hooraah

I'm starting to think it may be an overheating issue, but not the gpu die. Would these symptoms match overheating vrms or overheating memory? I could see vrms getting too hot causing the output to crash, but I'm not sure I see it causing the whole pc to turn off. Then again, maybe it would.

I've had an RX580 4GB card for a few years, never mined on it, bought it new. Ran it in an ASUS Z77 board with a 3770K overclocked to 4.2 for awhile. No issues. I recently was donated an HP omen PC that had an HP board and a 4770K. Tore out the internals and did a new build, overclocked to 4.1 or 4.2, and put the RX580. It has a 700-ish watt power supply out of a dell server. I don't remember exactly when I started having issues, but I first noticed that my monitor (1440P ultrawide on displayport) would seem to flicker and change settings randomly, changing color modes and occasionally have a flash of purple in the screen when watching youtube videos, but very intermittently. It could go days and be fine. Gaming seemed fine for the most part except for random crashes (usually just straight up power offs or restarts), which were not frequent (maybe once every 4-7 days). Sometimes when the PC crashed, it didn't want to restart and would power off during windows boot just after login. It might do that twice then be fine again for weeks. I finally got fed up and tried to get to the bottom of it. I did some stress testing and found the GPU temp getting a little high (high 80s). Redid the thermal paste and cleaned the heatsink, brought it down to 79-80C at load. During stress testing I was able to more frequently get the PC to fault where it would just straight up power off or restart. At time of one restart, the temps were running 81C. No blue screens, no instability beforehand, just straight up power off. Today I tried using the PC and all of the MSI afterburner temp monitoring stopped working and I had a wonky image on the far right side of the screen. I figured "This is it, its finally over". I pulled the RX580 and put it back in the Asus Z77 / 3770K build - it works fine. No flickering, can run stress tests for hours, no instability. I put an R9 380 card I had lying around into the HP omen 4770K build - no issues either. No flickering on the ultrawide, no problem with stress tests, no instability. If it were a power supply issue, I'd expect issues on the R9 380 card, and I wouldn't expect flickering issues in windows with no load. If it were just a displayport cable issue (which are common) then I'd understand the flickering, but that wouldn't explain the instability. If it were an issue with the motherboard/PCIE slot, I'd expect the same issues from the 380. At this point, I'd just like to troubleshoot one/both of these PCs and find the issue so I can either sell them or give them away. I've already moved on to a new system. I'd like to get to the bottom of this so that I don't end up passing on problems to someone else. I'd hate to just throw the RX580 in the trash if theres nothing wrong with it.

How often have you all gotten memory errors that have been fixed by re-seating? I still find it hard to imagine how ram can become 'unseated' by just sitting there. I have a few PCs that people have dropped off in my shop when they stop POSTing. On system 1, I try resetting the ram several times. I get it to POST once, its in bios for 10 seconds, then shuts down. Can't get it to POST after that. Shelf the system as I got busy with something else. System 2 comes along, same issue - won't POST. I pull the ram and try a few different pieces of RAM, none work. Then I try the RAM that came out of system 1 in system 2. POSTS, boots into windows, no issues. So I try the RAM from system 2 in system 1. Same thing, it posts and boots without issue. Strange. All the ram is similar - single DDR3L, single sided, might be a slight speed difference. So I swap the ram back (all ram back in the original machines they came in). They all POST fine now and all work normally with their original RAM. Now I don't trust either system and I'd like to run some serious stress tests to see if they are stable.

Oh I didn't realize the 1660 turbo'd that high. Thats pretty cool. That might be a good chip to drop in in the future. I think for now I'm going to give it to my son with the 1607, put his 1050TI in it and find a new side panel for it. Then if he ever gets into games that tax the 1607, I'll find a 1660 for it. I know OEM systems aren't looked upon well here but for free this is a pretty capable system I think.

