|
Ian is right, of course. Linus is also right that the ship has sailed on people saying it right.
|
# ? May 23, 2021 16:52 |
|
|
# ? Apr 24, 2024 16:11 |
|
Using Transfers is good too because it can ignore the bus width which would cause the bits/bytes per second to be wildly different. I occasionally get informed by friends on how pcie works due to LTT videos despite actually working on pcie devices for almost 10 years but I just smile and go “oh yeah that’s interesting” or gently reeducate them (linus thinks pcie gen3 uses 8b/10b coding for example and even then does the overhead calculation wrong) It’s just stuff to make people who don’t know the details feel like they do and that makes them feel good and smart so they keep watching.
|
# ? May 23, 2021 17:02 |
|
How does that even come up?
|
# ? May 23, 2021 17:10 |
|
pedantic doctor guy irate over a thing that happens every day, including by the people selling the memory vs tech channel that regularly fucks up the technical details truth in the middle I guess? but also like, if an official product description says something has "16GB 4266 MHz LPDDR4", then is it Linus's job to correct it?
|
# ? May 23, 2021 17:18 |
|
Reasonable centrist position: Don't watch LTT.
|
# ? May 23, 2021 18:11 |
|
Perplx posted:People use MHz and megatransfers interchangeably, they are wrong but I agree the train has left the station long ago and you can't fix it. Its like making hard drive manufacturer's use base 2 terabytes instead of base 10.
|
# ? May 23, 2021 18:25 |
|
Perplx posted:People use MHz and megatransfers interchangeably, they are wrong but I agree the train has left the station long ago and you can't fix it. Its like making hard drive manufacturer's use base 2 terabytes instead of base 10. Or gun nerds getting annoyed at people calling magazines clips. Yes, it's technically wrong, but the vast majority of normal people don't know and aren't going to care no matter what you do.
|
# ? May 23, 2021 19:54 |
|
Helter Skelter posted:Or gun nerds getting annoyed at people calling magazines clips. Yes, it's technically wrong, but the vast majority of normal people don't know and aren't going to care no matter what you do. It’s okay to call them a caisson, right?
|
# ? May 23, 2021 20:00 |
|
Both the ram MHz thing and base 10 vs base 2 storage nonsense have the shared property of being harmless to the lay consumer, for whom it remains clear that bigger number is better and that’s fine and what is important is that the scale mostly remains consistently meaningful and to the properly educated enthusiast who understands what is actually going on. In both cases it’s only actually confusing to, essentially those folks who are smart enough to go poking around or installing monitoring software but also too dumb to understand the outputs of said monitoring software or Google when it doesn’t seem to make sense. And, not to put a fine point on it, but I think it is okay and cool and good to aggressively confuse curious teenagers like that.
|
# ? May 23, 2021 20:10 |
|
The base 10 is confusing to the consumer because when you go to plug in your brand new 10tb drive, it lists as 9.09tb
|
# ? May 23, 2021 20:14 |
Charles posted:The base 10 is confusing to the consumer because when you go to plug in your brand new 10tb drive, it lists as 9.09tb Gubermnt stealing mah hard disk space for the 5G
|
|
# ? May 23, 2021 20:55 |
|
CaptainPsyko posted:it’s only actually confusing to, essentially those folks who are smart enough to go poking around or installing monitoring software but also too dumb to understand the outputs of said monitoring software or Google when it doesn’t seem to make sense. This is DDR4-3200 why is it only running at 1600MHz did I gently caress up a setting is there something wrong with my new motherboard -Me, for about 10 minutes before I remembered what DDR actually means
|
# ? May 23, 2021 21:07 |
|
Just wait when pulse amplitude modulation becomes standard on more interfaces! (pcie gen6 etc) Everything should just be in transfers/sec the clock speeds only matter at the phy level anyway.
|
# ? May 23, 2021 22:08 |
|
FuturePastNow posted:This is DDR4-3200 why is it only running at 1600MHz did I gently caress up a setting is there something wrong with my new motherboard I think this is probably a lot of what Cutress is frustrated by - he's written some articles over the years about how frustratingly opaque XMP is to people building their systems, so you're leaving a bunch of performance on the table unless you know to turn it on, and how the numbers are going to be different from what's on the box of the memory you bought, etc. I believe he's trying to act to get people better informed so they can make better choices and have well-performing PCs, not just to be a pedant.
|
# ? May 23, 2021 22:11 |
|
Arivia posted:I think this is probably a lot of what Cutress is frustrated by - he's written some articles over the years about how frustratingly opaque XMP is to people building their systems, so you're leaving a bunch of performance on the table unless you know to turn it on, and how the numbers are going to be different from what's on the box of the memory you bought, etc. I believe he's trying to act to get people better informed so they can make better choices and have well-performing PCs, not just to be a pedant. On the other hand, I am very grateful that there is now no chance at all to burn out the memory when you overclock the CPU unless you go right for the memory. "Finally think this is what the bus speed should be for this p3 and ... ah that's what it sounds like when all the capacitors explode on each dimm."
