#off-topic-tech
1 messages ยท Page 94 of 1
That type of reboot should only happen when the CPU/MB firmware notices uncorrectable error, in something with ECC.
So RAM shouldn't be able to cause that, just BSODs.
So, PSU/Mobo then.
Also PCIe has ECC if enabled.
Can be enabled/disabled in AGESA.
Not sure what the default "Auto" is.
Let me check.
I'll try to find to location, cannot reboot right now to check.
Auto sets ECC to enabled.
IIRC under the AMD CBS, NBIO Common Options
Not the "ECC" next to IOMMU, that is for RAM ECC.
Which does nothing without ECC RAM.
Something PCIe something eCRC or similar.
Need to eat before I can go check.
Not sure, everything on there is put on auto besides PCIe Ari Enumeration which is disabled.
Yeah, but point was that disabling the ECC for the PCIe might cause different issues to pop up, if the problem is PCIe errors.
Also the "Platform First Error Handling" to Disabled to get reporting for some of the corrected ECC errors in Event Viewer.
And there was another setting now to report "not important" errors.
I'll check after I have eaten.
Too many people complained about WHEA errors from marginal overclocking so AMD by default now swallows everything it can and doesn't report.
and just shut down the system?
Swallow corrected errors, reboot on uncorrected.
When old was to report both and let OS freak out about the uncorrected error (BSOD with specific code for Windows).
Great reply preview discord 
If it is shutting down, and not rebooting, that is either MB protections or PSU.
It's disabled.
or psu error, only reseting part of the chips
full reset on unpexpected partial reset
My gut now tells me... PSU .... but also still GPU (because it happens when gaming, and a lot when ALT + TAB but not always) and I fear it's motherboard (because that's more work to replace).
"happens when gaming" that would more hint psu then
gpu is the main power draw on graphic games
and it is bursting usually
yeah, I keep looking into it.
still need to find a better way to make system crash.
because by this rate, I might be debugging in a few years ๐
i live in arizona, and i've made chocolate chip cookies in my rear window
So: AMD CBS:
CPU Common Options:
Platform First Error Handling: Disabled
MCA error thresh enable: False
Log Transparent Errors: Enabled
NBIO Common Options
PCIe All Port ECRC: Enabled for ECC, Disabled for no ECC.
Advanced Error Reporting (AER): Supported for WHEA in theory, Not Supported for no error reports/monitoring for PCIe in theory.
PCIe ARI Support and Enumeration go into how the PCIe IDs are interpreted, more devices vs. more virtual functions for same device.
IIRC
Yes:
The ARI is a new feature in PCIe 3.0. When enabled, it changes the meaning of the hardware's 16-bit ID on the bus, or more accurately, it changes the meaning of its 8 LSBs. Instead of containing five bits of "device number" and 3 bits of "function", the device number is said to be zero, and the function number takes up all 8 bits.
As a result, there can only be one device on the bus it's allocated on (which isn't so bad, because there can be 256 bus numbers allocated). But this device can expose 256 function entities
So in theory with those first three set like that, the errors should get logged by WHEA in Event Viewer.
And tuning the NBIO ones should be possible to tune if PCIe errors can be detected or not.
So in theory with ERCR and AER enabled, with the other ones, the PCIe errors should also be reported, but no idea if that works on AM5 or not.
At least for ECC RAM and IF errors the AGESA at least used to just swallow and handle itself even with that "Don't do it" set.
I did check RAM with tool and it didn't give me any errors.
Custom view to get only logged HW errors:
prime95.exe
But check if there have been any entries done by WHEA?
Hmm..?
Why is the first "Number of events" very high?
Because with the filter I get this:
Ah, that isn't view, just filter.
Euh ๐
No, when I did filter the top is System...
Trying to understand if there are lot of entries that aren't just shown, or just counting other events.
It's counting all system events.
Same amount of events, so it says: there are 29 223 events. and 0 get through filter.
It confused me also for a minute.
Yeah, just that when I filter, the top stays as "System", and doesn't change.
Anyway, have to go out to diner. Talk to you later ๐
FIPS 204 is based on pq-crystals.org/dilithium/
Hi everyone,
I currently have a 13700k with an old Sapphire Nitro+ RX Vega 64 powered via a 750W PSU.
I'm looking to buy a new GPU (7900XT, probably also the Nitro+ line from Saphhire).
I know my CPU can draw up to 253W, the GPU I'm looking at could draw around 370 (according to google), so rounding up slightly, that's about 630W.
I also have :
- 3 non-RGB case fans
- 2 nvme ssd, 2 sata ssd, 2 sata hdd (one of which I'm gonna retire soon)
- 1 optical drive
- 1 RGB keyboard, 1 RGB mouse, 1 USB dongle for my wireless headset
But I have absolutely no idea how much power these devices use.
My question is : How can I figure out if I need a new PSU/Do I need a new PSU ?
what's your current PSU model?
also, how often do you put a full 100% load on your entire system?
Its a seasonic prime gold 750W
If its a good 750 maybe, i have an 850 for 7900 gre + 7950x3d and that is solid
plus: Nitro+ is almost always overpriced without any significant improvement over lower cost alternatives
Nitro+ is awesome shut up๐ญ
as for full load, I don't think I've ever put prod and gaming at the same time
as far as I can remember, i've only stress tested it to see the temps under full load, and that's about it
I didn't say it was bad, just that the price you pay extra doesn't give you anything meaningful in return
that's a high end unit, you're good
It is meaningful to me
Im just sad that it was too expensive
Pulse is also great tho and i am very happy with it
either one of the two can already make your PC run full load.
however, you usually don't have that - plus, you can always undervolt
especially with new intel power limits, your CPU should steer clear of the 250W mark
I've become an XFX fanboy with the new releases, they just look so clean
yeah I slightly decreased the frequency and undervolted by 0.1. Waiting for mobo/bios update from asus
that update should already be there
and... that's the only card that does that?
7900GRE is a 250W card under full load
versus 200-220W on the 4070 Super
? No but it is still nice while being the cheapest card at the time
if it's the cheapest, that's all good
Indeed. Thanks ! Quick question about that though. Can I use a random USB flash drive or do I need a freshly formatted one to update the bios ?
it's a good card by all means
it's just stupid to shill for a brand for anything other than warranty and design
always best to use a freshly formatted one
A'ight, FAT32 I guess?
Asus supports NTFS etc. too.
And can read from hard drives connected.
๐ฎ
If using the in-BIOS flashing, not the button press one.
I thought sapphire had really good support? I havent needed any so idk but thats just what ive heard
EZ Flash # from under the Tools in Advanced mode.
I still personally use USB-stick with FAT anyways.
But the EZ Flash loads the whole file into RAM and checks it before doing anything.
Also reboots into fail-safe-defaults for that.
I'mma flash it now, wish me luck xD
Weeeeeeee!
I was starting to get worried.
Sorry about that, did some memtest as well
also, first I tried to plug my usb drive on my keybaords usb pass-through, which I realised was kind of derpy
usb -> keyboard -> pc for bios update?
yeah I tried that
couldn't find the partition, so I plugged it straight into the back of the mobo
Upon further testing, it seems the intel baseline settings (or whatever they are called) sux donkey balls.
Benching my CPU with XTU made it go to 100ยฐC. Had to limit the temp in the BIOS.
i mean it sure is unique not that its a good idea however
i find triangle to be superior to arch
beams support the arch
They got a triangle in their logo it's all goof
oh i was just talking shit since the triangle is more structurally solid than an arch
triangle man triangle man
gadzooks!
gadzooks indeed
It removed undervolting the MB previous defaults were doing.
Which caused instability for some of the CPUs.
Forced some protection features on that had been disabled by MB makers to make those undervolts work.
And fixed some Intel fuckups.
๐ Join the split keyboard side 
it is split?
Crowdstrike President accepts the Pwnie Award for โMost Epic Failโ at DEFCON 2024.
Ok, but I was under the impression that 100ยฐC was a no-no :/
Yeah I read about that.
Not sure what the current Intel max is.
Probably that.
