Search the Community
Showing results for tags 'nvidia'.
-
Hello, I have an issue with my GPU where it stays at a constant 50%-70% usege mid rounds and between rounds as well as the end screen of a CSGO match it goes to 100% useage constantly. When I play baldurs gate 3 the same happens but I expect it more in that since its a much newer game etc. etc. My GPU also goes to max useage when doing heavy stuff in unreal engine but that is also to be expected. PC specs: i9-13900F ROG Strix B760-F 2x32GB DDR5 4800MHz I am playing on a 1440p monitor with Gsync enabled, I dont know if this has any effect. Here is a screenshot of the TechPowerUp GPU sensor data whilst running a match of CSGO. as you can see its very clear when csgo was opened, when the match ended and when csgo was closed again. I did this experiment after rebooting my pc with pretty much only csgo running and im using the newest graphics drivers from Nvidia My csgo settings: res: 1920x1440 stretched fullscreen multicore rendering enabled FPS limited to 400 Is this behaviour to be expected from the RTX 3080 or is there something wrong, and if there is something wrong does anyone have any suggestions on what I can to to fix it.
-
I just bought a Gigabyte RTX 3070 and the GPU part of it is working perfectly. But the RGB area is a bit sketchy. I installed Fusion 2, every RGB option works fine except Rainbow Wave, Color Shift, Gradient, and Wave. That's when the physical light on the GPU starts lagging/stuttering. What do I do about this? I tried reinstalling the software and installing Fusion 1, but that didn't work. I also tried searching Reddit and other websites but didn't find anything that worked. My specs: CPU - AMD Ryzen 7 5800X GPU - Gigabyte RTX 3070 Gaming OC 8G (Rev 2.0) SSD - Samsung 860 Evo 1TB RAM - Corsair Vengeance RGB PRO DDR4 3200 C16 2x16GB MBD - Asus TUF GAMING X570-PLUS (WI-FI) P.S. - Doesn't seem to obvious on camera but in person it's a bit laggy. IMG_0757.mp4
-
- technical issues
- gigabyte
- (and 4 more)
-
Some points to note: I don't care about dlss and ray tracing. BUT I would like to stream on Youtube from time to time. Also I would like to use this card for atleast 5 years (pairing with r5 7600) so confused as 6900xt is already almost 3 years old card. Which one will be future proof. Currently on 1080p but would like to aim 1440p in the future (probably not 4k)(play mostly competetive multiplayers right now but don't know what the future holds)
-
So i am in europe, so want to save as much power as possible. The RTX 3090 fluctuates between 30-100w just to show the desktop, which is not very cashmoney. I have an i5 13500 cpu, and the I-GPU on it is really decent. So If I connect both my monitors to the Igpu and disable the 3090 in device manager, nothing should be able to use the GPU, meaning it should be in deep sleep right, using far less than 30w right? Games and professional programs still allow me to choose the rtx 3090, even thought it has no direct connection to the monitor (windows 11), so that shouldn't be an issue.
-
Hi everyone, first time poster long time watcher of LTT, firstly, if this is in the wrong section please feel free to move. I have a pc I’ve been building up on and off for the last few years, it started out as an xp project for the Windows XP era of gaining, it consists of the following: XFX nForce 780i 3 way sli motherboard Intel core 2 quad q6600 2.4ghz 4x1Gb DDR2 OCZ sli ready memory Nvidia GeForce 650ti Creative sound blaster X-if Elite pro Corsair 1000w modular psu Now atmo I know it sounds like a real mix of different parts, the graphics card is purely a temporary solution as I’m thinking of going with a more era appropriate sli solution, but this all comes with time and money, I am doing this on a budget after all and am in no rush to finish the project. the issue I am having is with I think either the CPU memory controller or the memory it’s self, I’ve never come across this issue before personally so will try and explain it best I can. So the system will only post if one specific memory stick is inserted into dimm slot 0, we will call this stick A, without this specific stick it will not post and will always return the error code 1D and will hang there indefinitely, without this specific stick in this specific slot it will not proceed through the rest of the self tests, with 2 of the sticks called b and c they have to be in the slots as follows b in dimm 1 and c in dimm 2, again specially these slots or it will hang at code 52, I can’t swap the sticks around into the opposite slots as it causes the system to hang indefinitely and will fail to post, stick d is my odd ball, I have a feeling it’s going bad as it’s a 50/50 if it being installed in dimm4 will cause the system to fail to post hanging at code 52 or if it will post, boot and pass mem test overnight, this has been an ongoing issue since I started this project and I’m certain something is failing and given the age of this hardware it would not surprise me, I’m curious at to what though? my thoughts are the memory controller on the cpu or the memory it’s self, funny enough my board voltages are within tolerance except the 3.3v rail that according to the bios reads 3.01v however I’ve never trusted bios readouts 100% as the read outs from the same psi can vary from board to board and considering this psu has more than enough capacity and hasn’t seen a massive