|
# ? May 23, 2021 22:36 |
|
Friday Afternoon Alex / Linus vibes https://www.youtube.com/watch?v=u6zsHqdNRQ8
|
# ? May 24, 2021 10:23 |
|
Charles posted:The base 10 is confusing to the consumer because when you go to plug in your brand new 10tb drive, it lists as 9.09tb I guess maybe, but then your average consumer isn't taking apart their PC and installing new hard drives. Enthusiasts get itchy about this kind of thing but your average consumer works on the rule of thumb that bigger number = better. That does create some confusion in the case of say, cameras, where the yardstick of more megapixels = better doesn't necessarily hold true or TVs where not all 4K panels are created equal, but the people who will actually appreciate/notice the difference are likely already clued up on it anyway. It's like when people get irrationally angry about my choosing Bose headphones. Ok, so they're not as good as an alternative, but I'm not playing lossless music over bluetooth and I'm partially deaf so it's lost on me.
|
# ? May 24, 2021 10:38 |
|
Theophany posted:I guess maybe, but then your average consumer isn't taking apart their PC and installing new hard drives. Enthusiasts get itchy about this kind of thing but your average consumer works on the rule of thumb that bigger number = better. That does create some confusion in the case of say, cameras, where the yardstick of more megapixels = better doesn't necessarily hold true or TVs where not all 4K panels are created equal, but the people who will actually appreciate/notice the difference are likely already clued up on it anyway. Those aren't comparable at all. People buy computers or external drives and then wonder why Windows doesn't show the "full" capacity that is printed on the box. We even had a goon that got all indignant about it and didn't want to believe in the 1000/1024 bytes mathematical difference a few years back. Fame Douglas fucked around with this message at 11:14 on May 24, 2021 |
# ? May 24, 2021 11:11 |
|
https://www.youtube.com/watch?v=NJVGghP514E Part 2 of Hardware Unboxed digging into B560 motherboards, and now they've found a board, the Asrock B560M-HDV, with a VRM that's so underpowered that it seems like it can't go over 100 watts of power delivery, which means that putting a 125w part like an i5-11600K will run out of spec. As in, it won't even hit the base clocks that Intel advertises the part at. The root cause for this seems to be partly Intel's fault: as HWUB explained in a previous video, Intel's "spec" is a wide range - the base clock is deliberately set to be very low so that the TDP can be advertised as very low, and if you cap the power limit and cap the boost clock and boost time, you can technically claim that you're still within that low TDP, which Intel wants so that they're not scaring off customers with a part that looks like it needs a big-rear end cooler (even if they really do) On the other hand, the top-end of the range of their spec can go up really high, because the parts need to be able to clock up all the way to 5 GHz or whatever, and then run at that level all day, to remain competitive. Intel's trying to have their cake and eat it too. The problem, as I see/understand it, is that if the motherboard manufacturer looks at this scenario, and says "okay, I'm only going to give the board a VRM good enough to support 65 watts, because these parts are all technically rated for a 65 watt TDP anyway". If the VRMs are just good enough to run it at the minimum spec, but cannot handle a "no power limits" set-up, that's too bad, but technically still "within spec", just the lower end of that spec, as it were. But then, if the VRMs are so weak that they can't even handle 125 watts, but then the CPU support page still lists the 125 watt parts as being supported (just because the thing will technically boot and run in some capacity), then you get this situation - where the board can't even run the CPU at the base clock. Intel bears some responsibility for "stretching" their spec as much as they do, but on the manufacturer's side, if you're going to make a motherboard that cannot run some (high-end) parts even at the lower range of what Intel says they should run at, then the boards shouldn't include them as being officially supported. gradenko_2000 fucked around with this message at 13:10 on May 26, 2021 |
# ? May 26, 2021 12:50 |
|
gradenko_2000 posted:Part 2 of Hardware Unboxed digging into B560 motherboards, and now they've found a board, the Asrock B560M-HDV, with a VRM that's so underpowered that it seems like it can't go over 100 watts of power delivery, which means that putting a 125w part like an i5-11600K will run out of spec. As in, it won't even hit the base clocks that Intel advertises the part at. Does intel have non-K cpus specced at 125w?
|
# ? May 26, 2021 15:15 |
|
Since high end gaming boards effectively overclock and this board underclocks I wonder how big a difference you can get with the same cpu in 2 stock motherboards.
|
# ? May 26, 2021 15:19 |
|
Testing a basement-price mobo that doesn't have VRM heatsinks using an AIO cooler is pretty close to intentionally stupid IMO. But yeah, Intel is now in the position that AMD was in back in January 2017: the combo of high power use and consumer apathy means you get really lovely mobos from the partners. Intel boards aren't selling, so the mobo makers are looking to compete on price and save money on components. Much like how AMD FX got double-hosed by being both a bad chip and getting bad mobos. IIRC a year or two ago there was a news item that said while Intel had lowered prices for boxed CPUs, they hadn't budged on the chipset chips to mobo makers. This is the knock-on effect from that. An $80 mobo has the cheapest possible VRM and almost no components. Compare to Ryzen's launch when AMD must have been practically giving B350s away. (The good reason for Intel to keep chipset prices up is that they still sell a ton of 'em to OEMs like Dell and whatnot for corporate machines. They'd have a hard time giving cut-rate Z590s to Asus & Asrock for enthusiast mobos and then turning around and charging more to dell and lenovo. Not an option AMD ever had back then.)
