 doesn't use them.
Wait like I said, those values were in nanoseconds, and the timing is in clock cycles.
This is AM4 DDR4 CJR8gbit. Or c die
To convert to ns: 2000 * timing / ddr_freq.
For example, tRFC 250 at DDR4-3200 is 2000 * 250 / 3200 = 156.25 ns.To convert from ns (this is what you would type in your UEFI): ns * ddr_freq / 2000.
For example, 180 ns at DDR4-3600 is 180 * 3600 / 2000 = 324, so you would type 324 in your UEFI.
So is 136ns
Trfc 260 at 3800mhz is 136
So probably not much to tune on that.
Some RAM manufacturers put really loose tRFC into their XMP.
Yes, as that chart is baseline.
At 3800
Not the absolute max some specific chip units can do at low temperature.
Just what all of the chips should be able to do at sane temperatures.
Idk if its applying
And tRFC and tREFI are significantly temperature dependant above 60C or so.
I hit f10
My mobo has no diagnostic lights
But its just sitting there
I may be unable to adjust timings
But that URL has various suggestions for secondary and tertiary timings.
Might be that it stops applying some XMP caused auto setting properly when you do the manual tune to any secondary or tertiary.
As VSOC, VDDP etc. aren't directly set by the XMP, but as auto reaction to it.
Which is why the guide tells to use loose primaries to set secondaries, and only after that tune the primaries.
The XMP setting on this kit is pretty tight. By all standards.
As several of them have interactions with the secondaries and the secondaries are more important than the primaries.
i want to rant about how these lazy asses at JetBrains introduced this stupid frustrating bug into Jupyter notebooks view that makes it autoscroll to the same cell again and again despite all your effort to move away from it to do actual work. it's like if you where reading a book but it kept turning pages back. or writing in notebook that kept sticking your pen to another page.
the worst bug i have ever had
Hehe I cast starlink. ๐
I wasn't sure if that was my Aussie internets or not.
i remember the days that was the norm
XD my internet slow rolled it to me
Did I read it correctly, someone made a car that forces drivers to use touchscreen for blinkers? ๐
da fuq
i'd barely even use it to turn on my lights, or other things too ๐
but yeah, actual hot-seat operations ... i would never want that as a touch screen option
My car lights are set to automatic so I don't have to worry about that 
"lets put brakes on touchscreen!"
my wipers are automatic with when the window gets wet ๐
conversion kit... my car is a '84 ... wasnt standard lol
I think my parent's wipers are semi-automatic. You still need to turn them on but they will stop if there is no more rain
I support using touchscreen for multimedia but not for things like blinkers, wipers, ac controls, etc...
my GPS is touchscreen ... thats it ๐
My mother is furious and she bets it's the french who originally came up with the idea lol
LOL!
She never really liked french cars 
blame 'werid shit in cars' on the French
Ive never seen automatic wipers before
Nah, mr musky
if someone tried making autopilot probably whithin 3 years
Apparently not in the states
Wish I lived in the same city where I work. I would be glad to use a bike if that was the case
Set the speed settings manually here
I think some wipers are automatic in a way you don't actually need to turn them on. They will activate automatically as soon as they detect rain and stop when there is no rain
Others are like semi-automatic. You still need to turn them on
newer/higher end trim cars have auto wipers
i do yea
i hate touch screen stuff in cars
nah
i hate touch screen keyboards ๐
Hope this won't happen anytime soon lol
Imagine people standing in front of hospitals to get a new firmware
No money? You die!
imagine firmware failures on bioaugs ...
"i just bricked my brain, brb"
"Have you tried turning it off and on again?"
how does one factory reset a failed firmware update /??
'my clone will save me'
i was reading a book where people had a clone in a suspension chamber all the time, with a real-time sync link going, and if you ever died, the clone would wake up, and a new clone get assigned for the sync
immortaility, without being immortal
some things went wrong, and some people would wake up thinking they died ... but the other had not .... resulting in multiples of the same person running around
Isn't that similar to SOMA?
Yeah, it has a similar story
ah, ive noticed a lot of sci-fi stuff being very similar
EVERSPACE was kinda like it as well
wow, the somagame website is a broken pile of dren ๐
Wth happened to the website lol. I don't think it even looked like this before....
Best of both worlds with very minimal drawbacks
Must have mistaken it for some other game. Thought it looked much better than it is atm
Weird
i tried in a profile with and without adblocks... so thats weird
I'm running all ublock filters
Safari looks correct, Chrome not
I'm using firefox
looks fine on my firefox with ublock
which is hilarious, given Safari
guessing they doing something that Chrome is breaking on ..........
Chrome doesnt like the SSL
strangely enough, the resources come from different locations (not just a single CDN) ....
yeah, its the CDN they are using:
the cert is invalid-ish enough for chrome to balk
but Firefox and Safari are like "it looks fine, this is fine"
looks good enough, must be fine
600k filters 
WHo bought that $360 4060 Ti?
it was not me
the cert they are using is issued for "*.cloudfront.net" but the domain in the url they are using "cdn.thedarkswarm.com" which ... doesnt match weird ... safari usually is the one to bitch about that more
yup, there it pops ... now safari is breaking ...