amount of use I wouldn’t expect it to be struggling but it’s not out of the realm of possibility. Just more curious on anyone’s advice, it’s a weird old big heavy system and I started building it purely because I fell in love with the motherboards capabilities and the whole nvidia system, I’m by no means a nvidia fanboy, if it works I couldn’t care less for a daily system, but for a retro rig I guess I’m aiming for consistency in the theme and style lol. in terms of operating systems it’s currently using 2, good old windows XP for the true xp era games with EAX effects, hence the sound card lol and windows 7 for local connection over the internal network to my daily rig for file transfers and such, now while yes the system does indeed boot run and play games and will behave stable, it’s passed every test I can think of to throw at it, getting it booted is a challenge, it can be perfectly fine 1 day, shut it down and come back to it a day or two later and it refuses to post and I’m chasing myself round in circles reseating both the cou and memory testing each stick one by one until it posts and gradually re add each stick into there respective slots until it’s posting with all 4 sticks and can start playing some old school games. Any advice from anyone with experience with a similar issue, these particular motherboards, or anything you all might think relevant would really be appreciated, I’ve worked with quite a fe lga775 platforms in the past but always Intel based chipsets, never nvidia and so far it’s proving very capable when I can get it to boot, when being the key word here lol. Upon visual inspection nothing appears physically out of the ordinary, no signs of any bent pins on cpu socket, no visible bulging caps, no burn marks or scratches on the main motherboard, no visible damage to the RAM slots, bios has been reflashed quite a few times thinking it would help. Again anything you all can think of would be greatly appreciated and apologies if the post is abit lengthy, trying to rack my brain to give you all as much info as possible to help narrow the issue down.
- 6 replies
-
- old school pc
- retro
-
(and 2 more)
Tagged with:
-
I'm currently building a new PC and can get either an ASUS TUF Radeon RX 7900 XT or an ASUS TUF RTX 4070 ti, they're the same price for me, which one schould I choose. I know the Radeon has a bit better performance (at least from the reviews i have seen) and more V-RAM, however the RTX gives me more clearence in my Corsair 4000D Airflow, has better raytracing perfomance and is more power efficient. I would really appreciate your advice.
-
Hi everyone! So there is this GTX 1080 strix I have, it's 6+ years old at this point and the fans started dying which is normal at this age. Bought a triplet of fans from Amazon specifically made for this card, and they actually worked great for 2 days, and then suddenly they stopped. I changed it back to my old fans to (two of them still worked) but this time those doesn't even budge either. Usually when it's starts booting up all the fans start spinning and then stop because of the silent feature stuff under a specific temp, but this time they don't even start at the beginning. So basically I have a naked GPU now, and don't know what else I can do. (Don't see the point of ordering another set after acknowledging the old ones don't work either). I checked the connectors on the GPU, they seem fine, used DDU and did a clean install but that didn't help either. Any ideas? (Thanks for reading it all :3)
-
I had just recently built a new computer with a 4060ti (8gb) everything was running perfectly fine i would run valorant on 350-450 fps but just yesterday i was running it on 150 fps. Its not just valorant aswell, Destiny 2 I was running 150 fps but now im running it on 75-50 fps. I have no idea how this was caused and its scaring me because i fear that my gpu can break any day now.
-
Does anyone have any info about benchmarks on ML workloads, such as widely used model (i'm mostly interested in LLMs) inference speed for different configurations? What would be CPU inference speed on something like Xeon-W 2400? Or threadripper? Assuming any of those has comparable amount of memory to possible maxed-out M2 ultra config. How it compares against 2x or 4x 3090s/4090s? How all above compares against maxed out M1/M2 mac studio? How all above will cope with loading 30 (or even 60-70) billion parameter models? How they will work with 4-int/8-int/16-float models? And diffent inference engines (huggingface transformers/vLLM/auto-GPTQ/llama-cpp/GGLM) Aren't M1/M2 macs are actual secret ML monsters with their huge unified memory? And taking into account price of 3090/4090, M1/M2 macs can actually be a cost-effective (!) measure for this task. Here are my calculations : If I get 4x 4090, assuming each cost 1600 bucks, it's 6400. I will get 96 gigs of v-ram. And it's gpu's alone, not even talking about rest of the system. And inference speed might suffer a bit if I have to run model in distributed mode (I gotta do it even if I load llama-2 70b in int-4, for instance) If I configure m2-ultra mac to have 76 GPU cores and 192 gigs of unified memory, it will cost me .... 6600 dollars. For a bit more than 2 times of gpu-accessible ram. I need benchmarks. Does anyone have any? Mac's looking more cost-effective than PC is peak clown-world. And the funniest part, it might be a reality.