So I went by the electronics recycling center in my area where people can drop off old electronics for recycling, and saw this Dell Precision T3600 sitting there. I decided to pick it up and see if it still worked or if it had any useful parts in it. It had no RAM so I threw in some DDR3 I had lying around and it posts without a problem. I'm going to throw an SSD on it and install windows on it next. Here are my questions: 1 - Its got a LSI RAID controller card in it, and it also has sata ports on the motherboard. I don't know much about PCIE RAID controllers. Should I/can I use it for a standard SSD drive? Or would I be better off throwing it up on ebay and just plugging the drive into the mobo instead? 2 - Does anyone know what the best un-supported CPU in this Socket2011 board is? I got the official supported list from Dell, which shows it can take a Xeon E5-2665 8-core 2.4ghz, or an E5-1660 six core 3.3ghz, but I was wondering if anyone had tribal knowledge that better CPUs would work. I've upgraded a lot of HP/Dells in the past with unsupported CPUs, so I thought I'd ask. It currently has an E5-1607 4C 3.0Ghz. I've had really good experiences turning old Dell Precision workstations into cheap gaming rigs for people, but I've never worked with this model before (mostly older 775/771 xeons). I'd like to use this as an upgrade for my son who is currently on an i5-2300 with a 1050ti. Photos attached, for fun.

It seems like overkill to me. I run Emby, storage for my security cameras, a few VMs, and backups on an I5-3570K. Emby also records 2 TV tuners. I have a lot of users on emby, but most are direct streaming and not transcoding. The ones that do transcode use the intel quicksync and it keeps the load off of the main cpu. Even still, the most I'll ever see at once is around 4 streams, which it can handle fine. The UPS is definitely a good thing to have when you are serving video to remote people. 95% of my UPS benefit comes from momentary power flickers that would cause the PC to reset, but don't go on for minutes. Make sure your modem and router are plugged into it as well. I use a 350W unit and it doesn't last for very long, but its very useful.

Something isn't adding up. 4 users all simultaneously saving a 10mb file should be almost no load on the server, unless you have exactly 0 bytes left, or if one of the hard drives is failing. I would start looking at some stats and see where the bottleneck on your current server is - is the throughput maxed? disk i/o maxed? sitting there idle? If you copy a 1GB file to a file share, whats the transfer rate? You definitely don't need a new server to the tune of 5K. New hard drives at most.

I just have an extra windows 10 pro box with a bunch of hard drives shared.

I would just use the system as it is and install FreeNas on it and play around with it. Unless you have dozens of users hammering it constantly, I doubt you'll feel the RAM or CPU shortage. I do believe you can use a 4GB USB to boot freenas, but its been a long time since I've messed around with it. How many people do you expect to use it? If its just you, and your data isnt larger than a single drive, you may find its easier just to store it on your own system.

You need to change the main wireless mode from router to repeater or bridge. Not sure where the option is. Maybe network tools or administration.

You could make your tablet an access point and then use the router in wireless bridge mode. The downside to this is that the tablet will have to assign IP addresses, which will be inconvenient. Alternatively, you could try to use the asus router to connect to the tablet in repeater bridge mode, but then you'll have another NAT to go through.

While there may be some truth to price increases due to import tariffs, I would argue that the average increase will be significantly less than 25% as many products are made in other SE asian countries. Furthermore, I don't know anything about the guy that made the tweet, but I'm going to guess he's a tech commenter and makes money on pc component referrals? am I close?

Normally I would say it should work, but the last Lenovo I took apart had a very odd, proprietary PSU power connector. I want to say the one I took apart was an H50, but it was awhile ago and I don't quite remember. I had trouble putting a standard PSU in the case as well, but again, my memory of it is fuzzy. The problem ended up being a bad motherboard so I ended up stripping the system and thus never flushed out any of these unresolved issues. So check the PSU cable before you put a bunch of time and effort into swapping the motherboard over. Edit - just looked at your ebay links - yes, the H50-50 board uses a goofball PSU connector. You'll need to take the PSU from the dell. Make sure it fits in the lenovo case.

I had a 7850 die on me last year. My mother has a machine with a 6600GT in it still running. Results are random.

Its not a bad idea, but I wouldn't recommend the thinkcenter small form factor as a base. Newegg sells lots of refurb OEM systems, so they are always showing up in all different configs. That said, I'd look for a full tower and put in a 1050ti. That will be the best performance / $ in the $350-$450 price point. Anything higher than a 1050ti will require an upgraded PSU, and now you're either looking at a gaming PC or building your own at that point.