But just as explanation why the temperatures went up, as the CPU is getting more voltage below the new limit, as that MB maker undervolting has been removed.
12TB hdd for $350 or a 18TB hdd for $360?
whats the purpose
Its going to be for a data server. Cant say what for tho.
probably the bigger the better, just make sure you dont buy a SMR drive, but also larger drive means itll take longer to rebuild an array if you use any sort of parity
Is wd gold smr?
ยฏ_(ใ)_/ยฏ
larger drives are faster (if same rpm and platter count)
not always?
well, if I just spin a smaller drive at twice the speed, it'll close up in speed
tho I'm not sure on the platter count part
desktop HDD != server HDD.
What?
harddrives for desktop are not the same as one made for servers.
so make sure to check what type of HDD you buying.
Its not going in a rack server but a server nonetheless.
you putting em in raid?
Yup. Using a raid card to do it.
it depends how platters are provisionned, if the smaller size part is about the binning of the bigger size part, then it will have similar peak performance
if it is in raid and have somewhat steady activity
it is definitely server workload
you still must check specialized vs general server hdd usecase
and smr / archiving server hdd vs database server hdds
yeah
Why WD Gold? And not Seagate IronWolf Pro NT (NT001 as end of model number)?
"Nice" that WD doesn't have full datasheet for even those drives...
Just brochure with half the info missing.
why do you call it NT
arent all IronWolf Pro with "nt" in code ?
No, older NE series.
Same branding, different drivers.
NE has lesser specs all around.
Those are starting to disappear from retail.
IronWolf Pro ST6000NE0023: MTBF 1.2M, 300TB/yr
IronWolf Pro ST6000NT001: MTBF 2.5M, 550TB/yr
I thought ne code was for non pro or something
Non-Pros are VN
then are all nt helium filled ?
(quick answer, if you dont know you dont know, dont waste time on my googleforme questions)
So the Helium NTs are 2.5M hours, Air NT:s are 2.0M hours.
NT: www.seagate.com/content/dam/seagate/en_gb/content-fragments/products/datasheets/ironwolf-pro-12tb/ironwolf-pro-20tb-DS2129-4-2311US-en_GB.pdf
NE: www.seagate.com/content/dam/seagate/migrated-assets/www-content/datasheets/pdfs/ironwolf-pro-12tb-DS1914-6-1708US-en_US.pdf
VN: www.seagate.com/content/dam/seagate/migrated-assets/www-content/datasheets/pdfs/ironwolf-12tb-emea-DS1904-21-2207GB-en_GB.pdf
WD Gold: documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/product/internal-drives/wd-gold/product-brief-wd-gold-hdd.pdf
Missing Non-recoverable Read Errors per Bits Read at least.
12TB and above are helium filled for WD Golds
So WD Golds match to IronWolf Pro NTs when comparing helium to helium, or air to air, but WD doesn't tell that bit error rate.
And for their Reds it is lower than IronWolfs.
Both IronWolf Pro:s are available in Helium.
That NE is the outgoing old models.
NT is the new incoming models.
NT has helium too 12TB and up.
no nt starts at 10TB
ST10000NT001 is still Air in the spec-sheet.
oh sorry
Ah, read this wrong.
it's the other way around
there are 2 NE 10TB
and the 10TB NT is air
one air ne 10 one helium ne 10
were they testing ?
it's the same spec but one is helium
Yeah, seems I linked bad datasheet, that only contains air models, up to 14TB.
Didn't check it fully
The full one? www.seagate.com/files/www-content/datasheets/pdfs/ironwolf-pro-20tb-DS1914-19-2112US-en_US.pdf
Because Seagate pages don't list the NE:s anymore?
But they are still in stores as old stock, so need to keep eye out.
Ok, WD Golds in 2022 had the same 1 in 10e15 read error rate as the IronWolf Pros.
When the WD Red Pros only had 1 in 10e13, when even IronWolf non-Pros have 1 in 10e14.
But "nice" that the official "spec" doesn't list that.
I expect 21 > 19 means more up to date
Yeah.
1e15 means for full disk read 20TB you have 1/6.25 odds of hitting unrecoverable error ?
Each should be independent.
But it does matter for RAID rebuilds after drive loss, if there isn't two+ drive redundancy.
I dont really understand the context of that error rate "read"
does it applies regarless of if you read the same bit repeatedly or if you read others spreaded bits
if it wasnt independent we could not simply adding up the bit count
I'm counting 160T tries for a full read
AFAIK that for each read, there is that chance that the drive has to abort trying to read the data without a result and return just read error to controller. (after retries etc. up to the default limit).
that does not explain context
And if one hits while the array is degraded, all hell breaks loose.
SW raids usually just lose that specific block.
HW raids often just barf and stop rebuilding.
Lower chance better, not certain about the specifics.
i.e after successful reading of a bit, it is unlikely I believe it will become unrecoverable on next read
(i.e, is it head error ? firmware error ? platter error ? platter decay error ?)
Just combined probability by the manufacturer that reading of sector fails for any reason without permanent drive damage.
Because current drives already need extensive forward error correcting to not fail on all reads.
Big reason for move from 512B sectors to 4kB sectors, to lessen the amount of total data area going to the error correcting codes.
AFAIK currently it is very rare to be able to read all of block correctly, and the data is just parsed together from the data and ECC by the drive, before passing the resulting full data to the computer.
there is 2 stages I believe ?
the platter coding, handling some sort of "error"
and then the 4k block overhead to correct 4k data error (this one being reported potentially ?)
a bit like in line transmission
where link have data stream error correction in signal coding
and that data stream hold ecc headers in whatever protocol frames it use
anyway I mean 1e15 sounds very bare minimum reliability for 20TB+
Ah, yeah, that 4kN move was to lessen the amount of space used for the outside-of-drive ECC info.
And then the real magic happens inside the drive with another set?
And seems my BSODs might have been because of AMD Enhanced Sync.
Need to reinstall the security solution later to confirm.
But I got the same "All Edge windows flashed black" without the ESET, but with Enhanced Sync, that has been the precursor to BSOD so far.
And after I turned Enhanced Sync off, no crash so far and no new flashes.
Thing that switches between VSync On and Off depending on FPS compared to max refresh rate.
I would give anything to get BSOD 
This one doesn't help much, as it just tells that something with kernel level priviliges has trashed kernel memory structures.
Even when tried to run automatic analysis on full kernel dump, not just minidump.
how badly implemented vsync toggling can be ?
amd : yes
just start ending tasks in task manager until windows implodes
sounds like frame buffer interrupt and memory management is doing some use-after-free shit on vsync toggle then
Seems to be new thing in couple of the latest driver versions.
The old OS install had 24.4.1 or 24.5.1 driver.
And now on 24.7.1.
Not hit when Edge wasn't running.
If I quit all Edge instances before on that window flash, no crash until after next window flash.
So something weird between that and Edge specifically or Chromium in general.
Chromium in general doesn't like VSync Off.
When I was troubleshooting my PC/Build, I read a lot of suggestions that Vsync and AMD don't play well together.
Stuff break with even base Chrome if you force VSync off always.
Nvidia seems to just ignore owner settings for that for the broken applications.
AMD doesn't.
There is reason for that "Not Recommended" disclaimer.
Current default is Off, unless application specifies.
I had just cargo-culted that Enhanced Sync from before owning Freesync monitors.
if it is bsod : interrupt timeout / watchdog trigger
then it may be it is busy spinning to death
kernel_security_check_failure in few variations.
But all are about the kernel noticing trashed data structures like broken linked lists etc.
KERNEL_SECURITY_CHECK_FAILURE (139)
A kernel component has corrupted a critical data structure. The corruption
could potentially allow a malicious user to gain control of this machine.
Arguments:
Arg1: 0000000000000003, A LIST_ENTRY has been corrupted (i.e. double remove).
Example from net from analyze return.
Like I said, causing detectable but uncorrectable corruption to random kernel data structures.
Several different subvariants of that Arg1, that #3, LIST_ENTRY is just one of them.
And I have been getting at least two of the subvariants, possibly more.
basically
Need to make bug report to AMD once the computer has been stable for long enough.
