Jump to content

Jzuken

Member
  • Posts

    25
  • Joined

  • Last visited

Awards

This user doesn't have any awards

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

Jzuken's Achievements

  1. I am using Gigabyte B550i Aorus Pro AX motherboard and DDR4-3000 RAM. I just changed my 3600X for 4350g (don't ask me why), but now I'm having trouble with terrible performance of 4350g. It's supposed to be powerful enough for things like browsing web, playing simple games and whatnot, yet I find that it's constantly running at 50% load, I am getting constant stuttering on screen and in audio. It takes two and a half minutes to boot. That's compared to my 5-year old laptop, which has 2c/4t i7-7500u in it, boots almost instantly even with fastboot disabled and has no problem playing audio or running a few demanding apps. What can I do about it without completely reinstalling the system?
  2. If we are talking about SMART, they could do some simple checks, like seeing if GPU is connected to any monitor, or if the load varies with time as it would in game, whereas in mining the load would be constant. Of course that could be bypassed easily, but most miners wouldn't bother, and then they shouldn't be able to bypass load logging. Because they can do something like service-hours, or total load per lifetime. At least that would disrupt the used market for miners.
  3. Well technically running BOINC is no different than mining, but you can easily prove you've been running BOINC by showing your BOINC profile to the buyer. That would only "hurt anyone that wants to contribute to research" if they want to resell their GPU afterwards for a hefty price. But it's the same as if someone was running their 24/7 as a taxi or as sort of community project - it doesn't matter, a car with 300,000 miles would cost much less than a car with 50,000 miles, no matter if it was used as a taxi or contributed to community like delivering groceries to elderly people. Same goes for hard drives, doesn't matter if they ran them for 2 years to mine chia, store business data, serve ads or help scientists, a used hard drive is a used hard drive. Same goes for GPUs, if someone used them for occasional gaming - we should see it, if they ran GPU 24/7 for either reason - better skip it or get a much larger discount.
  4. That's pointless, as I already said miners will either bypass it with a $1.20 HDMI plug, find another leaked driver or tweak an existing one or find another buttcoin to mine that won't be affected by limiter. But at least it will hopefully ward off small-time miners who think they will jump on the bandwagon, mine some buttcoins for a few months and sell their GPUs for a profit to desperate gamers. It will also allow Nvidia to easily strip any miner of warranty if they found someone was mining, and it would make gamers more careful about which GPUs to buy second-hand, since mined GPUs won't have a warranty on them. Doing that would cut out some chunk of miners that think they can get a warranty claim if their 24/7 GPU dies, and that they can sell their GPUs back to gamers for whatever price they paid. It will also encourage more people to buy a fresh new GPU instead of digging for used mining scraps. As for the fear of mining GPUs, as an electronics technician I can most certainly assure you that most Ampere GPUs with memory running at 110C in mining won't live to see another generation launch in 2023 without major repairs.
  5. Honestly, it seems like the experience with 3060 either didn't teach Nvidia anything, or they are outright malicious in not doing the right thing. They are saying they are doing that to limit GPU mining, but I bet it will be bypassed within a month like previous one, and most definitely it won't block mining altcoins like kawpow and conflux. A really pointless "feature" to even spend resources on. I know that Nvidia wants to limit the hashrates for one reason, and that reason is that ultimately those mining GPUs will end up in hands of gamers for much cheaper price. But here is a better suggestion - instead of beating around the clock and tinkering with hashrate limiter (and not even limiting it completely) - just make a thing called S.M.A.R.T., hard drives had that thing for 30 years, why can't a GPU company make something like this? Show when the GPU was made, average temps it was running, for how long it has been running nothing but desktop, and how long it was running under load (games/mining/AI), show maximum overclocks, show any failures, maybe even show if it has been running mining operations (I guess it would be possible to detect it quite easily). You know, that will give gamers a huge advantage when someone can just look at SMART and refuse to buy a GPU from a rig that has been running at 100 C and maximum overclock for years in a dusty rig and is on a brink of failure, same as no person would be willing to pay much for an HDD from a server.
  6. No MSRP of CPUs are set by manufacturers, there are only 2 of them, that's what they want the GPUs to be sold for to the consumers. $999 for 5950x is actually quite good, that's just a 20% price increase. There are AIBs that increased GPU prices for something like 3060 from $330 to $500, a 34% MSRP increase, and then there are scalpers that add insult to injury and scalp those GPUs to $900 for $330 GPU. $799 is the normal price, where tariffs and seller margins are accounted for, $999 is the shortage price where seller has to pay more to the distributor to get an item, anything more (like vendors setting 3080 price to $1199 "MSRP") is greedy price. The shortage in the chip manufacturing industry is about 20%, not 150% like those vendors make it up.
  7. Largest GPU mining farm in Canada tour Do a GPU mining farm tour and interview with the owner. There should be plenty of farms that can be found in Canada. There is a lot of words thrown around about GPU shortage that everyone kicks to the other party - AIBs blame Nvidia for not providing enough chips, Nvidia blames gamers for excessive demand, gamers blame miners and scalpers. But the thing is, we had the same GPU shortage in 2017-2018, and that was without covid and "excessive gamer demand". I had to buy my 1060 for $450 back then, and that was the cheapest Asus turbine GPU. It would be interesting to know things like how many GPUs do the farm owners buy a month? How much do they pay for them? How many GPUs do they use? Which ones are more popular? And then the important questions: do they plan to sell obsolete GPUs? How long will they keep buying new GPUs? When do miners think the shortage will end, aka when there won't be any mining demand for GPUs? I feel like Linus is the guy that can pull this kind of thing off and he can get in some of the biggest ETH farms in Canada, and can also say something like "so how do those mining farms manage to get so many GPUs", maybe have a small investigation of sorts, and see what vendors seems to be the most responsible for selling their stock to miners out of back doors? https://www.youtube.com/watch?v=f0HC1Udk6-E https://www.youtube.com/watch?v=ajGljNRRAW8 https://www.youtube.com/watch?v=4ekOcDG2D8E
  8. Xiaomi Wiha is a good one for it's cost if you are a working on PCs as a hobby, not a job. It will probably last the same as iFixit in normal use, but costs less. https://aliexpress.com/item/4001045300206.html
  9. Jzuken

