it's your fault Firefox crashes and you should upgrade your trash PCyou are hurting Firefox market shareplease stop
>>108337867Thanks for sharing. I ran memtest recently and I was shocked that my RAM was faulty. I had to order overpriced RAM... not cool
>And for the record I'm looking at this mostly on computers and phones, but this affects *every* device. Routers, printers, etc... you name it. That fancy ARM-based MacBook with RAM soldered on the CPU package? We've got plenty of crashes from those, good luck replacing that RAM without super-specialized equipment and an extraordinarily talented technician doing the job.sucks for soldered ram fags
I wonder if memtest has been recently ``updated'' to include malicious code that reports RAM problems.
I was going to say that you'd see crashes in more than just the web browser with bad RAM, but then I remembered that most people only use the web browser on their computers these days.
>>108337940idk but I noticed there are scammers who use some shitbox AMD platform for testing ram used mostly on Intel platforms and they crank it up beyond spec and they're like>ayo dis ram be defective yosqueezing ebay sellers to get some money backwhen you sell ram that you never had issues with you should triple check the specs the buyer is testing at, most likely it's a scam, happened to me twice, sold 3200 rated ecc ddr4, then he cranked it up at 3999mhz and complained it's bad, meanwhile at 2133 2400 2666 3000 3200 had zero issues through 9000 stress tests.it's getting niggerish out there, stay safe.
One time I got faulty memory and didn't realize it for maybe a few months.After it got replaced, I spent the time to memtest and overclock it to the absolute maximum. It took like a week because of how long it takes to show/find instability during the tests.My bios offers the ability to backup your settings, so of course I did this. But surprise, the backups are binary files and they are not compatible with future versions. So all those settings got lost and I never had the time to fine tune it again.I'm still mad about it.Thank you for reading my blog.
>>108337867Linus said something similar when he was a guest in a LTT video, he said a huge amount of user issues are caused by bad RAM. He said one of the most important things in his setup is ECC memory
I got crashes when my memory overheated so I dropped the voltage down to a conservative 1.35V
>>108338073I believe this and it kind of scares me how it's /not/ a problem the majority of the time. Like we're running on a fundamentally unstable platform and just don't care because it's good enough. But also it's extremely fucking important that it not fuck up too, so ???
>>108338088ram is either stable or it isn't stableif it's unstable you fix your settings or put a fan on the sticks
>>108338073>a huge amount of user issues are caused by bad RAM>how come just firefox crashes and everything else runs fine?>shut up goy cattle! go spend 600 bucks on new ram preferably with rgb leds
>>108338109I think this Anon >>108337986 makes a good point.
>>108337867Surely firefox will print a warning, telling you that your hardware might be faulty, in case of such a crash with detected bad memory, right?It doesn't just upload it over the internet to some twitter (mastodon) faggot to write an essay about it, right?Otherwise, if the user doesn't check, it could be one of those rusting Intel CPUs or overheating laptops or over- or underclock.The RAM is not the only way how to get a bit flip.
I'm absolutely sure that modern ram with huge densities is infinitely more vulnerable to cosmic ray bit flips too. Same goes for processors with lots of cores. Manufacturers probably preferred to ignore the problem for regular consumer stuff because they consider a random crash every few months on average isn't much of a big deal, but if you add up all the millions of devices around the world, even a tiny probably can become super noticeable.
>>108338109I don't think Linus (TORVALDS) was trying to sell RAM sticks to me but what do I know
>>108338045>it's getting niggerish out thereAlways has been, butt thanks for the perspective.
>>108337867Faggots who don't care about static discharge assemble their PCs with no strap on carpet with their socks on. Static damaged components fail intermittently and unpredictably.
How does bubble memory fair in regards to error rates?
>>108338127Oh my bad, i thought you mean the other Linus
>>108337986Browsers are one of the ultimate ram tests because it involves lots of context switching for the CPU memory controller. A tab will eventually crash or the whole system.