FUN
so what happens in safari is ... if those are linked as resource, safari doesnt care that the name doesnt match in the certificate .... but if you go to the resources directly, it complains
My browser still likes it
yet in chrome, it just complains anyways
I was able to load it in chrome & firefox
heh!
yeah, my chrome isn't having any of it ... loaded from the website, or directly ....
my safari is half n half, it accepts it if its a resource link, but wont if its a direct
Oh wait
That's because chrome is defaulting to http on my end
If I force https it breaks
heh! yeah .... sometimes i have a hard time getting chrome to try the http version lol
HAX lol
Safari on mac is worse... it like ... totally ignores the existence of http
apple .... "thou shalt not go to http"
Http is still fine for stuff like that where it's more or less just a landing page with some links
exactly ......... !!!!
info pages, no point being all secured up the wazoo ... theres no logins, no forums, no chats, or comments
Had to restore router back to stock firmware from OpenWRT ๐ข
Everything on LAN was working perfectly fine on OpenWRT but there are a lot of issues when it comes to Wi-Fi
Guess OpenWRT doesn't work perfectly on all supported routers
Taking the voltage offset off helped
It's not that bad when it comes to Asus but I still wanted to test OpenWRT
Guess it's best to use OpenWRT for recommeded routers and not just all supported routers
any page script or data on these page can still be manipulated
it's not only a matter of privacy but authenticity
it was last openwrt update ?
because there a often Linux kernel wifi ap driver patches
Yeah, still had issues
I am going to get a new router and AP soon anyways
Edgerouter X SFP and Unify U6+
I find it hard to get really wifi ap
always are some kind of (stupid) all in one solution
doing everything bad rather than doing lan <> wifi good
Yeah, it's obviously the preferred choice
But not loading it at all seems like a poor fallback to just loading the page normally and plopping the little insecure icon in the top left
Although there is an argument to be made that you can't expect the consumer to know what that means
its allways problem after problem whit computers i swear....
major stability issues again it seems
well atleast its a fairly consitently recuring problem
the downside is every single time it happens theres just no signs of issues other than just everything crashing all at once
this video might help idk
its prettymuch just throwing a bunch of unspecified hang errors
It looks like first gen GDDR7 chips are slated to have 3GB modules, if roadmaps are to be believed. So Nvidia can potentially increase VRAM by 50% across the board. If they want.
All the same sort of memory configurations are possible (12 or 24 chips on a 384 bit bus, 4 or 8 on 128, etc.)
if they want
they likely wont though
5090 probably will. Nvidia always uses max available chips on flagship, if history is anything to go by.
Anything below flagship is questionable.
xx60 class cards probably not
I can see a 128 bit bus and 12GB for the 5060.
i think i got my answer and its not the answer i was looking for
I don't guarantee it of course.
PSU.....
What led you to that conclusion?
Plot twist, Nvidia makes a 256 bit bus 20GB 5080 with a mix of 3GB and 2GB chips. Thus bringing back concerns from the GTX 970 days, where the memory bus doesn't evenly address the memory.
I would hope nvidia learned from their mistakes
the fact that i hit 830W power consumption only counting CPU+GPU
What's your psu rated for?
well it says 850W so im gonna go whit 850W
Well yeah, that's definitely not helping then
850w Peak or sustained. is the diffrence
Well not really
Anything measured in software is going to have a long enough duration to more or less count as sustained for power delivery purposes
yeah
but ive seen some psus only say peak
kinda surprise the system dident shut down and recovered on its own
And you must also account for transient spikes, as older psus were not designed or required to handle them like the most modern ones are
The system not shutting down probably indicates poor psu design
Or it means that it isn't a psu problem at all, and is an issue with the wiring
wel is the PSU supost to imidiately trip at its rated wattage or crawl the system to a hault like in the video i sent
its supposed to trip at a few% over rated wattage
2-5%
The psu should immediately trip
If what you sent was caused by power delivery, it was due to a brownout
Even more usually
Again, not necessarily
interestignly tho mathematically CPU-GPU shouldent draw over 750W ever whit a like 100W window
Entirely depends on the unit
ahh
Most psus will trip at a hard current limit that's set somewhat beyond their rating, but they have other protection measures that will often kick in before the current limit
so either gpu was 200W over rated spec or 13700K was
for my unit specifically is 2-5%
What's your unit?
A PSU won't necessarily trip if a fraction of a second power spike goes above its sticker wattage.
something called transient spikes
THey're built for that.
EVGA GA 650
Or you had a software measurement error, which is extrememly common
it wasent a spike it hung for like 20 seconds before crashing
Depends on the design
Atx 3.0 psus much more so than older models
I recall seeing some PSUs call out that they can handle some higher amount of power during a tenth of a second or something like that.
GA 750 review lmao
Yeah, that's again very model dependant
Hence reviews
evga builds very good psus
Basically if it doesn't straight up short, it won't trigger
My 1kw unit from 2012 is still kicking and only just recently left warranty
so my 650w should be fine handling a 6900XT
here are the software readings under just a normal furmark stress test......
basically rebadged 1000w units WTF
Just because it doesn't trip until then doesn't mean that it'll work properly
Actually, it most likely won't
Those are numbers are still too high
hwinfo or hwm
Yup
Especially since you've overloaded the wiring at that point
HWM
use hwinfo
You'll typically see peak numbers around 1200W because that's where you'll start tripping breakers in the USA anyway.
HWM is vastlty innacurate
Should be more like 1500w
Those are numbers when protections kick in, not what it can handle
eh
1200W if your wiring isn't done great
ill be at 93% load
Which is an unfortunate reality that's out there.
with worst case spikes
was the 6900XT one of those GPUs that was prone to random millisecond power spikes?
thought i remembered that
Sometimes I regret not making my own music playlists
I end up listening to the weirdest shit ever
Whatever spotify feeds me ig
when my parents built their house they put 14 gauge wiring everywhere just to make sure
Fun fact, my Bolt by default charges at 8amps, and you need to specifically save a profile to get it to charge at 12 amps, because Chevy is so afraid of shoddy home wiring.
where do i find power in hwinfo
sensors
theres like 400
there arent any only ram
That's really interesting
Even a 16awg circuit with poor connections can handle 10a no problem
Heck, you could charge with a crappy extension cord at that level of current
Yeah, they're probably afraid of people using flimsy extension cords too
One of the issues with P750GM was that the protections were set way too high so instead of the protection kicking in, it just started burning
i managed to somehow not fry my pc by using a monitor's ribbon power cable at first (wasn't paying attention) which almost instantly OCP'd
for gpu
why cant you just filter by power....
no idea
never needed to
ight there
allright there
This is going to end poorly
its already poor in a way
I want snarky commentary from ChatGPT when it's grading a paper it wrote in the first place.
this is an amazing song
Diggy Diggy Hole lmao
I don't know what spotify is feeding me, but I've heard quite a few times
diggy diggy hole
digging a hole
Try the AsusWRT Merlin?
