Falkentyne

Gigabyte Ampere card, black screen... Were you playing New World MMORPG?

When 1660 Supers were $299, my friend was trying to talk himself out of buying one, even though they were supposed to be $249. Now with just 1660's at $500? Hell no.

Mastaflex. Reinstall windows. Sometimes you have to just deal with it and not say "oh i can't afford to do it". Making excuses doesn't solve problems. sometimes you have to just say "Screw it, I'm screwed. Gotta deal with it." Your life isn't going to end by reinstalling windows. You do have a roof over your head and a car or other vehicle, right? Then you have the basics you need covered. Priorities, man. I had to reinstall windows after a bizarre somewhat failed upgrade from 10 to 11 caused repeated spam nonstop "Firewall" errors in event viewer in windows 11, which I spent far too much time trying to fix. Let me tell you a little thing. I spent *MORE* time trying to fix the problem than I did simply reinstalling windows with a CLEAN install and reinstalling all my apps! First: Clean install of windows does not delete your files as long as you choose "keep personal files, apps will need to be reinstalled". It deletes what's installed but not the data. It backs up the folders for you. Windows 11 is even better about this. It is actually smart enough to keep the same username and in many cases, many of the files in users/documents folder still remain there. Program Files and program files x86 is backed up to windows.old. So all your data is still present. And you can always try a "repair install" first, by choosing "keep personal file and apps". If that doesn't fix the problem, then do a clean install. You'll lose a LOT less time and money than you will spending countless hours trying to fix a problem when you can get windows up and running in 15 minutes and spend the rest of your time fixing your data.

I heard about someone else complaining about Citrix...

Sorry, I meant some video cards have memory on the backplate side of the card (3090 Founder's Edition, RTX A6000, etc)....

WD SN750 and SN850 are good drives. I have a 2 TB SN850 and it has all my steam games on it. Samsung drives usually seem to be good but I did have a 256GB 960 evo, the boot drive in my laptop, just completely die out of nowhere after four years. No reason or explanation for that...it was just suddenly as dead as a doorknob (my desktop could sometimes detect it in the BIOS after a cold boot but then it would disappear and nothing could be read from it, and windows just showed a "type 10" error on the drive in device manager. My 1TB 960 Evo, 1TB 960 pro, and 1TB 860 Evo m.2 SATA (not nvme) and 1TB 860 evo SATA 2.5" SSD are still working. I also have a TimeTec 2.5" SSD drive I bought for Alder Lake testing that I may throw into my case with the rest of the system but I don't know how reliable it is...

Sounds like either the video card, a power cable or one of the memory slots isn't fully and completely secured. Check the mounting and secureness of everything, with the PSU power cable completely unplugged first. I've seen some people fail to properly snap in the memory, thinking they would break the slot, and then the system didn't work or crashed all the time. System memory and video card connectors (and other PCIE cards) tend to be the biggest culprit here, but sometimes SATA / drive power cables can come loose as well so check all your connections. A system should never freeze just by barely moving the case around.

I would go with Gelid Extreme if you are trying to deal with the front side of the card (the side with the GPU core). You can also try Zezzio (14.8 w/mk not 12.8 w/mk) thermal pads, which are definitely not 14.8 w/mk (they perform more like Gelid Extremes 12 w/mk, but are compressible as well). For the backplate (assuming your card has VRAM on the backplate side), Gelid Ultimates work very well.