First without ESET, then with ESET.
But off to bed ->
๐
Those damn a******
I am switching to another ISP and the current one decided to change my pppoe credentials so I can't use my own router to access internet
Now I am forced to use mobile internet since you can't change ISP modem back to routing mode without calling tech support
And since it's saturday, tech support does not work so I'll have to survive until monday on mobile internet....
I guess that's how they punish customers for moving to another ISP
The problem is, they just changed my credentials and decided to do it on weekends when the tech support is not available
Because they know I am moving to another ISP and this is how they mess with you 
Well, I don't need pppoe credentials anyways, I just need the tech support to change ISP modem back to routing mode
I could send a mail but you are not getting an answer until Monday lol
Just reset it to get it back into routing.
My ISP modem doesn't work that way unfortunately, resetting it won't change the mode from bridge to routing
The new ISP I am moving to at least has a customer website where I can change modem to routing/bridge mode without the need to call tech support
Did you reset it from software or hardware button?
Tried hardware reset too but doesn't work
Ah, sucks ๐
I am starting to think they did this on purpose because they know I am moving to another ISP and I use bridge mode atm
I meant, they didn't send you mail? Maybe it's in your spam ๐
No mail ๐
Fuckers.
Is the modem the same as the previous ISP?
Not sure how it works in your country. Here, we got two main ISP (Telenet and Proximus). And they both have their own hardware.
I mean, I haven't moved to a new ISP yet. They are doing the migration atm (which can take up to 30 days)
So, if I go from Proximus to Telenet, I need to get new router from them.
Yes, but it's the same router for both?
You mean ISP modem?
Each ISP has it's own modem. I don't think they even have the same brands
Oh, an ISP gives you a modem/router combo if that's what you mean
And if you want to use your own router, you need to switch ISP modem/router to bridge mode and connect it to your router
Yeah, no I though I might confused you by saying router. So I went to look up what the difference is.
modem: internet to house. router: internet to wireless devices. A router needs a modem. A modem doesn't need a router.
I didn't know that ๐
I think a correct technical term for a network device which has both modem and router is a gateway (correct me if I am wrong)
You are correct.
Basically, a gateway is an all-in-one box with integrated modem and router
- VoIP ports and IPTV ports if required
Never though about it.
My ISP offers some cheap Ubee gateway where you can't simply change to routing mode by hardware resetting the device
At least the new ISP allows me to change the mode through their website without the need for tech support
Basically this (translated version):
That's cool ๐
Hope this ISP is the one for you ๐
Wasn't it 'recently' that you changed ISP?
Doing it right now ๐
Btw, I am also an employee of the new ISP I am moving to so you can pretty much guess why I am doing it lol
*besides an old ISP simply being crap
Oh, it's still the same ISP. I get it.
Yeah
Fiber is going so slow in my country. It doesn't matter which ISP I go ๐ฆ
And regarding fiber, it should be available by the end of the year unless my crazy neighbour decides to protest again....
She prevented ISP from installing a street cabinet on a sidewalk near the apartment building even though ISP had all the building permits
Oh right, forgot about your neighbour ๐
ISP called the police, they wrote down the report and she got sued
I don't know anything about the lawsuit but I guess she will be crying like crazy when she gets her ass handed to her in court
I suspect she will pay at least โฌ15.000 or even more
court costs + compensation for ISP
Yeah, some people here behave like the whole city is their property
And you can't install anything on city property without asking the citizens first
*under the assumption you got all building permits and approval from the city
When they cut down a forest. I raise an eyebrown. But for a steet cabinet, I wouldn't care much tbh.
Wish I had a picture my colleague showed me, they were installing a telephone pole and some crazy old man climbed on the pole and prevented the workers from attaching fiber cables lol
A while back, two teenagers died in Belgium because they dared eachother to climb electric pole next to railway.
Oh, a lot of teens die here too
3000 volt โก
They are dumber though. They climb on top of a train and get zapped
I don't think they were sober either. Clouding their judgement.
People think you can't get zapped if you don't touch the overhead lines and that's why they do it
Being even 1 meter away from a high-voltage line is enough to get electrocuted
It's so dangerous.
Ofc, train company had to explain why the poles were accessible and not fenced off. This is the spot where it happened. So they had to climb over the concrete.
So I don't think the company is acountable. Looking at the image. They did try to seal it off and protect it.
It's just that the middle fence misses a piece because of old age.
I am even afraid to walk with an umbrella during a rainy day at the train station because you can hear an umbrella making a buzzing sound lol
I rarely use umbrella ๐คฃ
Lucky
it rains a lot, I just never use umbrella ๐
I also just realised my current ISP has a problem which could be classified as GDPR violation....
Basically, you can easily get someone's OIB (personal identification number), including their name and surname in order to check if they have any requests lol
Assuming they are using the said ISP
Here in Croatia, you can easily obtain someone's personal identification number, name and surname through cadastral online portal if they own a property
Maybe you can report it to EU
I should considering all the problem I had with the ISP 
I just tested it on my neighbour who we had trouble before and I can at least see he has an active phone line
No details but only an info there is an active phone line
Not sure if that counts as GDPR violation since there are no details but things like that shouldn't be available to the public ๐คฆโโ๏ธ
Don't need to log in, you can already see the information by inputing identification number, name and surname
Also, there might not be any requests if the person didn't send any
that's what I meant. Verify.
But you can use this tool to check if someone is using an ISP and some other additional data which I am not sure if it violates GDPR
To the left of the screenshot you can read what is possible no? It says "complaint"
No complaints in this case ๐
maybe make one to test it out ๐
I am still reporting them to see if it violates GDPR lol
Storage
Can I borrow some storage? 
F F F
new optic internet is causing frequent disconnects
i have no idea why - there's 0 physical movement of the network
and it already caused be some pain
fiber ?
maybe a missaligned bit of fiber when two fibers were joined together. had that happen to me.
Yes, for a price [:
My Patreon: https://www.patreon.com/buildzoid
Teespring: https://teespring.com/stores/actually-hardcore-overclocking
Bandcamp: https://machineforscreams.bandcamp.com/music
Twitch: https://www.twitch.tv/buildzoid
Twitter: https://twitter.com/Buildzoid1
Facebook: https://www.facebook.com/actuallyhardcoreoverclocking
#overclocking #intel #undervo...
Public Service Announcement from the Rambler about how to undervolt Intel CPUs correctly.
TLDW: Don't use the main VRM VCore offset or AC Loadline (Includes how to do loadline change correctly). Use CPU VID Request offset, which is much more hidden.
Baldur
I just done the actual calculation
obviously as the 1/1e15 is pretty small, the drift from linear dumb analyzis to actual exponential probability calculation is not that much
going from 6.25 to 6.75
anyway I've done the real calculation and ..
5 20TB 1e15 uncorrectable rate in raid6 (3 hdd storage volume)
still means only 96% odds to be able to read it all without unrecoverable data
this is :aware: moment
imo 1e15 is too small already
counting each bit of the array as independent
Yeah, reason why single-drive redundancy just doesn't cut it anymore and two or even three is the norm.
Seems WD has updated their Red Pro lineup to use same drives as current Gold.
Used to be < 10 in 10^14, now the same < 1 in 10^15 as Golds and Ironwolf Pros.
Red is SMR and Red Plus is < 1 in 10^14, like IronWolf non-Pro
Personally planning to buy 6 smallish drives and put them into RAID-Z2 (so RAID6 equivalentish).
But when even 6 smallish IW Pros costs almost 1000e...
4TB or 6TB drives on last look where the drive cost per TB stops rapidly dropping.
yep that's why I have a st6ne
2x 6TB NT in RAID1 in main computer for bulk storage.
Well, Storage Spaces Mirror.
With ReFS and data checksums set to on, which isn't default.
(1 - (10 ^ (-15))) ^ 160000000000000
0.852252771878 ok reading odds for full disk 20TB reading 1 pass
that's 0.852252771878 ^ 5
0.44961636694 (i.e unlikely) for 5 hdd
0.852252771878^5+0.852252771878^4*(1-0.852252771878)+0.852252771878^3*(1-0.852252771878)^2
0.541074979561
for raid6 ??
what did I miss
the grass you're supposed to touch wtf
Not going to even try to comment.
