|
Fun Times! posted:...others say you need to do a fresh install. People are paranoid about versions past. As far as I can tell Windows 10 has a greater chance to brick itself on an update than to complain about hardware changes, I've used my work desktop's install of 10 Enterprise on 4 different platforms (Lynnfield->Sandy Bridge->Westmere->Braswell->Westmere again) and not once did it do anything more than take a few minutes to switch drivers around.
|
#
¿
Dec 2, 2019 16:23
|
|
|
|
| # ¿ May 17, 2024 19:42 |
|
|
BangersInMyKnickers posted:1 and 2U boxes aren't terribly loud if you don't gently caress around with the power management features so the fans can ramp down. If you're throwing them in a mechanical closet or utility room it doesn't really matter. With that said, I personally believe that with virtualization and the high socket densities that are available today that if you are running more than one box at home you've hosed up I run two separate boxes for NAS and VM sandbox, but only because I like to be able to take the sandbox down for reimage/hardware swap without having NAS downtime (and the NAS hardware is too old for VT-d). Both are just normal microATX desktops and with 32-core HEDT systems being available at this point, I can't imagine anything I'd want to do at home that would have me going to rackmount.
|
|
#
¿
Dec 18, 2019 18:53
|
|
|
They might not bother anyway but given that the reason provided is ROM size limitations, boards with an uncommonly large EEPROM might still be able to do it. A lot of boards use 128Mb/16MB but the X470 Taichi uses a 256Mb chip, to give an example.
|
|
#
¿
May 7, 2020 14:53
|
|
|
EmpyreanFlux posted:I've been so used to the 212 being a normal, standard recommendation that it being mediocre is news to me. I get what you're saying, but there's no conflict in this. It's a standard recommendation because it's cheap, straightforward to install and has broad compatibility. It won't keep up if you go over around 120W but that's fine for most builds, I used them myself on an i7-920 and a 2500K. When I replaced that i7-920 with a six-core Westmere Xeon and started really pushing it hard, I had to get a big Noctua with six heatpipes because you can't push 200W through a 212.
|
|
#
¿
Oct 5, 2020 17:37
|
|
|
Yeah, it's definitely been surpassed but if the difference is so small as to be imperceptible and the 212 is on sale then it's not some great mistake to save your  . The key in my mind is just to make sure the capability of whatever cooler you buy is a healthy margin over your predicted CPU power consumption so that you can actually keep the fan on low.
Eletriarnation fucked around with this message at 17:59 on Oct 5, 2020 |
|
#
¿
Oct 5, 2020 17:56
|
|
|
shrike82 posted:i've been idly reading up on thin clients and noticed this is a pretty neat piece of kit Hell of a thin client - all it needs is a more serious GPU and they could sell it as a workstation.
|
|
#
¿
Aug 26, 2022 16:47
|
|
|
Inept posted:Idle use on a NAS is probably closer to 40-50 watts lower than an old desktop https://www.techpowerup.com/review/synology-ds920-4-bay-nas/12.html It depends on the desktop. I have a Socket 1156-based NAS that I'd estimate uses around 30W typically for everything except the hard drives, and I'd imagine a lot of other systems that have integrated graphics can do similarly well unless they're really old or based on HEDT/server platforms. If you really care to, a lot of systems can see substantial reduction through undervolting/underclocking as well. Eletriarnation fucked around with this message at 08:20 on Dec 15, 2022 |
|
#
¿
Dec 15, 2022 08:17
|
|
|
I broke it down in the NAS thread about six weeks ago, but that's measuring AC with a Kill-A-Watt and it's an 80+ Platinum supply at around 20% load so I think and really hope it's doing a lot better than 50% efficiency. Please, check my work if you feel so inclined and let me know if I missed anything. e: You actually responded to my initial statement about it, but you might have missed the follow-up. BlankSystemDaemon posted:Considering that the server's CPU is from '09, that's pretty impressive - processors back then had only very rudimentary powersaving compared to nowadays, so a new board off ebay/craigslist or their local equivalents might save you a not-inconsiderable amount of power. Eletriarnation fucked around with this message at 16:23 on Dec 15, 2022 |
|
#
¿
Dec 15, 2022 16:18
|
|
|
drat it, quote != edit. How long has it been since I did that last...?