RT-AC66U B1 uses the RT-AC68U firmware version.
Unfortunately, my model is not supported (RT-AC85P) ๐ฅ
I am planning on buying Edgerouter X SFP and Unify U6+ anyways
Hope AP will have enough coverage through concrete walls
I could buy a better model with better coverage but I am not sure if U6+ is not enough or perhaps more than enough
Anyways, I'm off to work
BG3 is still amazing
doing my 2nd "full" runthrough & i missed so much the first time
Thanks for reminds that I need to finish my honour mode run ๐
near as i can tell they did flight just about perfect
so many games get flying wrong, or do the "we can't do it because of X"
MZ4250 patreon
Paid 5โฌ for access to the google drive
Only need it for a month but I'll prolly just let it run for another 2-3 months just as a thanks
He got blender + stl files
He doesnโt have alternate ways to give money? So patreon wouldโt take their cut, though ig it isnโt that massive.
Maybe they'll get less overpriced now
Nikon is not really known for low pricing
Lower than fucking red
They do this thing where they price their low end stuff really really cheap but then the price scales up way faster than seems reasonable as you move up the product stack
That is true
I wonder if the red brand will stay ornamental Nikon will just phase them out and use their stuff in Nikon cameras
model, not brand ... ๐
Yeah, I'm aware
And nikon has made some bad stuff
But as a general rule, if you're willing to pay the price, I've never had an experience that did anything but blow past my expectations
My older body is at well over double its rated shutter life and still going strong
i know... Nikon (and many of the old OLD brands) are pretty decent
i was just being silly ๐ cause im always being told model not brand lol
ive got a Nikon camera still
Yeah, they've always been solid
I do find it a little funny that they've diversified so much
They have a division that does lithography machine mantineance
thats odd
It's because you use lithography machines to make imaging sensors, dot matrix meters, and phase detect af sensors
So even though they outsouced their imaging sensors as of ~2010, they still manufacture af sensors, matrix meters, and the processors for the cameras themselves
Sorry, you need to type louder
I can't hear you now
Perhaps too quiet?
No satisfaction of a mech keyboard now
Well, quieter than membrane lol
A lot quieter
af sensors? ๐
Yes?
"af" lol you guys introduced me to "af" ... and now thats all i see
๐
They're good af
AH! ... there we go
Although tbh I'm not entirely sure if they still do their own phase detect sensors for the z series bodies, since sony makes the imaging sensor and the contrast detect hardware, which have the phase detect system integrated now(if present in that model)
my Nikon was before all that... it takes film (which is hard to find)
Walmart sells it
I love fr
They'll develop it and run some prints for under $10 a roll
Just add frfr at the end of your sentence
hmm ... brave walmart ... hmmm
walmart around here is like pure ghetto dangerzone ๐
Film was the first thing that brought me to walmart last year, and they were out of stock both times I checked ๐ญ
if you dont get run into in the parking lot, you may get mugged at the door, or held up inside
That becomes annoying af quickly frfr
Nah it's great frfr
fraf ?
Also usable with the remove command:
rm -fr folder
sudo -af
The scariest thing you'll find at ours is a karen
Might be worse than a mugging
walmart is risking life and limb.... target is risking sanity and just losing it and ending up going to jail for your actions
sometimes i find its best to go to brick n mortor stores with big headphones on and the music turned up
rm folder frfr
folder gets removed frfr
Reminds me of a programming Syntax we had to make as an assignment in Uni.
We called it "Sprache HuSo" and it was quite easy to read if you just pretended you were talking to a Hauptschรผler
Basically a simpler form of a programming language
Pseudocode?
But no actual functionality behind it, just the definition of how it's supposed to be written
No, it was defined like a proper programming language.
It just wasn't implemented or anything, just the language definition
Interesting
The topic of the course was compiler building, so it was necessary for us to learn how a language is defined
The languages we actually implemented were much much simpler
anybody a idea to fix this shit? game is new installed
Wrong channel, but you have old mods installed.
They don't get removed when you uninstall the game or the mod manager.
my disapointment in discord rn is real
Imagine getting 1gbps down... going to host stuff and stream and it's only 30mbps up
asymmetrical bandwidth is the bane of my existence
my current internet is like that. it's amazing.
Microwave dishes are not very expensive ๐
we looked into laying fiber but it would have cost like half a million dollars
a $50 airMAX dish is much more reasonable
Longmont, Colorado has a city-owned fiber utility, which is unheard of in the US
Oh, sorry checkit
NextLight is
Symmetrical 1000 Mbps upload and download speeds
and it's under $70/mo
insanity compared to CenturyLink et al.
and according to someone i know who lives there, they aren't lying
Even worse... people buy that shit.
isnt the reason asymmetrical bandwidth exists bc of physical limitations
like the system was originally designed for cableboxes which are not going to be uploading much data at all
Asymmetrical bandwidth is a holdover from DSL days
It also comes into play with wireless systems, weather they're short range or satellite
cable sux ๐
they lopside it cause it cant handle much... so they try to put it most on download
they are super fast when compared to dialup
it was all the marketting trend here "broadband" vs "dialup" ... but dialup is dead, but they still seem to think they are the shit
now its moving more to 'fiber' vs 'cable'
๐คฆโโ๏ธ
I am getting to old for this.
then it will be 'subspace' vs 'fiber'
but I hope everyone who buys it has fun with it.
quantum vs digital ๐
It's getting there, close to half iirc
fiber took my city by storm and the cable guys are like 'oh shit'
more like 150 for gigabit lol
Nah, I pay 85 for symmetrical gig
But I do actually feel like at least for now, that cost is justified
im paying $60 for 600/600 ... sub gig speeds, but i dont wanna pay 80+
It'd better not do anything but go down in the future though
Yeah, high income but also high col
they constantly raising rates ... its like they NEVER go down
they wont go down until people stop paying it
Hdds went down ๐
That bad there too? Same is happening here.
every month they try to justify some new fee or price increase lol
Oh monthly even? We only have it yearly ๐
Sometimes
But to be fair, this is also the same price we were paying for 28/5 just a few years ago
My ISP is also very small, which helps
im locked in till feb 2025 .... but it will go up then, and i have to call in and negotiate a new 'deal' to lock in
right,i was paying $60/mo for my old 30/3 DSL lol .... so im not really upset
and my isp has twice now increased my speeds 'at no additional cost' ... last bump was for my upload to match my download
i have better internet than one of our work servers which sits directly on the net
theoratically: 84.9Mb/s down and 27.6Mb/s up. I pay around 85.68 euro's for it.
