AutoGen_Quest

Ahhhh right, CPU % would be reported low because of unused cores. But the game would be thrashing the half that's actually being called upon. Thanks for that.

FIIIIIIIIIIIIIIIIIIIIIIIIIIIIXED *89k me in the ]0@t >$% it was my stats overlay the whole time the_call_is_coming_from_inside_the_house.jpeg I wanted to do a sanity check and make sure other games are fine. I fired up one that I know my system runs buttery-smooth. But it's an Early Access game, and RivaTuner Statistics Server (the in-game stats overlay) often makes it crash. So I killed that, and just went with Radeon Software's performance graphs instead. Here's what I get with uncapped framerate. So far so good. Triple-digit FPS and a completely maxed GPU. Then I went back to Cyberpunk for comparison, and in order to keep it fair I didn't switch tools. When the game loaded in, I was dumbfounded. Easily twice the FPS, and neither the GPU nor CPU usage went up appreciably. But everything is so beautifully smooth now. So there you go, it was the observer effect the whole time. The tools themselves were changing the results.

Hyperthreading is only for CPUs, right? Forgive me for the dumb question. If poor hyperthreading were the root cause, then I'd expect my CPU to be maxed out while the GPU is underused. But that's not the case. They're both well below 50%. I've also looked into two community fixes, one for the RAM+VRAM allocation, and the other for enabling AMD SMT. Neither one of them fixed my problem. I'm leaning towards this game being shoved out the door too early just to make the holiday release window. This NPC is the second-closest to me at the start of the opening sequence, and they can't even get her leg textures to load reliably. Imagine what other kinds of bugs they must have been fixing that they couldn't get around to sorting out that one.

CPU usage is also very low. 28% in the example pic I posted. I'm not very knowledgeable with memory settings so here's the memory page from CPU-Z. I think it's saying everything's fine but frankly I'd be overjoyed to be wrong. Neither the GPU nor CPU ever work hard enough for my fans to kick up noticeably. I've set custom fan curves so they both run at a moderate speed until 70°C, at which point they stairstep up to 80% and then 100%. So if they start to draw a lot of power, the fans will quickly give me an audio cue.

TL;DR: What programs should I look at to double-check any settings on my video card which might be throttling its performance? GPU: Sapphire Pulse RX 5700XT (with latest drivers, 20.12.1) CPU: Ryzen 7 3700X RAM: 16 GB DDR4-3200 OS: Windows 10 Pro, build 18363 Cyberpunk 2077 is running extremely poorly for me, but when I put in a temp and usage overlay with Afterburner + RTSS, it's showing appalling underuse of the hardware. Seriously man, 23% GPU usage at 17 FPS?!? There are occasional bursts where it'll get up to near-max GPU usage, and the framerate will get up into a more sensible 60+ FPS range. But these are short, transient, and impossible to reproduce. I've tried applying the CP2077 community Performance Overhaul patch (which resolves a low default RAM and VRAM allocation, among other things). But that hasn't changed the performance in-game. So I'm trying to figure out if there's some power-draw limit, thermal limit, or other performance setting somewhere that's causing this behavior. Surely it can't be normal. I must have goofed around in the UEFI or Afterburner ages ago and not hit that limit until this game. But for now I can't find anything that explains why the CPU and GPU are so underused when I'm gaming. I can share performance logs from HWInfo or configs from various programs, but I really just don't know where to start. Any suggestions are much appreciated.

No integrated graphics in this rig unfortunately. I took it over to a friend's house and confirmed the GPU is the failed component. My rig boots up fine when I swapped in his GPU. When I put my GPU in his PC, it also doesn't boot and shows a "failed VGA" diagnostic light. For reference, the LEDs on the video card PCB are there to indicate that it's not getting sufficient power on that line. They lit up when I tested it with the plugs out. I still can't find any official documentation of that, but it's mentioned in a review here of a different Gigabyte card. It's an Nvidia GPU but the PCB layout, cooling system, and overall design are very similar to the RX 5700 XT I have. So overall, I've got no idea what killed this card. The event log shows nothing unusual until it stopped at 5:49 AM. Temp and error-state log from HWInfo was normal. I bought the card six months ago from the EBay store of a local brick-and-mortar retailer. Hopefully they honor the warranty and replace this via RMA.

Ahhh interesting, so maybe this is a PSU issue instead. I don't have a spare PSU, so short of putting this card into someone else's rig, how can I test that out?