|
|
#
¿
Dec 15, 2022 16:19
|
|
|
I said "with drive spin-down disabled" as in no, I have specifically configured the drives not to spin down and the 80W AC that I'm reading at the plug includes their consumption. Regarding power cost, I know that my power is cheap and I don't think I ever said or implied otherwise. I also don't disagree that the 120V mains makes a difference, although I think that difference is small enough that it is rarely going to alter your conclusion in this kind of decision. All I said is that Anandtech's 4770K system idled at 35W and my even older Xeon does ~30W, so if your "old desktop" is using 50-60 then either it's very old, poorly tuned for the purpose, or some kind of unusually inefficient platform like HEDT. Also, if your only reason to get a NAS over a desktop is power consumption you can likely still come out ahead on total cost by just buying a new platform instead of reusing your old gaming rig - I have an i5-10400 system that idles under 10W and it didn't cost nearly $560. There are other reasons to prefer a NAS and I'm not saying it's a bad decision - just that power consumption is not a strong reason to make that decision. Eletriarnation fucked around with this message at 23:13 on Dec 15, 2022 |
|
#
¿
Dec 15, 2022 22:10
|
|
|
I replaced my 3700X with a 5800X3D last week and I don't know if this is an expected outcome or if I just have a marginal 3700X, but the memory controller is definitely more robust. The 2x16GB set I bought back in 2018 was G.Skill 3600MHz CL19 and I always wondered if it was optimized for Intel instead or something; the best I was ever able to get out of it with the 3700X was 3466 on one particular EFI and more typically 3000-3200. After I put in the 5800X3D, I set it to XMP which would never even boot before and haven't seen a single problem. I also tested a 2x16 3200MHz ECC set I just got which kept throwing errors on particular addresses with the 3700X and the same addresses popped up on the 5800X3D, which was a pretty clear indicator that I should just return that set.
|
|
#
¿
Dec 26, 2022 22:03
|
|
|
a dingus posted:Did you upgrade your motherboard? I also had issues with xmp/3600mhz on my 3700x. Upgraded to a 5800x3d and a b550 and the memory runs with XMP perfectly. No, I'm still using the X470 Taichi. I thought I might have made a mistake by not going with X570 when my original kit wouldn't run at XMP, but then when I saw that the ECC memory was throwing errors at SPD I thought it was more likely a processor issue. I should have first attempted swapping the ECC DIMMs in the slots because the errors follow the same addresses and that is a powerful indicator in favor of the RAM being the problem, but I don't really regret the CPU upgrade since without it I'd still be unable to run the old kit at XMP. I will quite likely get another AM4 board at some point in the future though. My NAS is running on a 12 year old platform (LGA1156) and the plan always was to replace it with the 3700X/X470 once I verified that it supported ECC.
|
|
#
¿
Dec 27, 2022 01:44
|
|
|
Klyith posted:Does anyone even do 3-channel anymore? I'm not sure if it was ever a thing outside of LGA1366. Following generations of HEDT/server platforms went to 4+ and consumers stayed at 2. e: also LGA1356, the weird secondary Sandy Bridge server platform. Eletriarnation fucked around with this message at 18:30 on Jan 18, 2023 |
|
#
¿
Jan 18, 2023 18:27
|
|
|
Yeah... I mean, if you don't care about the features the TPM offers then you can just use Rufus to install Win11 on whatever you want. I can't think of any reason it would be less secure than installing 10 on the same machine (at least unless/until MS starts blocking updates on unsupported hardware). I have it running on a Surface Pro 2 with a Haswell chip inside and it works fine as far as I can tell.
Eletriarnation fucked around with this message at 02:07 on Jan 25, 2023 |
|
#
¿
Jan 25, 2023 02:03
|
|
|
Twerk from Home posted:So I'm really hoping this isn't your problem, but my piece of poo poo Vizio overheats if you feed it 4:4:4 4k input like a PC would do. It does it faster and more often when it's warm, and less during winter. People got motherboards swapped during warranty but the problem just comes back. Oh hey, I had a similar problem with a D43-F1 - once it had been on for an hour or two, the image would get this periodic double-vision sort of flicker. Tried a couple replacement motherboards from eBay but one was for the wrong model of TV, and one was stuck in store display mode. I ended up going back to the original motherboard and using a hole saw to cut a port for a 120mm exhaust fan. With a 4-12V input range fan running off the USB port, it's basically silent but moves enough air that I haven't seen any problems since.