    confused

    Bottleneck depends on the way that you are using your hardware. If you are mining Etherium you can plug 8 RTX 3090 into a Pentium G4560 and the GPUs would still be a "bottleneck". If you are compiling Chrome with 64-core Threadripper and have a GT710 in your system, that Threadripper would be a "bottleneck". If you are gaming, I'd say a good formula to keep away from bottlenecking is to have a GPU that is 1.5 times more expensive than your CPU (considering MSRP prices). For example Ryzen 3900x with RTX 3070, or Ryzen 5900x with RTX 3080, or Intel i5-10500 with RTX 3060. But of course if game is more CPU-bound, you need better CPU, if you are playing VR you need better GPU.
  10. Jzuken

    confused

    Get a ryzen 2200g and use that unless you really need a powerful CPU, but not a powerful GPU (What for? Compiling stuff?). That way when GPU prices go down you can upgrade both your GPU and get a 5950x or even something like 5950xt for decent price.
  11. I want to upgrade my CPU cooler that I use in a low-profile Fractal Node 202 case. Currently I have a puny Noctua NH-L9i sitting on top of my Ryzen-1700x CPU, which of course is not enough for it's TDP. I've devised that with some makeshift adjustments I can fit about 63~65 mm cooler in there, and I'm looking at two options: One is Scythe Big Shuriken 2 Rev.B TDP: 130 Watt Height: 58 mm Heatsink height: ~46mm Another is ID-COOLING IS-60 TDP: 130 Watt Height: 55 mm Heatsink height: 40 mm Unfortunately, there is no direct comparison between them, so it is very hard to estimate how different their capabilities are. Even looking at their design, on one hand Scythe has more heatpipes, but on the other hand IS-60 has much more larger fins. And trying to look at indirect comparison between those two coolers and other coolers yields that they should be within 3 degrees difference, and their rated TDP is the same. One advantage of IS-60 that I see is a lower heatsink height, which means I should be able to fit a full-profile 25 mm fan on top of it, and can even use a 140 mm fan. On the other hand Scythe used to be the recommended cooler for low profile builds. So, could people who used any of those coolers tell about their experience with them? Or maybe there was a direct comparison somewhere that I missed?
  12. Noctua has an extensive list of CPU compatibility: https://noctua.at/en/nh-u14s/cpucomp Going by that it should be absolutely fine to use Noctua NH-U14S even with some overclocking.
  13. I'm doing a small form factor build and I want to use AMD Wraith Max to cool the CPU, but unfortunately I'm not sure if it will fit into the case. Can someone measure the height of the heatsink on the Wraith Max for me? Or measure it from the motherboard and say how much height it is but say that it was measured from motherboard.
  14. Maybe not the right forum to ask this, but I will ask it here My problem is that I live in rural area where power outages happen sometimes, especially during severe weather. I have a powerful diesel generator set up in my house that can power almost all of the electronics in my house, but when the mains power fails it takes about half a minute to a minute to kick in, so I lose my PC, my router and all unsaved work, and it is also not good for the hardware. But classic UPSes are expensive and I don't need their capacity to keep the PC running for hours. What I need is a budget "UPS", that can keep my PC and routers running only for a minute while the diesel spins up, but can deliver enough power, as my PC and routers drain about 500 Watt at normal operation. Or maybe what I'm asking for isn't even called UPS, so do you have any suggestions for me?
  15. Well even at light loads CPU's waste a lot of power. So you can expect a 65w TDP processor like Ryzen 1600 to actually dissipate 30-35 watts of heat even at idle with default settings. So if you actually want it to dissipate less power at idle load your best solution would be not to undervolt it, but to underclock or even disable cores in software when your CPU is idling.
×