Hardware Debate

[thorn lagoon]

Please use this room for any hot topics so they aren't clogging up the main hardware room.

This goes without saying but Keep It Civil.

[gusty fractal]

Im gonna enjoy that one. XD

[true seal]

#sim-and-pc-hardware message of course I only showed the dependencies between the groups of brands and the flow of parts and did not evaluate the companies themselves, Asus is above AsRock not next to not below. Currently in all markets not only hardware there are many accidents of this type, 4090 with melting plugs? But why write about it? We all know how it works. Quality drops.

#sim-and-pc-hardware message obvious abuse in order to create a shitstorm, not that it is only can show errors during e.g. playing and pushing it to the limit of efficiency 90-100% sometimes even less, maybe and not that for sure, we only wrote about the increased risk of errors. Looking for topics to fight by force is very infantile, unless I write it so incomprehensibly that the meaning escapes, I'm listening.

[gusty fractal]

Warranty scams

Repair scams

Melting CPUs

Mosfets blowing up

Bios voiding warranties....

Using Intels hardware defect i225 LAN on the mainboards....

Its loooooooooooong list

[true seal]

Do you see any factual errors in this statement?#sim-and-pc-hardware message

[gusty fractal]

4090 is a cutdown chip with cut down parts inside.

Which would make your argument extremely funny cuz Nvidia wouldnt be allowed to sell any 40 series chip.

[true seal]

True. There are many brands that, changing their place of production or wanting to be more competitive, make a mistake and destroy their reputation at the expense of quality. Asus will recover from this. Probably, although the market will not forget them quickly.

[true seal]

You misunderstood, maybe it's my English that's to blame? We wrote about the production stage, not production planning. We wrote about the fact that huge amounts of products appear on the production line that don't pass tests and my first statement was in this context, i.e. we have in the Amazon link with a GPU that has a reduced amount of RAM, I assume that players are writing here and buying this GPU for SC or other games, so I wrote that in the case of a large load of such pre-sorted equipment, errors may still occur. Are you now clear that you are writing about the stage in which NV planned something and did something else? I am writing about the next step when the GPUs are already produced.

*I will not defend any brand/manufacturer/product here, for me personally what matters is the end result price/efficiency. I also do not tolerate private attacks and insults in discussions, keep it for your peers or someone who tolerates it and do not direct this tone at me - just a note for the future. I consider having your own opinion to be good, I do not forbid anyone from doing so, we can even agree that we disagree - as long as it results from a constructive discussion supported by facts, always.

ok, I understand, I won't hide the fact that I'm writing this quite quickly, I don't have enough time to precisely reduce each sentence in my head to the level that after translation will refer 1:1 I'm trying but I already have other super skills XD. Referring to your question. I thought I wrote it here already but ok, I'll repeat it. The production and testing process and selection are not 100%, load tests are not long-lasting, if you expect that even a top manufacturer will torture each Chip for weeks, no one does it like that, the risk of return/warranty costs is calculated. Considering that today's example 7900 with 20GB RAM is already a waste after not very heavy tests, as we wrote here with Brinnen, there is an increased risk that a player who will play and load this GPU to extreme values ​​will experience the problem of replacing such a GPU unnecessarily. To save you from that, I suggested not considering the higher risk GPUs, especially as AMD products have gotten cheaper and are becoming easier and easier to spend smaller amounts of money on - does that sound reasonable?

[gusty fractal]

This makes zero sense but feel free to tell that to yourself.

Used multiple 7900XT's for someones AI stable diff rig...
24/7 stress and not a single issue on them despite VRAM being fully utilized.

[true seal]

but it's such a statement with zero sense about nothing, why write it. You can think differently, what do I care, if you think differently you simply have the right to do so. There is no such thing as objective opinions, each is subjective based on experiences, on the sum of information that you obtain on any topic, I don't know why this persistent greed is storming shit.