๐ฆ
ooof
The upload is even better than my ISP tells me it is.
but the download is way worse.
I keep checking my ISP page of announced fiber projects. But so far, my street isn't planned
such a tiny country, you would think it's easy.
Probably the density of buildings.
You would think that would make it easier?
Here, they've done overhead fiber in places where they can't trench it
I am just trying to think why it takes so long. I just told you how much I pay for internet. It's a lot. So we know they got money.
What else could make it delay. Maybe not a priority. Or mismanagement.
They aim to have fiber for half our population by 2025.
in 2022 they provided fiber for 600.000 families.
We got 5 100 000 families in Belgium (2023), if they do 600 000 families a year, it will take 8.5 years.
probably ๐
maybe they upped the price of other packages?
or have other sources of income to cover it.
new concept to me
So how do they get bandwith? From corporations?
I am off to bed ๐
bonk
Here's one where my dl wasn't so slow
Symmetrical gig baby
I might have to check that out...
i have like 100 other people using the network rn too
bruh
where yall live to have this good internet
Id sacrifice 300 download for that kind of upload and ping
fiber areas
also @twin dew is it illegal for my isp to say price for life $xx but then theres a cost recovery fee ontop of that (its 2.99 usd/mo)
It is legal in my state but century link started using it but got sued for it
Oh, they do, but the requests get lost in transit due to packetloss
The various "administrative" charges etc. are basically that.
Extra fees to cover the basic taxes and costs the ISP has, over the official price.
except it's part of the service
so it's scam
you can say whatever
it's like selling "free" goods at $100 shipping
Yup, just very common at US.
IIRC airtravel tickets got special legistlation already that they must include all "mandatory extras" in the advertised price.
likewise floor shipping price must be 0
But similar regulation for internet providers and cable-TV is stuck in FCC.
one common way of achieving it is "free shipping for over $x order "
EU ruled that IAB is not respecting GDPR. 80% of european cookie pop-ups is managed by IAB.
Weird thing is that IAB has an alternative system, but they wanted to wait for the court ruling.
Because their current one is made to be as hard to use as possible, so that people would just click "allow"
This is the appeal btw. Now a Belgian judge has to make a descisions and I hope this gives a blow to these stupid cookies.
I've been hating this for ages. And everyone who knows me, knows this.
Some websites do it perfect though.
the problem was a TC-string, coupled with a cookie and IP, advertisers could still track you.
Now EU made the same ruling about the TC-string.
The ruling of EU seems to be that a TC-string is also considered personal data.
I mean not familiar with the concept of a TC-string. It looks something like this: "BOSSotLOSSotLAPABAENBc-AAAAgR7" and is the data of your consent. What you did consent to and what you didn't. Not sure if this is really personal data. But apparantly it is. Especially when they are coupling it with cookies and IP tracking I guess.
Personal data is information that relates to an identified or identifiable individual. If you cannot directly identify an individual from that information, then you need to consider whether the individual is still identifiable.
And seems that IAB has been determined to have used that their TC String concept to identify users, so it counts.
ofc it counts, why else would so many courts rule like that.
i'm really waiting for legislation that's a "either don't gather or use that info, or when you do, send that person a check" & see how long that industry sticks around
My internet speed, hello from France...
or just bypass that stuff entirely. make companies report the collection & sale on their taxes, & then give people a deduction on tax day ๐คฃ
U mean something like a universal basic income
more of a "knock it off with the useless lawsuits about stolen data settlements & security breaches that sit in court & appeals processes for years until the lawyers siphon off all the money" avoidance
it's a lot
It's for my comfort, I've only had optical fiber for a month and a half in my cave!
okay
Anyone else have these in their event viewer? critical -> "System maintenance detected issues requiring your attention. A notification was sent to Security and Maintenance." (event ID: 100)
Deadlines and notification Critical maintenance activity must execute within a pre-defined time window. If critical tasks have not been able to run within their designated time, Automatic Maintenance will automatically start executing at the next available system idle opportunity. However, if the task state remains behind deadline, Automatic Maintenance will notify the user about the activity and provide an option for a manual run of Automatic Maintenance. All tasks scheduled for maintenance will run, although the most starved tasks take precedence. This activity may impact system responsiveness and performance; therefore, Automatic Maintenance will notify the user that critical maintenance activity is executing.
Due to the nature of scheduling, some scheduled tasks may not finish: perhaps there are too many scheduled events to fit into the 1-hour maintenance window, or maybe the computer was simply not turned on. In such cases, you can define a task with a deadline. A deadline is defined as a recurring timeframe in which the system must successfully perform the task at least once.
Interesting, so if it can't finish the tasks in 1-hour, it stops and schedules the tasks for the next day. Maybe that's what was in my event-logs mhm.
If a task misses a deadline, then the maintenance scheduler will continue to attempt to execute the task during the maintenance window. Further, the scheduler will not limit itself to the regular 1-hour time limit. Instead, the scheduler extends the duration of the maintenance window in order to complete the delayed task.
Microsoft has been at least since Windows 10 expected everyone to keep their computers max in sleep mode during the night, with wake timers activated...
So that the computers would turn back on during the night and do all those things, and then go back to sleep.