- the success of fail is probably by 4kB sector?
But again, don't know.
ok I think I understood what's wrong
it must be evaluated by strip
i.e if there are 150000 unrecoverable but they all are not on same cluster
it's fine
going to need doing birthday collision calculation I guess
this is more complicated to guess than I thought
but giving me hope on 2 disk redundancies
it should make it extremely more reliable when it comes to unrecoverable errors
Talk about how to calculate etc. in first, second is old "scaremongering" article from past but contains the basics.
there are so many confusion and mixup in the first link that it makes me cringe
especially that big message explaining how the hdd recovers unrecoverable errors Kappa
Yes, and I'm not commenting who is right or wrong.
This isn't that well known area.
it still makes me scared though, because bit reading statistic is very poorly defined
which means that it could be likely to run into corner cases were it suddenly increases a lot
especially point that "average over time" comment
if that is how this is calculated for spec sheet
it is completely wrong
the spec sheet must report a safe error rate for any local in time operation of the disk inside it's warranty period
this is yet another example at how stupidly "statistics" are understood and used
Yes, and the Oracle persons talk about how long a drive tries to retry until URE is passed back is part of why consumer drives cannot be used in RAID anymore.
NAS drives have much stricker limit on that max time to not get timed out by raid stack.
Consumer drives can currently try for 30+ seconds and be completely unresponsive for that.
IIRC the old CMR WD Reds I have had 7 second default limit.
oh btw
the "if data is good but crc is bad, then data can simply be outputted as is" made me chuckle
How would you know which one is bad ?)
yep
But anyways, UREs are not really issue with with non-degraded RAID-equivalents, but can be issue during rebuilt depending on amount of redundancy and total storage size.
And too long drive internal retry time can kick the drive "out" and cause major issues, but is very much wanted for single drives.
Which is lowered by default on NAS marketed drives vs. consumer ones.
the oracle person comment made me think of something at least
: you want an external rebuild connector
i.e you connect externally the new drive for rebuild, letting the current array be safe as is, until the spare one is ready to provide increased reliability (be fully built, up to date, and checked)

I could even make that work.
But it would be just boot to Linux with extra kernel parameter that then gets passed to userland to launch the game as soon as possible.
i like how people are figuring out that the 0x129 microcode is doing basically nothing
Just limits the extra useless VID request spikes to 1.55V max. from 1.6+V before.
it ain't ๐คฃ
My Patreon: https://www.patreon.com/buildzoid
Teespring: https://teespring.com/stores/actually-hardcore-overclocking
Bandcamp: https://machineforscreams.bandcamp.com/music
Twitch: https://www.twitch.tv/buildzoid
Twitter: https://twitter.com/Buildzoid1
Facebook: https://www.facebook.com/actuallyhardcoreoverclocking
#intel #z690 #msi #14900k
Ah, that MSI fuckup, didn't look into it.
For Gigabyte it worked fine.
As long as you didn't then disable the Intel Defaults, because then the "fix" was also disabled on Gigabyte
Got to that part, so some kind of MSI voltage regulator failure.
Straight voltage, then to huge weird wobbling, then little lower straight voltage.
Some kind of ringing in the VRM feedback.
Causing excessive output swings in wave type form instead of normal spikes.
Possibly some other trace affecting the VRM feedback monitoring line (die sense).
So even that MSI one was now correct, except for the MB side VRM control failure in the thumbnail.
And that is on MSI, either the board design, or Rambler has broken board, more likely the first.
Discord Windows app is coded very shittily, news at 11:
https://x.com/da_wamwoowam/status/1824872498363523537
Running powershell commands to do system information gathering in very bone-headed way.
If i made a raid through windows raid and had to transfer the drives to another pc, will the new pc auto detect the raid array?
The raid software directly through windows
In most cases yes.
That chipset raid least likely, depends on the two chipsets moving between.
There isn't one.
Dynamic Disks are through Windows Disk Management.
Storages Spaces is through more hidden options in Windows, but clearly labeled as such, and show up in Task Manager etc. as Storages Spaces.
Any others are not from Microsoft.
So what is the device name in Task Manager, Performance, Disk in question entry at top right?
Or what is the name of the software you used?
RAIDXpert2 for AMD currently, Intel Rapid Storage Technology for long time for Intel, but with lot of versions that might or might not be compatible.
And no compatibility between AMD made and Intel made.
For those chipset raid ones.
The Dynamic Disk and Storage Spaces should be fully movable as long as they are never booted with just part of drives connected in any computer.
any good editors in here im just looking for advice, ive been looking for a way to remove the music only from any sound without ai but so far no luck
so that's kind hard to manage even with AI, however if you have access to the music you might be able to invert it and lay it over the top of it, which is effectively how noise cancelling headphones work. I can't promise good results and I obviously have to condone any illicit or piracy uses of this
no i dont
in which case I'm not sure if it is possible really, but someone else may know of an AI that can attempt it
hmm
as far as im aware every ai editor ive been to costs a fortune and only lets you edit 10mins (for free) of footage which make the editing time 3million times longer
Have you tried editing out the voice and then inverting the audio like Kitsuneko suggested ? Maybe there are some karaoke like software that can help ? No idea if its gonna work.
not possible without either using some ML algo or having access to the music (only works if the music itself isn't some sort of background music like in a store or at a wedding)
otherwise, nope no chance absolutely not
Do I really need to head to bed before 19:00?
I'm just completely out of steam already...
go read a book
Too physical, if you meant reading in bed or similar?
just anywhere
Go see a doctor, check your blood.
if you woke up at 01:00 then not really? but if you woke up at like 06:00 id be concerned
Cause is completely known, but will not be talking about it on public channel.
ah that makes sense
Baldur smoking too much of that good good
herg you would know a thing or two
Hehe
one million beer
Why does it need to be Oktoberfest?
Who needs a once-per-year occasion for drinking?
You mean "Who needs an occasion for drinking?", right? ;P
That sounds like alcoholism
Not of it's water, or orange juice ๐
not if you can quit at anytime
That sounds like an alcoholic coping
7900XT : Pulse version for 750โฌ vs Nitro+ for 850โฌ, which would you take and why ?
Is a i7 14700kf and a rx7800xt + 32 gb ram good for the game
(Updated bios and underclocked i7)
How much wood could a woodchuck chuck if a woodchuck could chuck wood?
Pulse cuz 100โฌ cheaper and nigh zero performance difference
a wood chuck could chuck a lot of wood if a wood chuck could chuck wood
For satisfactory, definitely ๐
Why are you talking about Chuck's wood?
You'd want 7800X3D for Satisfactory
๐ซฆ
das gay
SF scales incredibly well with AMD's 3D vCache - plus 7800X3D is a lot easier to cool than 14700K
Also: "32GB RAM" isn't particularly specific enough.
You'd want DDR5-6000 for AM5 and DDR5-6400 for LGA1700 (or rather, as fast as you can reasonably get)
If you like manual OC, just get 5600 CL46 Hynix JEDEC RAM and overclock it yourself.
For first two, CL isn't that important, it's primarily an indicator of the IC used
DDR4 I don't think ddr5 is worth the extra speed I don't have 650$ to drop on a motherboard/ cpu or the time to swap it
14700KF with ddr4 is kinda nonsensical - especially given the fact that there's up to 10% performance increase from ddr5 over ddr4 (even more in very memory/cache heavy scenarios like Satisfactory)
Anyways, there is a lot of super important context information that is not being communicated here.
What's your current hardware?
Like what specifically?
I have
Motherboard B660 DS3H DDR4 rev. 1.0
Teamforce 32gb DDR4 3200mhz
You mentioned spending extra on new motherboard and RAM, so that means
- You already have a PC that you are currently upgrading
- You have a limited budget in mind
Both of these are important factors to consider when telling you whether the decision you're about to make is a smart one or not.