>>108337867i know you are being sarcastic, but bad memory causes a lot of issues.suffering faulty 14900k from intel's "LETS NUKE OUR OWN CPUS BY SHIPPING THEM WITH FAULTY CLOCK TREES" the symptoms always looked like faulty software. random crash, random corruption (especially shown when archiving and un-archiving zip and rar files). did the game crash because of a bug? or bad hardware? idk, the symptoms are identical. ended up being the cpu was bad. when i had my first gen ryzen system, was a ton of memory issues. leave the room for an hour, come back to the entire system frozen. games randomly crashing. whats the issue? software? or bad hardware? ended up being ryzen's memory controller on the 1000 series was absolute ASS. now i finally have a rock solid system. but if you ever start getting random quirks, crashing, overall issues, check your hardware first. always blame yourself before you blame others.
>>108338149We should assemble our PCs nude as Lain intended.
>>108338122During a ram or CPU error the printed results cannot be trusted because the hardware is literally wrong about the calculation. Also the system will likely crash before you even see it.
>>108338195While not wrong, it's implied that their system captures and sends results after such a state.Which makes sense. In the event of a single incorrect bitflip rarely, potential results might be better than no results.You can imagine normalizing it too with the same telemetry ID or even across memory SKUs, etc.
>>108338109you under estimate how many people just use their shitboxes as glorified internet cafe terminals.
>>108337867I have good RAM and Firefox has not crashed in at least a decade. Checks out.
>>108337867DDR5 is more temperature sensitive than DDR4. Some kits throw errors as low as 55-60C which are laptop temps.
>>108338212>so how come Chrome never crashes?>goy i said shut up and buy the new ram!
>>108338109>>108338244Chrome definitely crashes harder because it's more hw accelerated than Firefox. Ram context switching isn't free it's coming from your CPU & GPU.
>>108338244In my experience Chrome is infintely worse in this regard. That's always been the case and continues to be the case.I'm not exactly thrilled with using Firefox but I sincerely do not like using anything based on Chromium. Gecko/Servo may not be the best yet, but I'll take it.>goyI don't think you know how to use that word. Mozilla doesn't sell RAM and in fact they're notorious for having high memory usage and leaks. It behooves their brand's image to reduce memory requirements, not increase them to the point you need to buy more.What I'm saying is your joke isn't funny and you didn't even try to make it funny, you're just writing nonsense. Come on man.
>>108337867I know that Linus Torvalds is using like a bit-correction RAM?But where can I find it? Are all RAM sticks like that or do I need to specifically search for it?Is it uncommon?Is it slower than normal RAM?Should I buy RAM like that instead of normal RAM?
>>108338267https://en.wikipedia.org/wiki/ECC_memory"ECC"
>>108338232Anyone building a new PC needs a ram fan now.
>>108338195If they were capable to upload a crash report, they are also capable to print a warning to the user.
>>108337867DDR4 and especially DDR5 really aren't as stable as previous generations used to be tbdesu
>>108338244Chromium crashed much more for me than firefox.It crashed so comically often that i just ignore it and hope an update fixes it. And it eventually did...Firefox is slow. Firefox hangs sometimes. But i didn't see it crash.
>>108338281The crash report is automatic and a unstable system isn't exactly reliable at doing automatic things.
>>108338264>goy>I don't think you know how to use that word.whose fault is ram priced at 3x? i'm asking seriously, if you follow the money, whose fault it is?>OpenAI executed two unprecedented RAM deals that took everyone by surprise.The secrecy and size of the deals triggered full-scale panic buying from everyone else.The market had almost zero safety stock left due to tariffs, worry about decreasing RAM prices over the summer, and stalled equipment transfers.hint: > born in Chicago, Illinois, on April 22, 1985, to a Jewish American family.
>>108338307>whose fault is ram priced at 3x?Not Firefox's.>>108338305That statement is pointless since it circles back around.The crash reports are automatic and executed during a point where the system is unstable.Meaning this happens anyway. It's 1 more function to call in the chain, messagebox("I'm dying"), that's it.It should work at least as often as the report feature itself.
>>108338322It can still generate a crash report but when a human looks at the report its all gibberish because the hardware was literally wrong about all the calculations it did.https://devblogs.microsoft.com/oldnewthing/20050412-47/?p=35923
>>108338343As I said before, that's only the case some of the time, and you can normalize it in various ways to derive value from all the samples aggregated, you're not just sampling 1 computer one time, that wouldn't be valuable anyway since it's an isolated incident.If this is a real fault that really happens consistently, you should be able to derive a point from it with enough samples even if most of the data is garbage most of the time.There will be commonality between them that you can surface with enough samples.
