Subj : Re: MSI H97 motherboard q To : Barry Martin From : Ky Moffet Date : Tue Oct 15 2024 16:18:00 BARRY MARTIN wrote: > Hi Ky! > KM> LOL. I just never got around to using it foranything in > KM> particular, tho at the time I probably had plans. As it is, it'll > KM> eventually accumulate enough temporary crap to fill it up. > KM> ("Only" 512GB.) > > That could fill up real quickly or take forever, depending on the > function. Its apparent function is to be a random temp dump. :) > > know, did a quickie Google search: > KM> Hadn't looked. > Bing? GoGoDuck? Duckduckgo :) I haven't used Google since Startpage/IXQuick and then DDG came along. > KM> Yeah, see, according to the article, it should work with 4 NVMes > KM> (and the 4-holer card I have appears to be the same as the > KM> Asus-branded one, and probably is, except for the fancy cover). > KM> But turns out there's also "bifurcation" -- the board has to be > KM> able to support splitting the PCIe bus between two or more NVMes. > KM> And apparently Zombie does not, while somewhat-older Silver does. > > Right: what semi-generic articles say and what the specific motherboard > does can be two different things. (semi-tangent:) I was reading about > eARC for a TV soundbar -- eARC was introduced with HDMI 2.1 and so > supports eARC but there's an 'asterisk statement': the manufacturer > decides what features incorporated and what are not. The problem here really is that the board only has half-vast support, and the deficiency is not documented. > KM> board, supports 192GB, more than most do today.) Used Asus P9X79 > KM> boards (initially popular with cryptominers, probably because all > KM> the x16 slots [3 to 6 depending on model] will simultaneously > KM> support a GPU) are now commonly being repurposed as budget server > KM> and workstation mainboards. > > "The manufacturer decides...". The mfgr didn't finish deciding.... BTW a lot of boards that swing both Pentium and Xeon can double RAM amount over what's documented... if you use a Xeon CPU. > KM> That's another thing Zombie doesn't support -- multiple GPU > KM> cards, which I expect is the same problem. However, it has pretty > KM> good onboard video (actually in the CPU), and allows up to 512mb > KM> shared RAM, so I haven't even bothered with a vidcard. > > Right: if the video is sufficient for the viewing needs why bother with > a video card? The card would take away a PCIe channel which could be > used by the NVMe. Yeah. Right now Zombie's video output has no job other than making me tear my hair out, because neither Win10 nor Win11 will see the rest of the network, other than very intermittently. And it's something in the hardware, cuz it's set the same as the other 10/11 boxen that work fine (for Win10/11 values of 'fine' -- that is, access tends to be only one way) and it does no better with Westworld's "portable" Win10 (that was installed on Westworld but doesn't care what PC the HDD is in) that normally works perfectly with the local network. Same with onboard NIC and MSI's own wifi card. Both see router and internet. Will sometimes see linux box, but not always. Nothing else. Can sometimes be seen by Silver, but not accessed. Couldn't see the old PCI (not e) 3COM NIC I tried for troubleshooting, and don't have a newer one instantly to hand. > The only problem I've found with the built-in video is they don't > allow more than two (maybe one) video ports. Here I'm starting to find > a third monitor would be handy. I suppose it depends on the board. Some have 3 (or 4 with multi HDMI) ports on the board and claim all work at once. Yeah, extra monitor is in my future, but first, a better desk. > KM> Well, heat-trips are real enough, but you can buy most NVMes > KM> either with or without a mfgr-attached heatsink. And... > > I'd tend to go with heat sinks but then I read when I was looking at RAM > about a year ago the material the manufacturer used to keep the heat > sink on the memory stick retained heat -- didn't pass the heat > effectively from the chips to the heat sink. The only RAM heatkinks, er, heatsinks I'm sure work come on RDRAM (Rambus) modules, and RDRAM runs so hot it needs 'em. Otherwise... they use either foam or silicon "heat pads" which are INSULATORS, how exactly is that supposed to work?? I think RAM heatsinks are mostly a selling point to the idiot overclockers, and provide a spot to waste power and add heat with fancy colored LEDs. > > KM> Had the 1TB NVMe > > KM> (on a PCIe card, no heatsink) working for a while the other day > > KM> and afterward it's barely warm enough to tell it was powered on. > KM> Couple days ago two hours of continuous use, barely warm to the > KM> touch. > > I'd say it was sufficiently cooling. Yeah, got itself right at the top of the "not to worry" list! (WD 1TB Blue, happened to be on sale half price and was deemed adequate.) The older pair in Silver get warm, but not really hot. Oh, you can now get an NVMe hotswap bay, but they're still pricey. > BTDT! I don't consider the PSU fan(s) as chassis cooling. The chassis > fans should provide a directional breeze (vs. gale-force winds) to move > warm air out. I think problems arise when people neglect to note that the average case is effectively full of baffles, just from how they're constructed and from being full of cards, and don't realize that you have to blow air INTO a confused, er, baffled area, not ACROSS it (which may prevent circulation entirely). > KM> Most of the higher-end cases of the past had an intake fan, so > KM> I'm not alone in this. (Silver, Paladin, and Bullet all do.) > > I think all of my 'new' desktops have at least