"all those things" all kind of things
coming from Microsoft
"Maintenance" is :
- removing Feature
- increase telemetry
- add pop-ups to connect to Microsoft accounts
- reinstall one-drive
- install "pre-installed" adds for office and Xbox pass
- update call home strategy
- uploads the "security reports"
- ..
I don't understand why this is critical event. PC shutting down unexpected, yes, that's critical. (kernel event 41)
i had a chair once with arms like that... and yes, they annoyed the crap out of me because i couldnt lower my chair enough to get under the desk
i ended up buying a whole different chair =p
that too
when arms are not adjustable themselves... it leaves you sitting on the floor ๐
ive got a Komene chair right now, adjustable arms
hell i just want to remove them
but as u can see in the image they are part of the body of the chair
good grief
yeah, my old crappy chair is that way .... the back is held on by the arms, which are mounted to the seat
then the 'rubber' parts on the arms broke ... so i was left with just metal rods to put my arms on ...... so i tried wrapping towels on them lol
weird. Crealities slicer produces shit results on windows and macOS, but the linux version just gave me the best print ever.
linux ftw
apparently the setting are way more granular and the preset "hidden" settings are set up perfectly
(it also looks way better)
Im getting a lot of coil whine from my gpu while running local llms
doesnt happen during games
Because llms have a much different power profile
Interesting
I should try it, because its more or less useless onwindows
it's basically cura but specced for creality printers
i know that but its kinda annoying that this task causes it. and nothing else I do
my CPU whines pretty hard on the start menu of DRG.
your CPUs whine?
lmao
cuz u know
if I wanted to do simulations of aerodynamics of 3d objects how would I go about doing that
THATS THE THIRD ONE LIKE THIS TODAY
new pioneer or what?
nsfw discord link
can discord make a rule that stops people from posting invite links
like a permission on a role, and if you dont have that perm no invite link for you
That's a thing already
And that person got automodded
For some reason it was just slow
Eh it does it after the post was sent
In talking not even letting them post if it detects a link or part of an invite link in the message
Yeah... They can't do the detection client side, and on the API side it's apparently way more efficient to implement it this way
Someone need to train an Ai model solely for regex
I spent quite a while learning regex about 6 years ago and now I really don't remember any of it
Regex fog
even shows you how long what needs in the time estimate
Skin ๐ณ
Oh that's pretty cool actually
Mightโve been posted but this is interesting
really useful to reduce print time if you wanna do something quicker
I didn't say it did that in DRG.
I said it did it on the start menu of DRG, once I click a button to get out of the menu the coil whine dissapears. It's consistent.
and it doesn't go away either, tested it. If I just leave my PC running with DRG start menu, it keeps on making a terrible noise.
You got any light leaks?
You see at the end of the cave the light area?
That's a glitch because of the shaders I use. Once you go further down it's pitch black.
I use optifine.
go fabric and use iris
benefits? or your opinion why you switched.
iris just does shaders, and requires sodium, which does the performance part
and sodium is just straight up better than optifine in everything
and optifine still advertises that MCPatcher is not needed
aight, I will look into it, thanks ๐
Can't wait to see how this horror game will turn out: https://www.youtube.com/watch?v=R2uW-9BKzuU
Get a tease of Hideo Kojima's newest game OD in partnership with Xbox Games Studio.
#tga2023 #hideokojima #od
Apparently, it's going to be very scary
Atleast they are "transparent" about it.
Guess we'll see soon. Kojima hasn't dissapointed me so far
You liked walking simulator?
Yes 
re: sodium
if you have a recent Nvidia GPU you can also use nvidium which is an absolute game changer
doesn't work with shaders though
AMD here @rotund hamlet
aw foo
Thanks for the tip anyway ^^
tldr nvidium uses mesh shaders to blast fps through the roof
300 fps at 32 chunk render distance on a mobile 3050 my beloved
What's the difference between Death Stranding and FedEx? In Death Stranding, Sam actually delivers the package
why is it nvidia only?
I would have kinda expected this been done in minecraft clien itself. That Microsoft would have enhanced it graphics somehow. But I guess they don't feel the need to do so since other mods do it for them.
im guessing they mean aio pump whine myn does it too but only on startup
Outdated tech.
Fine if thatโs your only mod, but sodium is preferred if you want more, like any shaders.
they invented mesh shaders and AMD hasn't caught up
they have been making improvements to lighting and loading, which made a couple fabric performance mods obsolete in 1.20
there's also a mod that makes it use Vulkan but I can't get it to work
Mozilla/5.0 (X11; GNU/Windows 10.0) AppleWebKit/537.36 (KHTML, like Gecko) Chromium/79.0.3945.130 Chrome/79.0.3945.130 Safari/537.36 Tesla/2020.16.2.1-e99c70fff409 cursed real useragent
Man I really wish they could figure out how to do parallelized slicing operations for 3d printing
because who would want to spent 1000$ a month just to have a working 3d printer
3d printing hasn't made a ton of advanced in capability, but in terms of consistency, reliability, and usability for the price they've improved massively
Well, AMD is missing Mesh Shader support in the OpenGL driver.
Not in totality.
got a dumb question since i rarely mess with stock
when you undervolt a cpu, does it crank up the amps in order to push the frequency higher?
No, amps are "byproduct" in a way in this case.
That is the point of the undervolt, to get the power lower.
And Power = Voltage * Amperage.
And as the resistance is almost static, the amperage goes down as voltage goes down.
right, i didn't know if it was a "lower the voltage to prevent burnout, but the wattage will stay the same (which would mean upping the amps)" kind of thing
i assume there's not a "let's fuck with the amps" setting for cpus
Only when power limited in software.
Lowering the voltage for specific frequency bin allows it to boost to higher bins.
kk
Which raise the voltage.
Just couldn't hit those higher bins before inside the temperature and power limits.
been fiddling with the ryzen master curve optimizer software & was curious
plus i saw it go to like 1.35v when it was screwing around doing its thing
kinda tempted to cap it at like 1.1v to see how it'd do
5600X btw
Point is to try to hit the highest frequency bin the CPU allows for.