Yeah, that board ain't gonna run a 14700KF at any meaningful speed
I assume CL16 for the RAM, CPU probably 12100F or 12400F
The chip is peaking at 95 on performance, 80 underclocked with updated bios and I have a 360mm aio
That sounds straight copied from chatgpt
๐
You sound like a client request. Our rate is 230$/hour
I'm talking about power delivery
It's called autism
And even if it was a ChatGPT output, I'd double-check the info and verify that it's correct
Sorry it wasn't meant to be an insult
Thanks ๐
i have a 7900 GRE pulse and it is very good
You might find another non-Sapphire option for even cheaper, that can also be worth a consideration
The most important skill when picking PC parts is to sort by price ascending
True. That being said, I'm buying second hand, because the retail prices are beyond ridiculous here.
Like you can add 100โฌ for the same cards as above, but brand new.
and finding lower prices is basically hitting the "I'm feeling lucky" button.
And buying used card at so low price difference would be just pure idiocy in my eyes.
If used really is 750/850, and new would be around 850/950.

Because of this file?
sources.debian.org/src/systemd/256.5-1/tmpfiles.d/home.conf/
github.com/systemd/systemd/blob/main/tmpfiles.d/home.conf
Added 9 years ago.
--purge
If this option is passed, all files and directories created by a tmpfiles.d/ entry will be deleted.
Basically the default SystemD has been set up for 9 years to create /home if missing, and never touch it otherwise.
But that --purge removes all files with tmpfiles.d entries.
q /subvolume-or-directory/to/create mode user group cleanup-age
Q /subvolume-or-directory/to/create mode user group cleanup-age
Seems that that --purge is very recent addition, either 255 or 256, not in 254 yet.
It is recommended to first run this command in combination with --dry-run (see below) to verify which files and directories will be deleted.
Warning! This is usually not the command you want! In most cases --remove is what you are looking for.
Added in version 256.
Maybe, but the lower prices don't ship, and as I don't have a car, I sure as hell ain't paying for train tickets.
In what?
That explanation wall of text with bad translation, so I couldn't understand most what you meant.
Windows doesn't boot.
You get that BIOS boot selection menu at one point.
But Windows doesn't boot.
Otherwise it was just gibberish.
I'm not native english speaker either.
But that had marks of bad machine translation from bad base text. ("nova button"?)
I need to understand what is happening to say what might help.
And you cut off the part that would tell what button that is.
But the point was that I expected that that "nova" was some untranslated word from your own language.
So that is just "boot to boot selection" type button with Lenovo specific branding.
And you don't have the System Recovery option in there or something like that.
And no, this channel isn't general technical help channel.
And I don't have time today to even try to sort this out to even level of understanding what is happening and if you have the tools to do anything about it.
Because most likely doing anything would mean at minimum full OS reinstallation, possibly drive replacement.
And you haven't specified how the laptop has been damaged either.
As in what happened...
Not what is causing it to not boot.
But I would expect from the symptoms that the classic type hard drive is broken.
And would need to be replaced and the OS reinstalled.
The system recovery lives on the drive too.
I am stuck, and for some reason I thought the satisfactory discord was were to go.
I am trying to connect some PC fans to a wall socket via old adapters.
Fan: 12v 0.16a
Adepter: 12v 1.5a
Fans work when I test them with a common 9v battery, but I tried 2 adepters, and redid my wiring a couple of times.
Also used the markings on the cords to determine+/-
What am I missing here?
Should work, if the adapter outputting the voltage, and the connections are right.
But impossible to say which is the faulty point.
From here, if you don't have multimeter or other way to test the output.
I'm finally back I think
You aren't sure where you are?
I'm not sure where everything else around me is. Everything is relative.
He is in a quantum superposition 
I've just been wildly busy, working 4 different jobs in addition to freelance software development
Lots of 60-80 hour work weeks
I have been keeping an eye out and am currently planning a PC upgrade around 6 months from now, with current cpu + mobo + ram going to a nas/das
I'm waiting for Radeon 8000, but it looks like a used 4080s after 5000 series launch might be my best bet
I am going to murder my ISP 
They probably know I am moving to another ISP, messed up something on my modem which is in bridge mode and now I can't use my own router since saturday
Been using 4G atm which is meh....
"customer retention"
The modem is so bad you can't change routing mode to bridge mode and vice-versa by yourself
Only tech support can do that
So, I am stuck with bad mobile internet until friday when they must switch the modem back to routing mode....
What's hilarious is that the tech support could literally do that in like 10 seconds during the phone call before unless you decide to move to another ISP
Then they have to send a ticket to "management"
I got so pissed when the tech support guy literally asked me: "Couldn't you turn bridge mode on modem by yourself?"
Maybe if you give normal modems that don't have almost all settings locked for customers, I could....
I have been doing finishing touches on 560 line / 15k character javascript thingie in Notepad for last few days.
(Including comments that are probably about 1/3rd of that).
Refactoring and rewriting large parts.
On my free time.
I must say, I'm extremely pleased with the new laptop
I think it was a very good purchase
And price to performance is about as good as I think you'll get in a laptop
I found an ipod older than a lot of people in this server ๐
Is it older than you?
Oh yeah, I had one of those at one point
shows the charging icon so fingers crossed
what laptop did you get and what did you pay for it
the ipod boots
and shows up in itunes
theres no way it remembered the date and time
Got synced from the iTunes
Autosynced on connection.
guessing that charge was caused by me?
also 230 an hour is wild
2005? Holy fuck, that means I'm ancient
fr
well yea
This might look like nothing but considering there is rain atm and mobile internet is very bad here, I should be lucky I managed to pull more than 3 Mbit/s
Upload is much higher than my dsl line lol
Managed to get over 100 Mbit/s on my phone in the morning when nobody is using mobile internet
Also wasted 1 hour trying to setup my 4G router until my dsl line is fixed. Figured out a router doesn't automatically configure settings once it pickups up mobile signal
You need to manually choose ipv4 and type in APN
Strange, I'm reading through a syslog trying to figure out why my server crashed, and at some point for whatever reason the kernel decided to just dump the registers into the log
0-100 brightness adjustment for the monitor backlight.
Need to run at 5 or 6 for 100cd/m^2...
Recalibrating my monitors and dropping the max brightness from previous 120 to 100 cd/m^2.
I don't get how monitor calibration can work when most panels control and monitor communication are basic ?
is it you always get high capabilities monitor Baldur ?
I didn't find much reliable information on true monitor calibration
i need to adjust mine to be a bit more brighter, as it seems it may be too dark compared to most other people (images i make that look fine, are 'too bright' for other people)
No, don't have monitor side LUT.
But basically first the monitor settings are adjusted to be as close to the wanted result that can be done, and then the graphics card is adjusted to account for the rest.
Using light reader device(s) placed on the screen.
Spectrophotometer to read the backlight profile mostly, and then colorimeter to do the actual calibration.
As spectrophotometers are slow, and cannot read near-black that well, but can read light in 1-10nm slices to get complete picture.
And colorimeters are fast, but have much worse spetral resolution and usually only have three set color filters.
With high-end monitor, there would be adjustment table that can be programmed into the display to convert the signal to panel response.
But with monitors without that, you just change the computer output instead, but if you can only send 8-bit signal, that will limit.
But as these displays take 10 bit signal, that isn't such an issue.
so how to make sure it is using 10bit display/stream ?
as otherwise the 8bit resolution is hurt a lot
fair
on my way to check how I can see that with nvidia/windwos
yea
8bit on my side (reported by windows)
Not visible even on 8bit if the monitor is already close.
Backlight spectra for one of my monitors:
it will, because the only way to adjust 8bit content to 8bit screen
if by dropping some steps
so you mean the graphic driver is smart enough for the graphic engine floating point color point conversion calculations ?
or for things like jpeg in higher resolution colors
And profile stuff: basically the shown that all-white, instead of 255,255,255, is sent as 249.11, 250.98, 251.98.
As the backlight is little too bright for 100cd/m^2 target.
bit-edpth to bit-depth conversion loose at least 1bit of data if you are close
and 7bit is not great at all
going from 10 to 9 is much much better
and play 8bit lossless
Nothing visible on my 8bit only monitors when I have tried to look at problems with grayscale or color gradients.