>>108338358> that's only the case some of the timeits 100% of the time with cpu or ram error.the secretary and ledger are both wrong.
>>108338144>>108338045>>108337940
>>108338362All I can say is that's obviously not true. Mozilla and Firefox are not the only nor the first people to do this.This is basic statistics and analytics.I'm not really sure how or why you're trying to argue this. The existence and prevelance of normal memory versus ECC memory adds credence to this alone, then we have Mozilla saying it's the case, and OS vendors like Linus, Microsoft, etc.This is hardly a controversial idea or atypical practice.A machine that is 99% stable is still capable of doing something. You're acting under the pretense that each machine sampled is 100% unstable 100% of the time. I think you know yourself that this is not how computers work, neither at the hard or soft level.It's actually fine if the bit for dark-mode flips incorrectly once because a cosmic ray hit your box, and something like a stack validator found it. The bug report is going to go through after that, I promise you.Make your own test cases for this if you don't believe it, write a program and corrupt its memory and see how far it can go in execution to what percentages and ranges of memory you corrupt.
>>108338244go back to /leftypol/ with your antisemitism
>>108338267ECC is a lot more expensive then regular ram tho. Generally ECC is not that useful for regular users ( Compiling the kernel code for millions of users,important servers or some other extreme "mission critical" use case. is outside this 'normal' scope) If it irks you, you can do a memtest and see if you have any bad blocks in your memory and there are ways to isolate these bad blocks. You might want to look into that :)
>>108338415This made me wonder if the cost of redundancy would be cheaper than the cost of ECC.I.e. just build the kernel twice and see if there's a discrepancy.Obviously this doesn't cover things like mission critical servers responding in real time on-demand, but I still wonder about lowering the requirements for things like builders.Maybe it ends up costing more, who knows.
>>108338125cosmic rays dont cause bitflips. that was just an excuse a mario 64 cheater said for his speed run
>>108338439My guess is no. even purely from a compiling standpoint it takes power and time, and it adds up over time. plus reviewing to see if the code is identical takes up extra time as well. you could automate it. but then how can you be sure the automation would be 100 correct if it doesn't use ECC?
>>108338494akshtually it was most likely hardware fault as he later said he had to insert his cartridge at a certain angle for it to even work
>>108338591how convenient it gave him an advantage if the cartridge wasnt making good contact it wouldnt work instead of giving a bitflip
>>108338625>how convenient it gave him an advantageit didn't, it upwarped him in TTC but mario fell back like a turd and took damage slowing him back down (and iirc he was going for the red coins which were BELOW anyway)also who the fuck would cheat like this live on stream>if the cartridge wasnt making good contact it wouldnt work instead of giving a bitflipcartridge tilting on N64 famously makes funny glitches like characters spazzing out and earrape noisesI couldn't tell you the specifics of why though
>>108338095thanks mr intel faggot keeping ecc workstation-only for decades until amd's zen, you absolute moron cuck
>>108338415>Generally ECC is not that useful for regular usersare you intel shareholder?> and see if you have any bad blocks at that stage faulty data already entered your system and corrupted other data. gg you dun goofed
>>108337867ECC? You don't need it.
>>108338809Oh no anon you made me panic sweat so much my kippa slipped off :C I'm on your side actually all ram should be ECC Ideally. still would be useless for 99% of normies two things can be true.
>>108338809I know it isn't a drop in replacement for ECC. just wanted to help a fellow anon doing the next best thing. relax nerd
>>108337867I am not paying for more RAM right now.Did DDR5 include some amount of ECC or do you have to buy an overprised server CPU to get it?
>>108337867>firefox is le bad!yes but chromium is le worse
>>108338979>Did DDR5 include some amount of ECCon-die ecc otherwise it woudn't work at alltotally unrelated to sideband ECC as talked about by OP> or do you have to buy an overprised server CPU to get it?that's just for Intel.AMD Zen all support ECC, but the mainboard may not.
my pc is in the area of a decade old with 99% uptime that entire time so i don't even know what the fuck to think of crash reports and slowdowns anymoreyou pump electricity through a plate of melted sand for 10 years what else do you expect but friction breaking a few atoms off
>>108338182No, in this case you blame Intel for selling faulty CPUs.Intel is also responsible for pushing the consumer market away from ECC.Probably because Mossad needs Rowhammer for their backdoors.