-
Hey guys, I'm wondering if anyone here owns a PNY RTX4070 XLR8 and can maybe tell me or maybe even send me a video of how the RGB effects look. I'm planning to buy it but I can't find any good footage of the RGB effects. Peace
-
Hey guys, I'm looking for a RTX4070 to replace my 3070 which is not operational anymore. I was thinking about two models: The PNY RTX 4070 XLR8 Epic-X or the MSI RTX 4070 Gaming X Trio Mainly I was choosing these two because of the noise levels and the asthetics. I've read in techpowerup's review of the PNY card that it should be very quiet (which is hard to believe since it's just a 2 slot card) and also the MSI one is claimed to be one of the quietest. Also I like the RGB on both cards because I will mount it vertically. Honestly I don't give a damn about the PNY clocking at stock speeds because it's a 4070 anyways and since I won't own any 4070 again i won't notice any difference. The only thing I don't like is MSI's software. The whole MSI center thingy f***ed me off when I owned a RTX2060 Super from MSI because I couldn't get the software uninstalled because it's a trainwreck and some of my games detect it as hacksoftware. So I prefer lightweight Software like the one from PNY, even if I din't have any PNY card. But I know for example Gainward's and Palit's software. So what you guys think? Which card should I give a try?
-
Hello everyone, I have Gigabyte geforce gtx 1650 d6 oc (rev.2.0). Who had experience with thermal pad replacement on this card, pls tell which thermal pads are needed (size,thickness).
-
Hello Forums, I have my RTX 3070ti for a while, and I was gaming fine on it and today I experienced a weird issue only when I am gaming. As suddenly the screen disconnect and blank screen shows and I couldn't go to task manager or display anything. The only thing I can do is do a hard reset on my PC. My drivers are up to date (536.67) After quick troubleshooting found out this error in event viewer: Faulting application name: dwm.exe, version: 10.0.22621.1635, time stamp: 0x6201aae6 Faulting module name: dwmcore.dll, version: 10.0.22621.1928, time stamp: 0x26557d25 Exception code: 0x8898008d Fault offset: 0x0000000000119223 Faulting process id: 0x0x43E0 Faulting application start time: 0x0x1D9C0A8FF8F0DDB Faulting application path: C:\Windows\system32\dwm.exe Faulting module path: C:\Windows\system32\dwmcore.dll Report Id: dc6b0582-dbe9-472a-9748-e20d8cbd0812 Faulting package full name: Faulting package-relative application ID: + System - Provider [ Name] Application Error [ Guid] {a0e9b465-b939-57d7-b27d-95d8e925ff57} EventID 1000 Version 0 Level 2 Task 100 Opcode 0 Keywords 0x8000000000000000 - TimeCreated [ SystemTime] 2023-07-27T16:40:09.5667621Z EventRecordID 14334 Correlation - Execution [ ProcessID] 11268 [ ThreadID] 13560 Channel Application Computer Lynx-PC - Security [ UserID] S-1-5-90-0-1 - EventData AppName dwm.exe AppVersion 10.0.22621.1635 AppTimeStamp 6201aae6 ModuleName dwmcore.dll ModuleVersion 10.0.22621.1928 ModuleTimeStamp 26557d25 ExceptionCode 8898008d FaultingOffset 0000000000119223 ProcessId 0x43e0 ProcessCreationTime 0x1d9c0a8ff8f0ddb AppPath C:\Windows\system32\dwm.exe ModulePath C:\Windows\system32\dwmcore.dll IntegratorReportId dc6b0582-dbe9-472a-9748-e20d8cbd0812 PackageFullName PackageRelativeAppId Followed by: The Desktop Window Manager process has exited. (Process exit code: 0x8898008d, Restart count: 4, Primary display device ID: NVIDIA GeForce RTX 3070 Ti) + System - Provider [ Name] Dwminit - EventID 0 [ Qualifiers] 32770 Version 0 Level 3 Task 0 Opcode 0 Keywords 0x80000000000000 - TimeCreated [ SystemTime] 2023-07-27T16:40:12.5943874Z EventRecordID 14335 Correlation - Execution [ ProcessID] 1860 [ ThreadID] 0 Channel Application Computer Lynx-PC Security - EventData 0x8898008d 4 NVIDIA GeForce RTX 3070 Ti Now at first I thought it is something was due to diablo4 hotfix update but then played another game for 20-30 min and the issue came again, the different was I could see in the 2nd game fps drop to like below 5fps and then blank screen popped. I could hear the ingame music but no display. Now I am not sure what is causing that as the only suspicion to me was windows update KB5028254 which have updates about some display settings. Not sure if I need to do ddu or remove the update and go again? Have anyone encountered that issue?