No. The "issue" with very high w/mk thermal pads is their compressibility. The higher pads do not compress very well, making them very unreliable for surfaces that have multiple contact areas, where a pad of the incorrect shore rating or thickness can cause another section to not have proper contact integrity. This causes a problem on any surface where a heatsink piece has to cool multiple areas of different height, because you need a thermal pad thick enough to properly contact the area to be cooled by that section, while also not affecting contact of a completely different section of the heatsink! The most common scenario for this are GPU heatsinks, where a one piece block with different sections has to cool the GPU core (coldplate) and VRAM and VRM's. Then you need precise measurements and therefore, OEM's and AIB's prefer to use a very squishy pad, even if the performance is mediocre, since the GPU core will then make good contact with the cold plate and thermal paste. The absolute best high end thermal pad are fujipoly 17 w/mk pads but these are EXTREMELY expensive and unless you're rich, you should probably avoid buying them. They are also VERY hard and have almost no leeway for compression, and will just crumble apart after first use, if you disassemble the card. Best to use these on a surface where the heatsink is ONLY cooling that section. Gelid Ultimate 120mm * 120mm pads, or the "2-pack" versions of the 90mm * 50mm pads, are generally a better bargain for high w/mk pads. They are also a little bit softer than Thermalright Odyssey pads. Gelid Extreme pads are the softest of the higher w/mk pads, making them easy to conform to different surfaces if the thickness needed is a bit thinner than the pad you bought, if contact pressure is high enough (e.g. you actually have enough screws to compress them well). Unfortunately these can become gummy and melt a bit under high heat, so if you need a pad to tame very high VRM (mosfets) or VRAM temperatures somewhere (e.g. VRAM on a RTX 3090 backplate side), Gelid Ultimates will be better. But if you need a pad that can cool VRAM on the same side that the GPU core is on, without being so stiff as to not compress, Gelid Extremes are the best for that. Zezzio 14.8 w/mk pads seem to be very recent entry, and they appear to have about the same softness as Gelid Extreme pads. One user tested them and they seem to perform more like Geild Extremes do, despite Zezzio claiming a 14.8 w/mk thermal transfer. But as always, the THINNER a thermal pad is, the better its heat transfer will be. So Zezzio 14.8 w/mk pads may be an option if Gelid Extremes are too expensive. I do NOT know if the Zezzio pads are more high heat tolerant (e.g. won't turn all gummy) than Gelid Extremes however. I picked up one of the 2mm 120mm*120mm ones and used a section for the 3090 FE backplate "PCB hotspot" cooling, but for the backplate side VRAM itself, I put on Gelid Ultimate 2mm as I already know how well they perform. https://forums.overclockers.co.uk/posts/35151821/ I would avoid their older 12.8 w/mk pads. Seem to be much harder. Weaker pads tend to be softer and easier to work with and have more leeway on where they can be used.

System specs and OS, please? This seems strange. You can normally force maximum safe boost clocks by using MSI Afterburner, pressing Control F and then clicking (do not move the point, if it changes to "curve", reset the graph) on the 1.10v point on the graph and then pressing "L" then apply (again make sure the clock offset does NOT change to "Curve"). This won't use 1.10v, it will still use the maximum default voltage for the clock (usually 1.056v-1.081v, whatever the card wants to use) but will force full 3D clocks. Even then you shouldn't be at 1 FPS even if the card were running at 210 mhz (2d clocks).

Actually cards sag because of the design of the PCIE metal bracket of the I/O shield that goes into the bottom of the case itself behind the slot. Some video somewhere (not sure if it was Jayztwocents or someone else) showed this visually. The 3090 FE has absolutely no sag whatsoever.

No scratches, nothing sheared off= no damage.

Once very nice thing about the NH-D15 is its mounting system can actually be used with some AIO's that have atrocious backplates (like Alphacool...yikes), at least on the Intel side (that even includes socket 1700 if it fits on 1151). Not sure about AMD however as AMD sockets are completely different. Anyway whatever, you can't go wrong with NH-D15 or NH-D15S if you need PCIE slot clearance.

Henry Weinhard changed the formula recently as the manufacturing was taken over by a new company (Molson?). It's no longer the same stuff anymore. Now it's just bad.