|
|
#
¿
Mar 14, 2023 16:51
|
|
|
gradenko_2000 posted:AM4 has the pins on the CPU, with socket having holes Wibla posted:Always? For values of always >= 2005, I guess? I started to say PGAs had a much longer run than LGA has so far, but I realized I don't know when they started doing that instead of the CPU being soldered/a giant DIP chip. e: Looks like PGA was introduced in 1989 with the 486, so LGA passed it in lifespan (from an Intel desktop chip perspective) a couple years ago. Core 2 mobile used PGA even after it was discontinued for desktop so really they didn't kill it until 2009. Eletriarnation fucked around with this message at 20:07 on May 30, 2023 |
|
#
¿
May 30, 2023 20:00
|
|
|
You'd think that, but then again maybe the world is ready for hexadecimal model numbers. Couldn't be much worse, right?
|
|
#
¿
Dec 30, 2023 14:47
|
|
|
Klyith posted:So here's my hot take: if you buy poo poo with zero research or reviews, just because it has some bigger number or has some branding, you are already a sucker. A less-disingenuous CPU naming scheme won't help. You played yourself. As much as I hate bad numbering schemes, I also kind of feel like this in the end. Buying CPUs was never just about "bigger number is better" even back when the CPU was directly branded based on its clock speed and turbo wasn't yet a thing. You can ask anyone who took a Celeron 300A to 450, or who got a Willamette P4 with RDRAM. This has only become more true in an era of different core counts, heterogeneous architectures, multiple architectures sold simultaneously with the same branding, and widely variable TDPs. Fortunately there's lots of reference material out there explaining what each part number actually is these days, including first-party reference sites for both Intel and AMD. Before you buy anything you owe it to yourself to at least look the parts up and make sure that they are what you think they are, let alone looking up some benchmarks as well to make sure that they perform as you would expect.
|
|
#
¿
Dec 30, 2023 17:27
|
|
|
Cygni posted:While I agree that power users will figure out the difference between the CPUs and the days of simple naming schemes arent coming back, the fact that the naming scheme is intentionally misleading and designed to bilk money out of the vast majority of laptop buyers who don't know the difference should still be called out. If we, the turbo dorks, don't call the Big 3 out on their scammy naming schemes, who will?? Yeah, I agree, we should call this out as being suboptimal. The only reason manufacturers ever do this kind of SKU renumbering is because they want consumers to think older tech isn't older tech, and to thereby be OK with buying it. It's not a new practice though, and unfortunately it's fairly common. I don't think that selling an 8840U which is a 100MHz faster 7840U is fundamentally any different than selling a 14900K which is just an overclocked/binned 13900K, and I think it's a hair less dishonest even than letting reviews go out for a GT1030 with GDDR5 and then quietly launching a DDR4 version. I also feel like the harm caused by this kind of rebranding is pretty unclear. How many people actually see a 7520U next to a 7840U and think "the number at the front is the same, a 7520U is probably good enough" and how many of those people in turn actually have a use case for which the difference will be important? Eletriarnation fucked around with this message at 17:42 on Jan 2, 2024 |
|
#
¿
Jan 2, 2024 17:37
|
|
" when buying a laptop, a smaller minority will get to "its a 2000 series i5", a teeny-tiny amount of people will get to "its a Sandy Bridge i5", and a laughably small amount of people will get to "here, lets look at the marketing decoder chart to see what CPU this really is before buying", and the Big 3 count on that when they do their naming. It sucks, call em out!
|
I think if you are literally buying based on nothing more than "the number at the front is higher" when actual specs and benchmarks are out there, you are doing the same thing as walking into a car dealership and asking the first person you see "Hey, what's hot right now?" and buying that. You are admitting with your actions that you don't really care about the fine details of the actual object you're buying, you just want to get sold something functional with a good story of "something new" and hey you're gonna get that so what's the harm? If it's actually important to you that something have specific performance characteristics, then I absolutely think that the manufacturer must not lie about those characteristics or try to obfuscate them - but I don't think "well, if you check the decoder wheel then all this means is it's a 2024 model year instead of 2023" rises to that level. Someone who wants to play a particular game or run a professional workload can handle that, I think. It would be easier to track what's actually going on if vendors didn't try to play games like this, but as long as the practice helps their OEM partners move machines to people who don't care to know more I do not know why it would change. Eletriarnation fucked around with this message at 01:54 on Jan 3, 2024 |
|
#
¿
Jan 3, 2024 01:36
|
|
|
hobbesmaster posted:There are cars that literally no changes between some model years too. It’s clear what they’re trying to do, but if you’re going to just do model years then do model years? I kind of agree but suspect that if they actually started off trying to faithfully follow model years then you'd have an effect where uninformed consumers would not want to buy a model from the previous year, even if nothing newer had actually replaced it yet. This would lead vendors to fudge the model years as well, just like car manufacturers do: quote:The new model year typically begins in August to September of the preceding calendar year, though can be as early as February, such being the case for the fourth generation 2022 Acura MDX, which started production in January 2021. This was partly due to the advertising of a new model being coordinated with the launch of the new television season in late September, because of the heavy dependence between television to offer products from automakers to advertise, and the car companies to launch their new models at a high-profile time of year. Eletriarnation fucked around with this message at 06:25 on Jan 3, 2024 |
|
#
¿
Jan 3, 2024 06:12
|
|
|
I mean, there's always 5650. Heaven forbid we ever do anything with the ones digit other than 0 or 5 - that's a clear indicator of a trash product or something. Maybe Intel's new foray into three-digit part numbers will bear fruit, I see they have a couple 4s in the series.