|
# ? May 26, 2021 15:21 |
|
Asrock obviously shouldn't advertise compatibility with CPUs its board can't run at their stock clocks. But also, gently caress Intel for making CPUs that can't hit their advertised clocks with less than 200W.
|
# ? May 26, 2021 15:53 |
|
mewse posted:Does intel have non-K cpus specced at 125w? nope. all of the non-K parts are rated for 65w, with "T"-model parts rated for 35w (and even lower clocks) FuturePastNow posted:Asrock obviously shouldn't advertise compatibility with CPUs its board can't run at their stock clocks. yeah, it's a problem from both ends - the manufacturer shouldn't list a part as being supported if the VRM is too small to even run at the bare minimum spec, but also Intel's bare minimum spec is really stretching credulity. An 11400F can compete with a 5600X... but only with power consumption that Intel doesn't want to be honest with on the marketing.
|
# ? May 26, 2021 15:56 |
|
https://www.youtube.com/watch?v=CayBoXmyf0s&t=780s "I want the companies to know that we're insane, and we'll spend every dollar to validate their claims." God I love this channel.
|
# ? May 26, 2021 18:16 |
|
That video is basically Steve drunk with power and it's awesome.
|
# ? May 26, 2021 18:33 |
|
I believe in steve
|
# ? May 26, 2021 18:49 |
|
How many fuckin mod mats did they sell.
|
# ? May 26, 2021 18:51 |
|
Romes128 posted:How many fuckin mod mats did they sell. According to him, the store entirely funds the company at this point. Pretty much any new product they put out sells through its initial run immediately.
|
# ? May 26, 2021 18:59 |
|
That's awesome
|
# ? May 26, 2021 19:00 |
|
beyond cultivating an enthusiast fanbase the modmats are well designed products, honestly. exactly tuned to their audience of people who want to start tinkering, probably very inexpensive to produce and yet including content that is unambiguously GN even before the unique branding.
|
# ? May 26, 2021 19:21 |
|
CoolCab posted:beyond cultivating an enthusiast fanbase the modmats are well designed products, honestly. exactly tuned to their audience of people who want to start tinkering, probably very inexpensive to produce and yet including content that is unambiguously GN even before the unique branding. I want one of the large so fuckin bad but shelling out 100 bones is too much right now.
|
# ? May 26, 2021 19:26 |
|
gradenko_2000 posted:nope. all of the non-K parts are rated for 65w, with "T"-model parts rated for 35w (and even lower clocks) So.. the B chip mobos were always non overclock, targetted at business use, I don't think it's worthwhile to rage about the VRM can't run flagship enthusiast K series cpus at their fullest. I did get a Z370 mobo from asrock that had a super lovely VRM with no heatsinks, that's more annoying to me than a B series mobo
|
# ? May 26, 2021 19:32 |
|
Romes128 posted:How many fuckin mod mats did they sell. I've bought a modmat, a shirt, a toolkit, and a mousemat. Endymion FRS MK1 fucked around with this message at 19:48 on May 26, 2021 |
# ? May 26, 2021 19:45 |
|
Pilfered Pallbearers posted:https://www.youtube.com/watch?v=CayBoXmyf0s&t=780s holy poo poo steve you just keep getting better and better
|
# ? May 26, 2021 20:55 |
|
Modmats were showing up on other channels for years too. He’s had a lot of free advertising. I can see now why other channels are pushing merch more and more.
|
# ? May 26, 2021 22:52 |
|
Lol the intel thing is just like when they strongarmed microsoft into “vista capable”
|
# ? May 27, 2021 03:57 |
|
mewse posted:So.. the B chip mobos were always non overclock, targetted at business use, I don't think it's worthwhile to rage about the VRM can't run flagship enthusiast K series cpus at their fullest. I did get a Z370 mobo from asrock that had a super lovely VRM with no heatsinks, that's more annoying to me than a B series mobo It's my understanding that it's the H-series motherboards that are even more barebones-basic for business use.
|
# ? May 27, 2021 04:04 |
It’s very important that Steve wants to buy a 40,000 dollar fan tester AND does serious reviews of GPUs with pretty anime girls on them.
|
|
# ? May 28, 2021 17:17 |
|
|
# ? Apr 24, 2024 16:11 |
|
I will give Steve credit for doing a serious review to let those looking at them whether they're pieces of poo poo or not. But dear god. Perfume? Why?
|
# ? May 28, 2021 19:23 |