TL;DR What can I do to try and diagnose why a video card failed? Is there anything equivalent to the diagnostic LEDs on a motherboard? more info - my GPU stopped working overnight. Previously the system was pretty stable after I built it 6 months ago. No recent changes to firmware or drivers. Possibly there was a Windows Update that applied itself with a reboot last night; I saw the icon in my system tray yesterday but it would disappear if I tried to hover over it for more info. I also noticed the GPU fans kicking up to a higher level than usual within the past week; I opened up the case last weekend and blew out all the dust, so that struck me as odd. Normally the GPU fans run at 20% or 40% (I have a custom fan curve in MSI Afterburner and set it to a staircase behavior) and I can't hear them at all. But now when I'm gaming, even on less demanding games like Terraria and Civ6, I would hear the GPU fans kick up to 60% and stay there for a minute or two. I didn't notice it moving up to any of the even higher fan speeds on the stair step, so I doubt it was getting into more stressful temp ranges like 80°+. The video card is a Gigabyte RX 5700 XT Gaming OC 8 GB. The motherboard is an MSI B450M Mortar Titanium mATX. When I start the PC, I get no display output at all; I have tried 2 different monitors on DisplayPort and a third on HDMI. The motherboard has diagnostic LEDs and the #3 LED stays lit at boot, indicating a "VGA not detected" failure. For troubleshooting, I have tried removing and re-seating the video card, as well as unplugging both of its power cables and re-seating them. That was done at the video card end and at the PSU end. The behavior is the same as before: the GPU fans spin up, and the LED behind the logo pulses like normal, but the motherboard keeps lighting that "VGA failure" LED. I noticed two LEDs on the video card's PCB, right above the power connectors, labeled LED1 and LED2. neither one is lit. I can't find any info on Gigabyte's support pages about what they mean; the only manual I found is a fairly bland quick-start guide. if anyone knows what those LEDs are for, or if you can suggest other ways for me to test and troubleshoot this issue, I'd be grateful.

I've reached out to them but their support is a bit slow to respond lately. Understandable, but I'm itchy to have a working mouse, so if I need to go ahead and order a new one I'd rather do that sooner than later. Anyway thanks for your info, and Semper Fi.

Also, side discussion - what other manufacturer makes reliable mice? I don't want to put up with constant component failures, I've had it with Logitech. I'm looking for something that has several thumb buttons, a fairly large grip, an ergonomic tilt (though it doesn't have to be vertical), and preferably Bluetooth insetad of using a USB radio dongle. I don't do any competitive FPS or RTS gaming so I'm not too fussed about sensor type or refresh rates.

TL;DR when you get a replacement device under warranty, and that replacement fails, which service length applies? The original warranty, or the warranty on the replacement unit? Back in 2017 I bought a G703 Lightspeed Wireless mouse, which has a 2-year warranty. And good thing it did, because there was a major problem with a bad batch of sensors, and many people had the scroll wheel fail. Logitech eventually admitted the manufacturing defect and offered a warranty replacement. At the time they did that, the original was out of stock so they supplied me with a different model, which is currently sold with a 1-year warranty. Predictably, that replacement is also having scroll wheel problems. Ugh. So now the original 2-year warranty is expired, but it's been less than 1 year from when the replacement was sent. Is Logitech within their rights to tell me I can't get the new one fixed or replaced? Or should they have to honor its 1-year warranty regardless of how long ago I made my original purchase?

I think if I did that it would have to be an additional fan. The bottom intake is the only realistic option for the inwin 301c. A case that blows air back from the front position would have to draw air mass thru the side panel, which is just so congested with cabling. I did turn the rear intake to exhaust. CPU and GPU temps didn't change much, but the case ambient temps are better. My leg no longer feels warm from radiant heat off the glass 10cm away, which is its own special form of progress.

The case definitely has airflow issues; the cable management is in the exit path for that left side-panel grid, and even after being pretty careful to bundle and tuck the cables when I built it, there's clearly not a good open channel there. Right now I have two 120mm fans blowing air up from the bottom; they take air in past a filter and supply it directly to GPU. There's also one 120mm fan drawing air in from the back to supply the CPU cooler with fresh air, because my GPU is fairly long (3-fan unit) and I don't think the CPU cooler gets much from the bottom fans. But at the very least I might look at turning that rear-panel fan to be exhaust, since the normal exhaust path isn't working very well. EDIT: logged some more data from a 2nd gaming session (same game). Now that the case fans are confirmed to be pushing at 100%, even though room ambient temp is up two degrees to 29, the max CPU temp (Tctl/Tdie) recorded is only 83.9 (down 6.7). A pretty good improvement for just turning up the fans. At that timestamp, here are how the other values compare to previous session: GPU: 76 (-5) SSD1: 56 (-7) SSD2: 52 (-7) HDD airflow (interpreted as case ambient): 44 (-4)

Room ambient temp: 27 In a 3-hr gaming session, I measured some case temps with a thermal probe. Highest recorded values: Tempered Glass at upper right (near drive cage, not GPU): 43 Case top surface: 44 Case exhaust airflow: 56 my peak CPU temp was 90.6 (Tctl/Tdie). At that timestamp, GPU temp was 81 and my SSDs were 63 and 59. (max GPU and SSD values in the whole session: 84, 66, and 59.) I don't think I have a good, independent measurement for case ambient temp, but the HDD reported 'drive airflow' of 48. One problem I've noticed is that the system fans don't seem to be running at 100% even though I set them for it through MSI Command Center. Further digging indicates that the Command Center program only controls fan speeds while it's open. So there must be another fan curve in BIOS that needs to be adjusted. For now I'm going to collect more data with the fans definitely at max.

Mate my leg feels hot just being 10cm away from the tempered glass. The thing is def not turning over the case's air quickly enough. If I play an hour of games the entire case is hot to the touch: the tempered glass side, the front panel, the top panel, the back panel... it's cooking. Cheers, I'll go check it out.