[true seal]

and here we have two things, first of all I did not write that 100% GPU with 20GB RAM has some failures. I wrote that there is an increased risk and I gave why I think so and even from my experiences some such facts result from it. The second issue is not only about filling the number of RAM cells. It is also about the intensity of downloading and sending data there. Graphics in games burden memory cells with permanent life, it is a good stress test if you play for a very long time. Therefore, for professional server solutions, e.g. in banks where any failures and errors should be minimized, ECC memory (Error Checking and Correction, Error Correction Code) was created not because they were bored but because there are different quality standards and different values ​​of calculations that you perform on such bones. Players do not require it, you will see a black screen a few times, well, frustration is the maximum you will experience.

[true seal]

I can't agree with that. I'm sorry, but there's no logic in that. If a product fails in the first tests, you can't compare it to a product that passed the tests. What you wrote is like this: if I ran down a small hill and fell/I can't fall, then if I run down a bigger, steeper one, I won't fall. It's completely the opposite - if I can run down a small hill, it makes sense to run down a much steeper one. That's why XTX, passing basic tests, has less chance of making a mistake. After all, that's what tests are for.

and adding infantile icons won't change that

lol

you've already shown the lack of logic and the lack of a dictionary and now the childishness, what else do you want to show us?

don't worry about me. I have no problem understanding what you are writing, but I suspect that on the other side there is resistance and some kind of fight for the sake of fighting, which you wrote about earlier. For me it is a senseless waste of the day and its resources.

[gusty fractal]

Just ignore that guy.

Its just random waffling with marketing stuff mixed in.

Not worth my time explaining someone the basics of binning and QC when he thinks TikTok is the correct source for everything and cant be wrong.

Its amazing how wrong someone can be and still think he is always right.... Quite baffling tbh

[true seal]

for lack of logical arguments private invectives. What to expect from the internet

[true seal]

nothing like that, rounding, patching up defects, releasing whole series of defective products (here see Intel's latest exploits), is this the future based on quality? NV cheating customers for years? Only pelicans like to swallow it thoughtlessly.

this is a very cultured form of getting attention compared to how you behave and what you write.

[gusty fractal]

So, question to you again.

Since the 4090 is a cutdown from the server full size...
Its a defective product by your standard?
Yes or no.

[true seal]

ok, this is a question for you, do you see a difference between: A) development work, B) production planning stage C) and production itself? You write and ask me about: B and I described a specific situation: C on the example of a pasted GPU from Amazon. Interestingly, all the current speculation and leaks regarding NV rev. 50x will most likely be averaged out at stage B, ultimately minimizing losses.

[gusty fractal]

So still no answer from you.

Got it.

Third time you avoid my question.

[true seal]

writing in such a way also makes no sense, NV went a different way than AMD. The idea for PR of NV products included aggressive marketing including black PR. Have you seen AMD be in any way concerned about it or react? No, because it resembles a small dog barking at a big dog (only the relations are reversed, the big dog howls and the small one doesn't give a damn about it). They simply did their thing, and attention I am not glorifying AMD here, I am only comparing both companies and their solutions, We have reached such a funny moment when a smaller player on a set market made a cheaper product with the same or even better performance - it depends individually on how we want to compare products. Of course, excluding the price, not touching on the issue of TFLOPS, omitting all FPS performance results in one single point and provided that the game supports raytracing NV has slightly better parameters. It's just like having blinkers on and trying hard to find out if she's better at anything XD for some it's an argument, for others it's nonsense. You know my opinion.

[true seal]

Your question does not relate to what I wrote. Cutting off what we are writing about and creating a separate topic what NV released for production, released what they mathematically considered a compromise of quality/loss and profitability, the resultant of these three created 4090, it's very cool that they made such GPU for there are people who do not have a $ limit but for most they priced themselves very badly and what more should I answer you?

[true seal]

only because I know what you're getting at. I'm not avoiding the answer, I'm just avoiding pulling up stage B and treating my description C as equivalent, these are two completely different situations. Clever, but not this time.

I am also glad that your dictionary includes difficult words like baffling and not just insults, that's a huge plus, at least we have a semblance of conversation about technology, production and sales and not just empty statements about nothing important.