>>108337867he looks european so his opinion can be safely ignored
hes not wrong and iv been saying firefox crashing is because of your bad memory since before that tweet.
>>108337867No, they need to fix the memory leaks in Firefox
>>108338073bad software crashes ram
>>108337867>webdev blames hardwareThat takes me back to the good old days of the webdev general where a 12900K was their solution for websites running like shit. That generals a ghost town now, full of AI bitching just like the rest of /g/
>>108340377horse shit AMD has had ram issues up the ass since time immemorial and still has ram issues up the ass
Fuck off negroes. People have been dooming on Mozilla for decades now. I will NEVER stop using it EVER. It’s still the best. Firefox + ublock origin is literally all you need. You can improve the UI on your own too with a few lines.
>>108337867thanks op those were good threads
>>108338278ironically wendell from level1techs did a video on this last year i believe. some mobos have a bracket for point micro fans at the ram kit. granted this was more epyc/threadripper not the ryzen stuff
>>108343287amd always had chipset issues. theyre great for price/performance but they had trouble getting thunderbolt speeds lol
>>108337867Ah yes, the good old "it's your fault not ours" plea.I mean, i haven't really had crashes with FF in many years but still, why spend time doing this instead of addressing the issue and making the browser more resilient to memory errors? Why yelling out it's tje users' fault?Mental illnesses I swear.
>>108343981Firefox has been shit for over 12 years.
we should post memtests
>>108338045>it's getting niggerish out there, stay safe.American culture has been half judaized and half negrified since the 1930s or maybe earlier.
>>108338088>it's extremely fucking important that it not fuck upName any person who ever died due to faulty RAM.
>>108344065Me.
>>108343981how the fuck would you do that? do redundancy in user space which is completely retarded and double Firefox RAM consumption while still not addressing anything that would actually crash the browser instead of a tab? are you retarded?
>>108344040I think you’re ignoring elephant (worshipper) in the room.
>>108337907>digikey/mouser lookup for datasheet for the chip>set hot air gun to appropriate heat according to part specs>flux and heat edges around chip until it comes looseJust go for it, absolutely nothing will go wrong. You can always repair the bonding wires and silicon die if something does break. If people can do it in a dirt floor 3rd world workshop for youtube clicks, so can you.
>>108338088>on a fundamentally unstable platform haven't CPUs gotten so fast and optimized that a lot of calculations are off by pretty worrying amounts? its just not really enough to matter since perfect accuracy isn't needed most of the time
>>108345552Unstable ram will return wrong data when it does matter THO, eventually.
>>108343278>That takes me back to the good old days of the webdev general where a 12900K was their solution for websites running like shit.lmao
>>108344040Thanks carl now tell me about the giant penis in the basement of the church
>>108345552I think it all basically boils down to the difference between high and low signal processing. Sometimes the hardware will have consistent ranges and stick within them, sometimes you push it beyond the limits and start having spikes in voltages, or maybe some component is broken with some electrical fault, etc. whatever interference in whatever way that fucks up your ability to interpret a signal as binary patterns / data.This is around the layer my insight and abilities break down. More knowledgeable of software than hardware.
>>108337867Could it be that this is an issue with Firefox only and not Chrome, because most Firefox users are retards running Loonix on decades old Stinkpads?
>>108338073He uses a macbook air with linux on it doesnt he? I dont think they have ECC
>>108345977Could be a correlationIDK how many times you see some retard complain about their 10 year old SSD being "chewed up" and it's always a Firefox user.Most of the complaints I see from the web being "bloated and slow" is from fucking Firefox users on old as fuck hardware.
>>108337867I can confirm. My Firefox crashed twice a day, and RawTherapee sometimes froze. I ran Memtest86+ just for fun, and after a few seconds, red [FAIL] appeared. I replaced my faulty 8 GB DDR3 RAM with a secondhand (but new, factory-sealed) 16 GB DDR3 RAM, and I haven't experienced any crashes on my Ubuntu since then.
>>108338861if blaming Intel is correct, then Apple Silicon Macs should have lower rates of memory-related fuckups. I hear they don’t have _all_ of ECC’s fuckery protection, but they have some.Do Apple Silicon Macs have fewer general memory errors?
>>108337867makes sense to me I have firefox running for 16 hours a day and I don't remember last time it crashed