Which can be done by combination of raising power limit, improving cooling and lowering the voltage each bin uses (negative curve optimizer).
i'm also trying to keep it from doing those weird cooling ramp-ups whenever some background process kicks in
do newer intel chips with e-cores not do that kind of thing?
Cooling ramp-ups?
That is problem with your fan curve.
yea true
might crank up the case fans a bit more, & then keep the cpu fans down until it gets really toasty
brb bios fuckery time
ok that's a lot better i hope
& i know, it's not a top end cpu, so it doesn't really matter all that much to performance, just nice to be able to keep it calm when there's nothing to freak out about
I want to host a game with 1-3 friends I have a GPU of 16GB and a Radeon RX 5500 XT. Will my PC be good enough?
you never learn by keep using icue after all these years
Where did you get 16GB RX5500 XT?
Normally that GPU comes with just 4GB of VRAM?
I was talking about two different parts of my hardware. The RX 5500 XT is my graphics card
But it could be that I am switching some things around in my head
And you were writing shit, on wrong channel.
So you have some random CPU you didn't specify, and 16GB of RAM, but you stated you had GPU with 16GB, and then the actual GPU separately.
Let me look it up again
This is one of the off-topic channels, and I'm if you aren't asking about Satisfactory, then you would have needed to specify the game too.
I am talking about Satisfactory.
I have a Radeo RX 5500 XT GPU and a a RAM of 16GB
Not related to player run hosting.
But RAM limits mostly the save size, CPU would determine how fast your own FPS would be.
As each player adds about 50-100% more CPU load over just single player.
You got better specs for play run hosting? Can't find it atm.
There aren't.
For the clients it is mostly the same as with dedi, but the hoster with bad CPU gets crap FPS from that extra CPU load.
If it's not on the wiki, maybe it should be. This questions get's asked every x weeks.
If CPU limited, FPS will drop that 33-50% per client.
If not CPU limited, depends if it causes the game turn CPU limited.
What's the difference even between hosting a session and running dedicated server? Doesn't it just spin up a (hidden) "dedicated" server (so same technology more or less) when you create a session, and automatically stops it too when the player quiets. A.k.a. a sessions is a opinionated dedicated server with complexity hidden for player?
No.
The game is always running host-client even in SP.
But the code is different from the dedi to some extent.
Right ...
And seems to be much more CPU limited when actually hosting multiple players.
For the hoster.
lol, reddit disagrees like always ๐
And I can be wrong.
Just lot of people who have been hosting with really bad FPS after first client joins.
But the hosters GPU seems to run at lock-step with the processing still even when hosting.
And when hosting, the CPU time goes way up.
oh wow, finally found a good reddit post.
yeah, I google a bit later. Want to know where they differ technical.
I have to work now ๐
Well, with dedi the rise in the "game" time doesn't really affect anyone so directly as when hosting as player.
@twin dew Connected the robot motors with pi and ps controller, it worked until L298N burned down. How could I make a H-bridge out of these? Or would you recommend using smth else? www.amazon.de/-/en/gp/product/B0CGVB651Z
No idea right now.
And why not just mod the 555?
The 555 was for the pump, and there's noone with soldering skill here
The 555 wouldn't solve the reverse speed problem
I can use those mosfets for one direction without making a H-bridge
We thought the L298N could handle the motors, but it clearly couldn't as the nice orange flame showed.
Those are just low-side units.
You would need separate type for the high-side.
And then two of each set in H-bridge.
As the VIN+ is directly connected to the OUT+.
And the controls is on the - side.
So would need separate type that has the VIN- and OUT- direct and control on the + side.
And then not use the direct connections at all.
And for H-bridge, you usually want that the low- and high-side drivers have same response time when switching.
To not get momentary shorts on one direction of change.
Of course can be fixed on SW side with suitable delays.
But as low-side MOSFETs are cheaper and faster for same power, you might not find high-side controlled module like that.
Cheaper for same power, faster for same price.
Wdym low-side?
As N-type semiconductors are just better than P-type.
For H-bridge you have two low-sides and two high-sides, controlled as two pairs.
Okay...
Haven't studied electronic, seems complicated
Maybe I just make the motors one-direction
Though not ideal
Not that complicated, just needs right parts.
Do they sell such premade high-side drivers?
And for one direction you turn the A-pair on, and for the other direction you turn the B-pair on.
No idea.
But these are the reasons why motor control H-bridges are usually sold as units.
Or built from scratch.
20โฌ tho ๐ฅฒ
When good quality high-current FETs can cost 4e each in small amounts via reputable company (not Amazon/Ebay/AliExpress) and H-bridge needs 4.
Well that shit I sent claims 36V 15A and it's 20pcs/10โฌ
I can buy the 20โฌ thing ig
Not like there are many options
Currently Pi is fucked as we're not getting display out, but working otherwise
Also the MOSFETs on those modules need 4.5V minimum for control signal.
And are designed for minimum of 10V control.
Most of the specs are specced at 10V control.
"power: 400W" when the spec sheet only goes to 50W for single one with 10V control at 25C ambient...
And that board has two.
So 100W from that in theory.
No, Tcase, so even worse.
So those modules you linked before aren't suitable to be driven by Pi directly, and the marketing spec is much higher than what they should actually be able to do.
Ah, whops, with that 0.011 ohm internal resistance, at 5V 15A it would "only" be dissipating 2.5W for single one with 0.165V loss.
Mistake in my head between the full current and the loss in the fet.
This perhaps? www.amazon.de/-/en/dp/B09FKXRCVY
Nice for them to take pics of broken ones... (the pins being angled from bad handling)
The motors are 24V 1.6A 22W, though L298N literally burned down
That should work, based on two BTS7960B half-bridges, which are logic level input, 16mฮฉ.
So with 1.6A current, there should be just 0.04W heating in each of the half-bridges.
And the actual spec sheet says 33A as minimum (That 43A is the typical internal protection shutdown limit).
And not dual h-bridge, one h-bridge made from two half-bridges with the normal two inputs (one for forward, one for backwards)
No idea where they got that "Stepper motor" part either in the marketing crap.