Old 8bit native IPS monitors.
Remember that it will just offset values by needed value.
So 255 can become 254, 253 etc depending on how much adjustment is needed, and then next value will be almost same offset in most cases.
So you lose some values, but not that many, and there shouldn't be more than one missing value between in some spots.
If the monitor isn't just horrible.
yes, but missing one value on 8bit depth is a lot imo
when on 10bit is completly unoticeable
But still only a problem if it becomes visible issue.
now if graphic driver is smart, you dont loose any value on anything floating or downscaling (10b+offset to 8b, graphic float to 8b, etc)
though I dont know for these
imo that kind of stuff should be able to be negotiated in monitor by graphic driver
there is so much quality lost by guesswork on either side
when for same hardware it could be transparent
I want the monitors to push panel profile to graphic driver
would be great
and optionally give a configurable values table
including the usual stuff, like brightness setting
Already done as basically only file in monitor drivers, ICC (color calibration profile) file.
But as there is so much variability between identical monitors, that is just "somewhat close".
icc is limited
I'm talking, pixel shape and size, color and light time distribution, zone behavior, pixel color time response
And in theory there is standardized API over HDMI/DP etc. to adjust monitor settings.
No-one just bothers to support it.
Some of these are part of the info the monitor sends to computer via EDID.
But for lot of the fields usually just contain crap defaults that don't match the monitor...
everthing so the graphic driver could make educated decisions if it wants to, without asking the monitor to process anything more
I'm tired of the default craps, when it ask nothing more in the hardware to simply push real values
Yeah, if the fields were missing, at least you would know that if the data is there, it is usable.
Now you cannot trust what the monitor reports for physical size etc.
imo monitor hardware and graphic card drivers
is under used
because of a mountain of historical architecture crap
converting a to b then to c then to a then to e then to b then back to a
when it could just report that A and go on
but simply reporting A will probably get patented by Nvidia or something
clown world
Clown world where the minimum brightness of these monitors is about 90cd/m^2...
Boeing grounds 777X test fleet after failure of key engine mounting structure:
theaircurrent.com/aircraft-development/boeing-grounds-777x-test-fleet-after-failure-of-key-engine-mounting-structure/
only thing that I recall is that mozilla and firefox account merged
poop
No cut of price for the new gddr6 4070 /wccftech.com/nvidia-geforce-rtx-4070-gddr6-launch-same-gpu-specs-price-performance-tweaked-g6-memory/
Factual
Ok, I could get accurate sRGB colors in Windows HDR mode, and even wanted SDR brightness.
But the contrast goes to absolute shit because the backlight is on full blast, so the minimum brightness is up from 0.09 to 0.30 or so.
Giving overall contrast ratio of 315:1 for SDR in brief test...
heres something i have been wondering about the intel microcode update
does it only actually have an effect on the default presets they have been roling out or are thoes seperate things?
there seperate right?
At least for Gigabyte, the VID request clamp only worked with Intel Defaults enabled.
great so basically my options are have a limit of 100 something wats or just not care...
But Gigabyte also exposes manual setting to set clamp to whatever value you want (has for long time)
What? Why 100W?
gigabyte set there default preset way bellow intel spec unless they changed it again
No? Unless you are thinking of that one long pulled update before Intel actually specified the defaults.
The actual specs that should be selectable with the new ones:
maybe that was a bit ago but i remember when the defaults started coming out people where mad at gagabyte for the intel defautls preset staying under 200W i cant recall the exact number i want to say 165w
Probably on the Baseline, which Intel doesn't recommend if the MB VRM isn't complete crap.
As there isn't one "Intel Defaults", there are 9.
2 for most CPU models, and 3 for 900 not KS:s.
i wanted to ask because yea i havent done any of the bios updates since a few months aftera built it
i tough i did one a few months ago but apparently i only preped the usb i nerver actually did the update
Asus exposes all of those at least.
Gigabyte based on Buildzoid videos does too IIRC.
MSI I think didn't.
Haven't looked into AsRock and Buildzoid hasn't done a video with one of their MBs.
Even MSI might have but had named the better ones as MSI <something> Performance and MSI <something> Extreme, didn't look close in the video and Rambler didn't investigate.
im guessing this is what i need
But if that is the case, that is MSI fuckup if they don't use the highest of those Intel sets for the specific CPU that the MB VRM can support.
its the most recent and its the only one i see with anything microcode related
That is the latest with all three planned actions in it.
First was implementing those "Intel Default Settings" variants.
Second was microcode/firmware fix to eTVB behavior.
And third is that 0x129 microcode with the VCore VID request spike limiter.
Which was released on the day that AMD 9000:s came out, so the reviews for those couldn't be run against the new BIOSes, that nerf performance.
Compared to the original "undervolted to hell" defaults.
And with the eTVB fix limiting max clocks above 70C correctly now.
Not allowing for the last 200MHz of boost anymore when CPU is hot.
ok its updating
it updated i think
oh great they changed the entire layout
and they got rid of multicore enchancement so back to 100C central i go
and i dont have the preset golden told me to set so now my ram is gonna be unstable again

i fixed the intel defautls bit but the ram is gonna be for another time
because i want to test it again before fixing it again
i really feel for y'all with raptor lake
honestly the only painfull bit iv had is having to reconfig the entire bios each time other than that no real issues... all of them where caused by other things acting up
Damn, no more password sharing on older TV ๐ญ
Netflix finally noticed password-sharing protection doesn't work on older devices and decided to cancel support
sad
i did end up contacting gigabyte support and basically the conclusion is the board has no MAC adress and the only real thing they suggest is a RMA attempt
assholes
kinda reminds me of this:
now i'm waiting for one of the big streaming services to buy up the patent for disk drives, then not make any anymore
Fuck windows
idk which i hate more, evil windows or incompetent windows
That Secure Boot bypass vulnerability was blocked 3 years after it was fixed in Grub2?
Again...
That should have been long enough for distros and people to update to fixed version...
It's also happening on patched systems
Ok, MS did fuck something up, as that patch wasn't supposed to be applied to any systems with dual boot:
This SBAT update will not apply to systems that dual-boot Windows and Linux.
Poland Can Into Space?
www.axiomspace.com/news/ax4-countries-crew
Axiom Mission 4 to ISS will include India, Poland, Hungary
i don't think that's the first pole in space
Yeah, multiple during the USSR times on Soyuz.
and i think there's some polish-descent US astros
Ok, only one in 1978, second one was just standby for that.
And third one is in ESA waiting for a turn.
But this was just to the "Poland cannot into space" countryballs meme.
oh wow that's an OLD one ๐คฃ
Bro showed his age
Just a minimum.
Various other ones too.
And even: https://store.steampowered.com/app/441050/Polandball_Can_into_Space/
Your main objective is to help Polandball here! Break through Earth's atmosphere and reach for the Moon to prove other countryballs once and for all they were wrong and conquer the final frontier!This is not going to be an easy task as you have a huge target on your back!Crazy Features - KURWA!!! Upgrade Polandball's Rocket Build a strategy...
$2.99
441

This old CCFL backlight is just so pain to work with when calibrating...
Unfortunately DisplayCal doesn't like the newest Argyll CMS, so the colors are missing from that image.
But there is one line per the three primary colors, and then combined total.
That center green peak has additional small peaks left and right of it, and the blue is very wide, so the green blue-side peak messes majorly with the blue, and to lesser amount with the red...
Old HP ZR24w.
So direct light for green and red, but phosphorus for blue, and that has been steadily losing the battle to age.
But that monitor has 47k hours on in by the inaccurate internal monitoring (inaccurate clock).
that was so pog to see this meme in tv show in such a serious tone
this is bullshit since password sharing protection have very little to do with the devices
this is most likely about drm
That's weird because I noticed Netflix immediately detects you are in a different household when using a new device for watching Netflix. When you use an older device (in my case Panasonic TV) , it won't detect a different household
it's likely policy
i.e : you relogin -> you get kicked
they probably dont want to break pre prohibition logged devices
Or just finally depricating something else that is needed for the very old program version, when the TV manufacturer has dropped support long time ago.