one chassis fan. Recall > at least one old computer where the CPU fan had a hood on it to blow the > hot air out of the case. Those hoods are an invention of the devil. The hood arrangement was most of why that expensive Dell kept overheating. > KM> Dell's "engineered" cooling on what was in 2003 their > KM> top-of-the-line $4000 system was probably the worst I've ever > KM> seen. Tinker came to me as a two year old because it wouldn't > KM> stay running even with then-newfangled water cooling, but still > KM> using the dedicated air funnel... I threw out all Dell's crap, > KM> gave it a normal HSF and a case-intake fan; temp dropped 40F and > KM> it ran stable until the capacitors went a few years later. > KM> (Probably the most overpriced motherboard I've seen, too.) > > But it's a Dell!! I'll admit to not being a fan of Dell only because > back when I was buying used/reconditioned computers I could not find the > specs for the units. "It can come with ..." doesn't tell me what I > needed to know. Their consumer PCs are still kinda crap, but their business PCs (Optiplex and servers) have been very good for the past decade or so, and their business laptops have always been pretty durable. But yeah, in the olden daze they were terrible for proprietary parts (including the power connector from the PSU) and it was tough to replace anything but with an expensive Dell-branded part. Tinker was the first one I'd seen that could use off-the-shelf parts. Some since then still have proprietary PSUs, and good luck finding which without letting out the magic smoke. > Makes sense: one wants the heat to flow from hot (the chip) to cold (the > air). If the heat sink retains heat that doesn't help cooling the chip. > There is going to be some residual heat in the heat sink just because of > 'inefficiencies' but overall the heat sink should pass on the heat. Right. I have a couple of those passive copper server heatsinks, made for older Xeons (which ran rather hot). The foot is 1/3 inch thick and they have about 50 very thin fins. They will not take a fan, so it was all ambient air movement. > Yup: on a previous system (it's still in use around here someplace) the > AMD CPU was overheating when the room temperature got warmer in later > Spring. At the time used the heatsink assembly provided by AMD with > their CPU. At the time to me it was if the manufacturer is fine with > this assembly so am I -- I'm not overclocking, gaming, etc. A little AMD assumes their retail market is all gamers, who are going to overclock so will be replacing it anyway. So they always provide the cheapest crap that still more or less appears to be a HSF. -- I may have been the birdie. :D > birdie told me that wasn't correct. Swap for a big heat sink assembly > with a 120mm fan -- dropped the normal running temperature by about > 50øF! Yup. Doesn't even need to be high-end, just better than what AMD gives you. I've never seen one of AMD's default HSFs that was more than just-barely-adequate if the CPU is doing absolutely no work. And if you heat-trip an AMD CPU, you'll probably kill it. Intel doesn't do that. If a retail CPU ships with a HSF, it will be at least adequate. Not up to overclocking, but good enough for ordinary consumer use. > > KM> And even the most basic heatpipe outperforms the best of the slug > KM> type. > > More cooling area I'd guess. Nope. They usually don't have much of a foot and often not a big fin array either. The secret is that the heatpipes move heat really efficiently, basically operating like tiny fridge condensers (there's usually liquid inside the pipes). With newer CPUs, that's the only sort I'd use. And doesn't need to be fancy -- the $20 HP castoffs in Silver and Fireball do a stellar job, both idle just above ambient and about 45C if busy. Paid about the same for the one in Zombie, a little fancier, and that slightly-faster Xeon runs at about 45C. > KM> No, it had definitely been heat-tripped (remember it came to me > KM> because it was at best flaky and had been deemed dead), and > KM> probably that set a flag in the CPU so now it believes if it's > KM> powered on, it's that hot. It takes seconds to get there, even if > KM> it's barely warm to the touch. Likely it's meant to indicate it > KM> ought to be replaced, but it behaves fine with better cooling. > KM> But the "new" Xeon is so much faster, I don't care what the i5 > KM> thinks. :D > > It did get replaced: well, the owner did! LOL. Yup. PEBKAC in action. :D Moonbase's temporary innards (dual P3-500, yes it's ancient) started falling over after about 20 minutes up... turns out that's how long one of the CPU fans lasts before it seizes up and heat-trips that CPU. It is temporarily a single-CPU system, not that it matters when its sole job is to play DOOM. But the overheated CPU is none the worse. Oiled the fan and now it turns much better; will have to check if it plans to stay working or if I'll have to dig one up somewhere. Those P3 CPUs need the fan. Or I could just replace the innards, since I have a new board to replace the one that needs tired capacitors swapped... I ordered the $170 "used" board with some trepidation about the possibility of bad capacitors on a 20 year old board. I received the exact same board but brand new (current manufacture) that presently retails for $460 direct from the mfgr, and has solid capacitors, and came with a core2duo instead of a plain P4. Same model number, tho. Someone screwed up at the salvage yard. :D þ RNET 2.10U: ILink: Techware BBS þ Hollywood, Ca þ www.techware2k.com --- QScan/PCB v1.20a / 01-0462 * Origin: ILink: CFBBS | cfbbs.no-ip.com (454:1/1) .