Fuck, it doesn't deliver. More expensive listings of the same product do tho.
That specific module is even sold in Finland in real electronics stores.
Such product has very poor reviews: www.amazon.de/dp/B07WT2PPWZ www.amazon.de/dp/B08JM8DCJ8
Ah, and has backside cooling on it, didn't notice.
That "BTS7960" is down to 9e in Amazon.de
Depending on seller.
Aliexpress has it for 4โฌ/piece, maybe I'll buy like 2 extra if some don't work or arrive damaged
But seems that the actual half-bridges used are chinese copies, as Infineon has stopped manufacturing.
So the actual specs can be anything.
Well, that's why I look for overmarketed shit so if it can do a fraction of what it claims it'll work
BTS7960B was Infineon product, so as the real ones haven't been available for some time, those are just chinese production copies probably.
For the two half-bridge chips used in those modules everyone sells.
I'll risk it ig, don't see many options without doing it myself, which takes too much time and knowledge, which I currently don't have.
Cannot be complete crap if Partco is selling them in Finland for 20e each.
Lets hope so
Because they actually have to provide warrantly.
For stepper motor use, you would need two of the modules.
For normal DC motor, one.
But wouldnโt this just work for pump one direction speed with pi?
Cuz 3.3V input and then whatever hz it gets determines the output
That FET is only specced from 4.5V input up.
Maybe thatโs logic power like the L298N was?
Ok, moment, need to check if it talked about logic level or not.
I might have misread the datasheet
Cuz Pi can deliver continious 5V logic power, we did so to the L298N for the 10sec it worked before burning
That doesn't need that at all.
Just the motor power and then the control / signal, need to check what the actual signal input spec is for that part.
Yes, Gate-Source voltage is specced only 4.5V and up.
I can order one piece for like 1-1,5โฌ and 3 pieces of the motor drivers in case one doesnโt work.
So at 3V it will just be starting to open and the max current is few amps if even that.
As the FET isn't really conducting yet, so the internal resistance is very high, and so the voltage loss is also high, and that causes excessive heating.
Ok, seems I was wrong, missed this line:

So should work down to 2.6V control.
For all of them.
To some extent.
Both seem to agree to some extent.
That the MOSFET is conducting, just not very well?
And the drain current will be limited with nominally 3.3V control voltage.
Which is why the actual part is only specced from 4.5V control and up.
But it also can be that the chinese copies work with lower control voltage than the real thing.
Because just one D4184A from the actual manufacturer costs 0.60e each or 0.23e each in 2500 unit reels in "retail".
And is being EOL:d, marked as "Not for new designs".
(actual end of production for EOL of such part can be 10 years after the announcement).
But this one ships from Amazon and at least ships to Finland, for 9e each before delivery fees.
www.amazon.de/-/en/BTS7960B-Double-Stepper-Driver-H-Bridge/dp/B09HGBM5D2/
The chinese are mostly talking crap, actual instructions that seem to match the boards (except the PCB with extra regulatory markings)
psc.byu.edu/00000179-d3e7-d7a6-a5fb-ffefc2af0001/bts-7960-motor-driver
Just search for that BTS7960 in the Amazon.de?
Or at another source
The aliexpress reviews seem legit, mention it arrived with the pins bent as you saw in pic due to poor packaging, but after bending them back the product works
Point was if the wait time from Ali was too long.
Oh, itโs just alright
so it was as described in the picture ๐คฃ
I need it in exactly 5 weeks and aliexpress choice is pretty solid at delivering under a month
they just made the picture with dent pins on purpose ๐ to avoid claims
Yup lmao
First one finished 
thank you ๐ only 3 more PCs to make and then all the npcs
but this one was the most complex one and also too the longest to print ~10h
How do u make the 3d models?
the schema show 5V+ Input power
isnt what it is about ?
maybe you are confusing input voltage with control voltage
transistors usually needs some tensions from source power to work correctly
and it could match that graph
Gate-to-Source is the control, Source-to-Drain is what is controlled.
Just chinese doing bad specs compared to the actual original part datasheet.
But as they probably aren't using the actual part either, but some local copy...
Problem is that you really don't want to have the FET only partially open for high power application.
That is the whole point of using PWM instead.
are those sunglasses?
no ๐ฆ
eyes is just hard
what do you mean
I mean if its input power graph, then it match board spec : input over 5V
if I'm right in that, then it's the open power passing graph, which would makes sense since it's in Y:Intensity
so as long as input power is 5V+, it's fully open
Vgs, Voltage, Gate-to-Source.
So control.
With Vds at 5V, as in the controlled voltage at 5V.
So the FET at 5V controlled, needs 4+V control to get fully open.
then what is this I(A)
Currewnt
but not gate current
Through the FET, from source to drain.
When the gate-to-source voltage is at specific value.
Mainly stepin for the RDSon.
which is in milli-ohms.
So at the 100A value the RDSon is 50mOhm.
And fully open it is only 11mOhm with 4.5V control and 5V controlled.
And with 10V control it is only 7 mOhm IIRC?
So if the chart continued to 4.5V, the current would be about 455A
With 50mOhm RDSon and 2A current through, the voltage loss in the FET would be 0.1V and it would need to dissipate 0.2W as heat.
they basically take the value for which it starts to be pseudo linear then I guess
that match 3.3
At the 3.5V Vgs the RDSon would be about 125mOhm, and with 2A current wanted through the FET, the voltage loss would be 0.25V and power loss 0.5W.
@edgy hazel this is how you do eyes.
inb4 this is a face blind shapeshifter
"idk, this is probably what a human face looks like. I guess?"
I know i dont need it but the sandisk 1.5TB micro sd card is tempting
150 MB/s ๐ So depending on the use case, maybe?
The old HDD's with 7200 RPM had around 160 MB/s. And m.2 SSD is about 7000 MB/s.
150MB/s is very good for a micro sd card
๐คฆโโ๏ธ
Have fun gaming on 150MB/s. Like I said: depends on use case.