Which is why Smart Anything is currently so broken concept, when the support periods for the software are so short compared to the lifetime of the device itself.
I have a 1998 Camry as my daily
my sister had a 2004 camry for 12 years
Had to do little rust repair and replace cv joint & boot to pass mot this year
never needed a thing
Wish I've looked earlier into Rock Auto in the USA for parts, they had both left and right whole cv-axles shipped to finland for about $280. Did not have time to wait shipping from USA ๐ค
It's fun to stay at the Y A M L
Finished transition to a new ISP today. Mobile phone number has been transferred and I am already impressed by 4G speed compared to my previous crap ISP
Waiting for a technician to come with a modem so he can setup vdsl too
Hopefully, I will be getting fiber soon so I can have those speeds ๐
And at least Asus is doubling down on having just 2 PCIe slots and ton of M.2 slots on all their new AM5 MBs...
For the high-end stuff.
The low end has more x1 slots.
But again set in a way that they cannot be used with GPU...
What is the point of that 16x physical, 1x electrical PCIe slot right below the GPU slot?
Just have all the slot-position M.2 slots below the GPU slot, and the normal PCIe slots at the bottom?
How hard can that be?
Because there aren't lot of PCIe devices beside GPUs that need more than 1 slot of space...
Asus - in search of incredible
Water
You ain't putting 2 high end GPUs under each other without suffocating them
Only the top slot is 16x electrically.
All the other three are just 1x:s.
On that MB.
But point was, why aren't the PCIe slots at bottom, and the M.2 slots below the GPU slot?
M.2, GPU, M.2, M.2, PCIe, PCIe, PCIe.
Instead of the pictured:
M.2, GPU, PCIe, M2. PCIe, M.2, PCIe.
Because they don't have good receipe for M.2 ssd cooking ?
In most cases AFAIK the M.2 drives benefit more from the forced airflow from the GPU than they suffer from the higher air temperature.
Tech guy just came with a new modem and activated vdsl2 for now until I get fiber: much higher speeds than before ๐
Before, I had 43 Mbit/s download and 7 Mbit/s upload on another ISP
im still faster tho ๐ญ what cable does bros router run on
"sir did you illegally modify your fiber cable to get speeds you werent paying for?" - ISP famous last words
At least I am finally getting what I paid for. The old ISP used the exact same line but couldn't deliver more than 43 Mbit/s
And I am going to roast them for turning off my internet before transfering to a new ISP. That's literally not allowed according to the law
didnt the fcc rule that internet service providers would have like
When you move to another ISP (using the same existing infrastructure), the old ISP can't turn off your internet until transfer has been completed
nutroiuns (im tired) facts like labels for them
almost like those
so you couldnt hide most the fees
That reminds me of an incident I had with my ISP a couple years ago. Lost fiber for 8 months.
Yes, because in US the ISPs liked to play it so that the actual price was just pure profit and any actual costs were separate hidden fees...
of course
they are gonna maximze profits off you
after all its the land of the free, why stop the isps freedom from harvesting the most cash dollars off you?
@twin dew I just got back from vacation (gone for a week) and now my fiber is a tenth of what it should be
If rebooting the router and ONT doesn't help, then almost certainly something ISP side.
Thanks
Already rebooting router
How long does it take for the er605 to boot up?
Power light on and nothing else
Ont ethernet light is flashing now
okay its back on now
thanks!
You are one of the lucky 90% of people who can fix an internet issue by simply reseting the router
I am always 10% of the people who have a complex issue 
Just that even low end "enterprise grade" gear shouldn't need such reboots...
Which that TP-Link ER605 is.
My former ISP modem required reset from time to time and especially after power outage
Since the internet would work exactly 24 hours after power outage and then it would stop working unless you rebooted the modem
Weird tbh. Never seen anything like that
Yeah, consumer grade and ISP supplied gear can be crapshoot.
But stuff sold to companies for the company use should be better.
Not that the current one is any better though (Arcadyan Magenta home box)
Forgot to ask ISP for my new pppoe credentials so I can use my own router
Got a spare iPhone 12 Pro Max from my sister since she doesn't need it anymore (also managed to scratch the screen lol). I should test 5G speeds on it
Looks like I won't be needing to ask ISP for my pppoe credentials. Username is already visible in the modem configuration site and the password was hidden but I used the inspect tool and changed input type=โpassword" to input type=โtext" in order to reveal the password ๐
my dorm has 10gbe.. I can't use it cause I only have a 1gig nic on this laptop ๐ข
And I'm still only paying for 50/50, when I could get up to 1000/1000 for 35e more per month.
Getting a Pixel 9 Pro XL, what's everyone's experience with cases like this, where there's a ring stand/grip attached?
https://www.amazon.com/gp/product/B0DC5MTMVP/
https://www.amazon.com/gp/product/B0D86Y8SVC/
Some dirty ISP's love to use a trick where they offer customers 2 Gbit/s on fiber and don't have to worry about oversubscription ratio anyways because most devices have networks cards which support 1 Gbit/s
And that is the minimum speed available so you are actually paying more for a speed you can't use due to hardware limitations
This is fiber to the building and ethernet to the apartments.
Old build so the ethernet cables cannot do more than 1Gbit.
IIRC done in the 1990s.
When my house was built, an idiot electrician decided to cover the phone cable in cement instead of installing a conduit....
And I had to drill holes through ceiling so I could put ethernet cables because of that mistake
That was the norm until 90s or so.
And sometimes used even after that.
My house was built in 1988
quit flexing!
*Flexes with OLT instead of home connection 
Wait, 73 Mbit/s is not enough. I need more!
And according to this, I should have more lol
I suspect it's because of vdsl profile and ISP doesn't want to pass max speed since it could make the connection unstable
Yeah, I was thinking it might have been because of that
For the at that moment measured conditions.
At least my connection is perfectly stable atm
With ADSL and VDSL, you can in theory adjust the amount of margin.
Too much and you have less speed than possible.
Too tight and you get disconnects on the phone cable link that break the internet for 30 seconds or so each time.
My previous ISP offered me some cheap Ubee modem and it's a complete joke
Yeah, I get 250 GB per month on my sim card and there are discounts on services (internet and mobile plan)
I can also buy smartphones for much cheaper prices (without VAT)
Finnish mobile service providers currently really hate that they switched to unlimited for everyone long long time ago when no-one used mobile internet much yet.
As they cannot roll back without huge uproar.
I once used ChatGPT to win an award for the 15 fastest employees to answer questions related to euro 2024 lol
The only condition was to finish the quiz in 1 week and there were no restrictions regarding AI
So I used ChatGPT and won some football prizes lol
And I have no clue about football
Would just like more money.
Remember, even with mobile flat monthly usage amount is just about money, not about service limits.
As the price of a bit transferred when the network isn't maxed is trivial.
So only the peak times actually cost anything, mostly in upgrade costs.
eSIM made a huge impact here (unless you visit a country where not a single operator supports eSIM)
So even my bottom of the barrel "4G at 20Mbit/s max" contains unlimited amount of data in countries where the ISP has their own network, and then 29GB/month in rest of EU.
As the data roaming costs between mobile service providers are still high inside EU, just not as insane as they used to be before price regulation.
Oh, I remember the look on my face 10 years ago when mobile data was really expensive
Now I can get a week of unlimited surfing for only โฌ4,99
I remember the good old days in 2005 when flat rate (unlimited data) broadband internet was really expensive and I used to check how much data I wasted every day lol
Oh, you want a tourist sim card with unlimited surfing for 10 days? That will be โฌ10 please 
Finland never had anything but "pay for the max bandwidth" for anything after landline and/or ISDN, where you paid by time connected.
Wait, internet cost was based on time spent rather than the data wasted? ๐
Old legacy phone line modems.
Like any phone call.