Eletriarnation fucked around with this message at 17:27 on Feb 7, 2024 |
|
#
¿
Feb 7, 2024 17:22
|
|
|
hobbesmaster posted:Apparently a ryzen 5 5645 is a thing. https://www.amd.com/en/product/12186 As far as I can tell the 5645 is just the Pro version of the 5600X. Same core count, clock speed, cache, architecture, and TDP. Same thing for 5700X/5845 and 5900/5945. No idea why they couldn't just have the same model numbers with Pro at the end.
|
|
#
¿
Feb 7, 2024 23:53
|
|
|
Makes sense considering that the 5700 comes with a cooler and the 5700X does not.
|
|
#
¿
Feb 13, 2024 21:30
|
|
|
I run ECC on my NAS because: 1) It greatly reduces the chance of outages due to soft errors. 2) It greatly increases the chance that I will notice if the RAM actually starts to fail before I start getting mysterious outages. 3) The cost increase vs. non-ECC DDR4 (which is the only downside of ECC) was only about $60 for 32GB, less than 5% of the total cost of the system after I include drives. I know that there are hypothetical benefits for file data integrity, but it's a distant fourth as priorities go - I expect ZFS to play a greater role there since file data spends a lot more time on the disk than it does transitioning through system memory. I have found ZFS to be very good at handling the aftermath of sudden outages, so to me the importance of this system's reliability is more about uptime and avoiding the work of recovery than avoiding data loss. Of course, if your data is all easily replaceable then this is even more true. On the other hand, if you are using something less resilient than ZFS then you might really want to avoid sudden outages. I think it's totally valid for a home user to not give ECC serious consideration, if your preferred platform doesn't support it (e.g. Intel) or if you aren't very concerned about the possibility of soft error outages. Soft error outages don't seem to be especially common in my experience even now that we are getting into the tens of gigabytes of memory, so that's fine. However, if your platform does support it then the cost is not bad. For that reason, I think it's a bit apples vs. oranges to say "you should have an offsite backup first" considering that could easily double the cost of the overall solution. You should have an offsite backup for anything critical or irreplaceable, but that is not most of the contents of many home NAS systems. I also don't understand the claim that it's comparably important to have ECC (or periodic testing) on your desktop vs. your NAS. It would be nice, sure, but a soft error on my desktop only affects me and maybe the data on its one SSD. Most of that data is Steam games and can be replaced in a heartbeat, and that which isn't easily replaceable is backed up... mostly to the NAS. The vast majority of the data on the NAS isn't even transferred through the desktop, I download it through VMs also hosted on the NAS. Eletriarnation fucked around with this message at 16:35 on Feb 28, 2024 |
|
#
¿
Feb 28, 2024 16:04
|
|
|
Yeah, most mATX cases are like "standard big-OEM desktop" size. Noticeably smaller than a full tower, but not really "small form factor". It is a nice sweet spot if you're confident that you won't need all of the ATX expansion slots but you still want a relatively normal build experience, especially since mATX boards tend to be a little cheaper than full ATX boards with the same feature set and they're definitely cheaper than ITX. You can find some SFF-ish cases out there for them though, especially if you're willing to go low profile. I use a Silverstone ML03 for my HTPC and it's great since it looks kinda like a DVR and fits easily into a TV console, but it wouldn't be ideal for a gaming build. Eletriarnation fucked around with this message at 16:28 on Apr 12, 2024 |
|
#
¿
Apr 12, 2024 16:25
|
|
|
|
| # ¿ May 17, 2024 19:42 |
|
|
Really, are we going to have to do the thing of putting "AI" in all of the product names for a few years? It's going to be like raytracing all over again. All the marketing will be saying there's something revolutionary right around the corner which will totally need this new functional unit, and then if and when that (unnecessary) killer app appears we're all going to realize that the first gen hardware isn't actually fast enough to do it effectively.
|
|
#
¿
May 7, 2024 16:31
|
|