I wouldn't event want it for my camera. But I am not going to have a discussion about it.
it's very fine
problem of gaming performance of hdd is because of random read
flash is already an upgrade
I think you may be confusing 150MB with 150Mb
which yes is few
Yeah, ofc it's so great. Everyone has microSD in their 5K gaming rigs.
what are you on about
sata ssds are enough for gaming
they are just slower for big updates (and yet epic games store is toping at 500MB/s disk speed)
all the micro sd I seen are more about 150Mb/s
than 150MB/s
150MB/s is fine
Just saying you are 100% correct. M2 is garbage. MicroSD of 1.5TB at 150MB/s is best.
bandwidth wise it works
never said it's best
the problem is write and lifetime
Its slow AF and does nothing but cause issues
HDD exists
Depends on card.
Class 10 means that it can do 10MB/s sequential writes.
HDD's are better than MicroSD for MOST purposes
microSD is king in one specific scenario: small formfactor
good luck transporting a bunch of HDDs in a shopping cart across an uneven parking lot
Also HDD's are great for mass storage
For the spec, not actually in general usage yet.
microSDs are great for tiny storage
I never seen market place here that didn't wrongly report Mb/s as MB/s
Oh are we specifically talking about portable storage?
No.
Somebody said they are interested in 1.5TB, I just warned them about the speed. And said: depending on the use case, it might be fine.
you don't consider microSD for your desktop PC
I wish samsung made a 512gb and 1tb bar
If they want to PC* game on it, I would recommend something else for that price.
Yeah
you consider microSD for your FPV goggles, phone storage expansion or stuff like that
Steam Deck.
ooooh yeah
Also why not just use SD anyways, isn't it able to be theoretically faster since more space means you can fit more IC's?
the main issue is not the speed
if speed is 150MB
for gaming
then Letterdief lost it over that
the usecase of microSD directly implies that SD isn't feasible
bro, get out of here ๐คฃ
Almost no-one is making SD cards anymore.
Most you can get are either MicroSD with adapter, or just SD sized permanent outer shell around MicroSD card
150MB/s random is decent for gaming
neither my raspi, my FPV goggles or my phone accepts regular SD
Nah, A1 is 1500 IOPS random read and 500 random write.
A2 would be 4000/2000.
๐คฃ ๐คฃ ๐คฃ bro
As in that card cannot do anywhere near 150MB/s random read probably.
That "Up to" is for sequential reads.
no I know I'm just saying that if a microSD could theoretically hit 150MB/s random read that is absolutely usable for gaming
For the cases when you don't have PCIe 5.0x4 M.2 slot available but you do have UHS-1 card slot?
and the old " 7,200 RPM hard-disk drive tops out around 160MB per second"
Bc the alternative is ~3MB/s on an HDD in a Triple A game (at best)
and that was my point
if i say anything dumb i just woke up
Even that 10MB/s the A1 requires would be better than HDD for random read.
Not my point. You can yell at me all you want. I don't see how one would say "for a gaming rig, go for this 1.5TB microSD instead of ANY other option". That's why I said: depending on the use case.
Now this discussion turned into, people saying microSD is "fine for a gaming rig". Which I totally will not agree with sorry. Waste of money, there are many better options out there.
Gaming will kill a microSD with the reads
Just purely in terms of speed, if it really actually did hit 150MB/s random read consistently, it would work
It's not only that, some people eventually use their computer for other stuff. Running a server. Coding. Streaming. Etc.
Letterdief on a lore creative rampage
People who are using their computer as a server or for streaming are smart enough to not use microSD for that
Have any of you checked how much games actually use for random read speed?
The amount isn't anywhere near 150MB/s.
Its less that they NEED 150MB/s random, moreso that it helps with streaming game assets
starts here btw #off-topic-tech message
no one but Letterdief talking about micro sd being "best"
Also what do you mean you wouldn't even want 150MB/s for your camera? Wtf are you recording 8k video?
Me?
Yes you
I mean, I kinda said something along those lines
#off-topic-tech message
Nah, I don't need 1.5TB for my Canon 5D.
Oh well yeah most people don't
I thought you meant you wouldn't use a speed "that low"
That's why I said I don't need it even for my camera.
but you didn't say for gaming and it was long after Letterdief doing "sarcasm" about something only them thought the conversation about
Nah, it's fine for that purpose. Just a bit too big for me ๐ Can't imagine me filling up 1.5TB ๐
That's what she said
btw, what you're referring to is upper midrange to highend Gen 4+ NVMe here.
m.2 can also be SATA or shitty PCIe Gen 3
So 5800X can do 5.05ghz with pbo?!
ok
And thats not being misreported
And can anyone of you even find NVMe SSD that can do 4KB QD1 random reads even at 100MB/s?
Except Optanes?
7000MB/s actual is high end
The reason I posted that, was to show you the insane difference.
even high end Gen 3 (which in context with Gen 4 and Gen 5 is basically midrange at most) doesn't do 7GB/s
and then it's still sequential speed
150MB/s ---- 7000MB/s
Corsair's 12000MB/s or whatever the number was is rediculously high and I consider that to be esoteric not high end
debatable. There are dramless MAP1602 drives that do this which I'd definitely count as great drives, but wouldn't rank as high end among the E18/IG5236 competition
MP44L, NM710, US75 and the likes
until you learn it only does 12000mbps for the first 30 seconds its on
The dramless drives are only hitting that speed in sequential tho
that's gen 5 sequential which is basically never sustainable
but even if you look at lower ssd, they still have a good speed (compared to micro).
Yup, 990 Pro 2TB can do about 100MB/s in QD1 4kB random read.
this has been about sequentials the entire time, you don't get reliable real world speed data from manufacturer specsheets
wasn't Crucial P5 2TB a low end SSD?
Because you are looking at the sequential numbers, which is almost meaningless.
I've seen a lot of people consider the P5 to be mid-range
nope, upper midrange to highend. P3 and P3 Plus are low end
Oh, my bad then.
P2 is garbage
regex