Oh
My parents had to cancel a phone line because my sister was always on the phone and caused a huge bill
That was like early 2000 I believe
I remember when adsl was being offered to customers in 2000 or somewhere around that year and it was considered a revolution back in those days compared to dial-up
I'm not a specialist, but you seem to have some serious signal attenuation. How far are you from your ISP nearest DC/repeater?
Perfectly normal for VDSL2.
And the speeds are about half the theoretical max for very short line.
Oh, for vdsl2, dslam is installed in a street cabinet
Normal phone cables just aren't designed for the high frequencies used by VDSL2.
I think the distance to the cabinet is less than 500 meters
Probably using 17a, which tops out at 150/50 theoretical max for runs less than 150m with very good quality legacy phone cable.
IIRC for the distance.
Yeah, ISP replaced all the copper wires in the city 10 years ago since they were in terrible condition
That 30<something> VDSL2 profile is not really used by anyone.
If they didn't replace the wires, there is no way I would be able to use vdsl at all
Only in less developed countries like germany
Before they replaced the cables, my attic had a phone box where thick wires were connected together with screws
Now it's thinner wires and punch-down telephone block
Because phone providers were forced to wire up everyone with phone cable, and there still isn't similar requirement for fiber in most countries.
So only the "fast profit" places are wired by the legacy operators.
And they fight tooth and nail to not have to wire anything if they don't want to, and to prevent anyone else to do it either.
As 5 year ROI is way too long now...
And with fiber, the ROI might only be 10-20 years.
For some places.
Return-on-Investment.
How fast you get the money back you used to build the infrastructure.
Because any money you get from the subscriber isn't profit until that point.
So with 5 year ROI, you build something, and your company books are in red for 5 years before you start getting direct financial benefit from the investment.
So for most current companies, max ROI for anything is in 0,5-2 year range.
Costs more in materials and takes more time on the initial installation.
Just makes repairs and replacement much easier, but that isn't benefit for the installer.
this man above me, cat2 react him.
Found the exact thing you mentioned ๐
starting to think everyone upgraded there internet at the same time lol
i personally 12x'ed+ my speed just from swapping to ethernet
*their
this is why everything is going to go to hell, because long-term thinking doesnt have impressive ROI, so everything worthwhile in this world where growth is ending is not going to be done
yeah people will eventually adjust but the process is going to be ugly and much will be lost
In the same way that thief is not supposed to be worth it but often is these days (short term)
long term planning have better roi in the end, but it takes into account more things than the digit dance of marketing "lessons"
'digit dance'. lol never heard that one before. i'm stealing it ๐
you think my m2 mac will like office 2011
i have a physical license key and i dont wanna deal with the subscription BS
Truenas isn't supported on arm (yet) but it looks like open media vault is. Think I might revisit my pi nas project
The PCI bus is so limited idk if it's worth it
Anyone got any tips for good monitors? Budget: max 250 euro's. 1080p, purpose gaming, 24/27 inch?
I bought asus VG248QE back in December 2015 and I'm still using it ๐
But it's a little bit outdated by now, there are probably better models which have recently been released
Not sure if asus actually made longer-lasting products back then or I am just lucky
Not to mention asus strix 970 (in working condition) which I sold 2 years ago because I was upgrading my pc
I got one ๐
the budget is also 250 MAX. My niece is 15.
i bought another laptop so no more nuclear battery
And she got my old PC. Which has a 1080 card ๐
That doesn't say much, it comes down to the games she plays
Yikes, that's a scam
200โฌ for 165Hz 1080p
And how many of those people know how to tell a good monitor from a bad one?
Don't use that store then
These are not only better, they also cost 119โฌ and 125โฌ respectively
cheapest one, full HD
How bout you don't then?
enough, since the reviews are all about gaming ๐
That doesn't say anything
Let me look into it.
Do you know how to tell a good monitor from a bad one?
I used tweakers.net set to Belgium for these
I used tweakers already.
They only have best buy guide* about "general use case monitors"
not gaming monitors.
Then how come you came around with garbage?
Not for reviews, for price comparison
Because I have that monitor and I love it ๐
Ideally you'd look up reviews of the monitor from a trustworthy source. These are a few examples:
- Hardware unboxed/monitors unboxed
- tftcentral
- rtings (measurements only)
- techless
- aperturegrille/a5hun
- pcmonitors.info
- optimum tech
Not exactly that model, but close.
and my monitor had awards from some tech websites.
Anecdotal evidence is bad practice
You do know that seemingly reliable tech websites also sell those awards, right?
And "close" can mean that they swapped the panel for an entirely different one - which is the equivalent of buying a whole different monitor all together
look
4K and Quad HD.
What is my niece suppose to do with a 4K monitor ๐
I will figure it out ๐
AOC 24G2SPAE, no reviews on tweakers.
You never even specified the system she has let alone the things she does
Skill issue
#off-topic-tech message
Customer reviews can be widely ignored, except for analysis of commonly occuring issues (which can still be user error).
The people writing these reviews often either don't have a comparison at all, don't know enough to tell good from bad or compare it to something wildly outdated and significantly inferior
Nah, I am not ignoring tweakers. They are leggit, the reviews.
Fuck that "great value" bullshit thing, that's not how you find good electronics picks
I know, was just looking around ๐
The professional reviews they do are solid albeit not great.
The customer reviews are bullshit
Then again, tweakers are just one reviewer of many
I wonder how healthy a 1tb HDD from 2008 is
Didn't know you were dutch Fireworker ๐
I'm not
tweakers.net/productreview/255078/iiyama-g-master-g2470hsu-b1-zwart.html (example of full user review)
they add photos and everything. It's pretty professional.
But Dutch isn't too far from German and scientific data doesn't need a language
It really is not. They collect zero actual data and just tell you how it feels to use
A good professional review gives you all the data you need to make your own conclusions
This is an opinion piece, not an actual review or test.
It's a glorified unboxing
Look at the reviews from these sources and you'll see what I mean
Yes, they bombard you with data that you don't understand but that means that they actually care about giving you a good conclusion.
"I tried it, it's nice for Trackmania" is not a review, it's an opinion from someone with zero trustability
The Iiyama G-Master G2770HSU-B1 Zwart also looks alright for 169 euro.
Exception to this is rtings. They have some weird algorithm to calculate the "conclusion" and gives you bullshit results.
Their data is good, what they do with it isn't
First two are an instant elimination unless you sit over a meter away from the screen
I did btw ...
I got 27inch with 1920x1080, no issues with gaming ๐
And? It's still not recommended to use.
Also: your opinion does not matter
And? Those guides mean absolutely nothing
and 18 reviews from users. Making it one of the more popular monitors on tweakers.
From THREE YEARS AGO
well, I have to pick something
That's hella outdated
all I read here so far is two 24inch screens
I named four monitors.
All four of them are better picks
and my niece knowing she wants probably 27 inch.
Popularity is not an indicator of quality, otherwise Thermaltake Smart would be a good PSU
I feel like we've been over this before ...
You did not, at least not since you asked about monitor recs
Either you made up your mind, then don't ask for an opinion, or you are open for advice, then be ready to discard what you believe to be good
No, we've been over the "reviews don't say anything" and if I don't read reviews y'all are like "read reviews"
Ah yes, because Valorant at low settings is the exact same as Cyberpunk with Raytracing.
"Gaming" is not a usecase.
Reading reviews is still better than going in blind.
Collecting extra data is even better. And listening to opinions too.
There are customer reviews (meaningless, ignore) and professional reviews (good, read them)
I said she had 1080 card
you think that has raytracing? It's my old build ๐
No. Going in misled is not better than going in blind
Unless you meant the GTX 1080, that's not a GPU
And in the context of the discussion, "1080 card" means "card that's best suited for 1080p resolution"
I'm saying that the info you provided is incredibly lackluster
And not enough to give any meaningful advice based off of
looks modern ngl lol
Then you just didn't give enough info leading to this misunderstanding
Then we gotta upgrade ๐ฆ
i wonder how a data soicentst from the 1980s would react if i showed them a 1tb microsd card
Don't say "1080 card", say "GTX 1080"
and how much it would be worth
RTX 2060 equivalent, depending on the games she plays that's still capable of 1440p