People trying to run ADL with DDR4 are going to run into the same drawbacks as Rocket Lake (10nm backport to 14nm) had with regards to CML cross compatibility and RKL's structure actually designed for dual memory controllers. People ripped into RKL hard when it came out, not understanding that sacrifices had to be made for how it was designed, and RKL is still the MOST stable platform I've ever used. ZERO random WHEA's, ZERO parity errors, ZERO L0 Cache errors. ZERO. If you're unstable you just hard lock or BSOD (or an application crashes with an internal error). Now those same people want to run ADL on DDR4, thinking "Gear 1" will save them with latency. Yikes... Complete utter Yikes...

Sounds like an SSD problem or a USB Devices problem. Such freezes can happen when a polling error occurs, like if a device is suddenly not accessible and the system waits for it to become accessible. Check your drive health and run some scans. Also have you checked windows event viewer for the date/time of the freezes? Many times you will see an event logged there. (go to system, in the windows event viewer). (or you can run it by start menu-->run-->eventvwr.msc )

Please listen to what other people are saying here and WAIT until you're old enough. And DO NOT trade in someone else's name!!! Learn about trading on your own. Wait until you are in University or independent. Didn't you hear about that one young guy (who was OVER age 18), some college guy, who ended his own life because he "thought" he owed hundreds of thousands of dollars on Robinhood (which turned out to actually be false...it was a "temporary" financial hole)? It was all over the news. Seriously man, be careful about what you're getting into !!!!!

TFX is even better than SYY-157 but the prices are highway robbery. However if you need a stable, best performance paste at any cost it's worth it. I've tested it vs Kryonaut Extreme and unlike Luumi, I get better temps with TFX. Note that TFX, TF8 and SYY-157 are all "carbon nano" type pastes, while pastes like MX-5, Kryonaut, Kryonaut Extreme, GD900, GD007 are all silicone based (with varying compositions). That doesn't mean that all carbon nano pastes are going to be great either. The new 'Coolermaster Cryofuze" is carbon nano, but is WORSE than SYY-157 in every way. http://forum.notebookreview.com/threads/which-thermal-paste-to-buy-and-apply-traditional-and-liquid-metal.806840/page-81#post-11125037

I noticed a missing fan connector silkscreen mark on the PCB. That seems to be the same template of PCB as used on the 3080 Founder's edition card (you can tell by the positions of the shunt resistors on the front side of the card). That missing fan connector is obviously designed for the 3080 FE production. I do find it odd that they are using the 3080FE template rather than the 3090 FE template even though this has RAM on both sides like the 3090 FE does (the 3090 FE PCB is larger and you can see that the 8 pin shunt resistors are in different locations).

Clear CMOS with the power PSU cables unplugged (or PSU switched off in the back) with whatever the method is for clearing CMOS on your computer. If there is a button or jumper, short the jumper pins for 30 seconds. Try connecting through HDMI after doing clear cmos and see if it boots.

I was not aware of this. I do not know the differences in VRM layout between the 3080 Strix and 3090 Strix. I would have to ask someone else and I'm a bit too busy working and studying... I haven't seen any posts about the 3090 Strix dying unless someone can point me to any of them.

Looks like either a burn mark from a hotspot, or liquid metal oxidized residue that wasn't fully cleaned up properly with a 2000 grit sandpaper wipe + isopropyl alcohol.

The only cards that have been dying in new world are those cards using crappy AOZ power stages with imbalanced phases, like eVGA FTW3 and some Gigabyte models.

Hotspot temp throttling isn't the same as the throttling temp shown in MSI Afterburner. The throttling temp shown is for the core average temp. Hotspot follows some undisclosed undocumented rule. I've been at 95C on the hotspot with 90C MSI afterburner throttle point without throttling, because core was 82C. Also note that memory temp has its own throttle point also, again not available in MSI Afterburner.

poop case dimensions so the motherboard ATX 24pin cable fits right? Generally you need just the measurements of the PSU housing itself versus the PSU length. Then add about 3 inches to accommodate the ATX, PCIE cables etc so there's room to actually plug them in.