-
Budget (including currency): 1300€ Country: Germany Games, programs or workloads that it will be used for: Apex Legends, Cs:Go, GTA 5, Rainbow Six Siege Other details (existing parts lists, whether any peripherals are needed, what you're upgrading from, when you're going to buy, what resolution and refresh rate you want to play at, etc): Old pc died =new build after 8 years... Need a PC for 1440p/144hz gaming and im unsure about the CPU and GPU. Is it worth going for the Ryzen 7000 Series even though I'd need a more expensive Motherboard and more expensive DDR5 Ram, for future proofing my PC or go with maybe a 5000 Series CPU? Same goes for the GPU. Are the new 40-Series GPUs really worth it or is let's say a 30-Series more than enough for gaming? Any recommendations and/or tips are welcome
-
Hello everyone! I'm creating a topic here because I need some really good help. I've been planning to upgrade my GPU for a while, and my mind was focused on either a 3060 or a 3060ti, having watched the reviews on the channel I was pretty convinced. But lately I've been reading a lot of articles regarding the fact that nowdays 8GBs of VRAM are not actually that much anymore, so now I don't really know which card I should go for. I sold my old PC to a friend of mine, and I need to get my 1060 out of the system and put it in the other one before giving it to him, so I need a new GPU asap. My budget can go really up to 400ish euros, and my current rig I have a: Ryzen 5600X ASUS ROG-STRIX B550-A GAMING 16 GB 3200 750w ASUS ROG STRIX 80 GOLD PSU Since I don't have much knowledge of the current gen GPUs, I'm asking you guys help, either from team green of red. Thanks in advance!
-
So I have dual GPU system, a 2070 super and GTX 980. My intentions were to use the 2070 super for modern gaming, then the 980, as a capable card for rendering and outputting analogue video for retro-gaming through Batocera But, batocera cannot work with a dual Nvidia GPU setup. Unfortunately, the typical ways of disabling a GPU on linux are not available on Batocera because of how it's distro is setup. I cannot use windows, firstly because Nvidia drivers are bugged and causes a crash when setting super resolutions (needed on windows for GTX 980 video output to CRT) and Retroarch CRTswitchres, which is needed for convenient CRT video display, practically doesn't work on windows. On a previous topic I made, I was recommend virtualization and to only passthrough the 980; which I made decent progress on using virt-manager to successfully run batocera in VM; but my current barrier is GPU pass-through. I'm frankly not sure how to do it. Last time I tried, I think I accidentally bound the SSD with my linux install to the passthrough driver because they were in the same IOMMU group as the 980, breaking my install (Pop-os) Woops. Now I'm unsure what to do before going back at it. I could continue to struggle with linux and virtualization, but it's wearing on me. I think I should get the 980 in a IOMMU group that won't break linux if I have to bind everything in it to the pass through driver; and maybe there's a magic linux command that can help or I could install pop-os or a different distro to another drive and see if that helps me get the group bound so I can access the 980 with virt manager, or I could swap what slots the GPUs are in, or some combination of the above... I'm just hoping to get some more direction and help, or sane alternatives to anything I'm trying. I was also recommended to build another PC (Price concerns, a shell of a pc to put the 980 in at $75 is all I would really be able to afford, and I lose the compute of my 5600x for emulation) or sell either card and pick an AMD alternative (Selling stuff is annoying and I don't want to move to a mixed vendor setup or full AMD, and the 980 is already optimal, I can go on) Which is why I feel like I'm running out of solutions. If virtualization is a flop, for performance reasons or just a plain lack of working, I'm not sure what else to do that doesn't involve spending a bunch of money. My dream product rn would be like a PCIE to PCIE adapter that has an off switch so I could disconnect the 2070 super from the system without wearing out the connectors, so I could just use normal Batocera.
-
I updated my nvidia drivers (last version was from 2020). Now I tried to play games and after ~10 minutes of playing, my pc shut off suddenly. It keeps happening so it wasn't a one time thing. How do I fix this.
-
No, this post is not a how-to guide. I am literally making this thread just to tell you all about the Android tablet Nintendo released back in 2017 that comes with a docking station to plug your mice and keyboards into, and it also allows for video output. What do you all think of this? Is it pointless to use an Android tablet with a full keyboard, mouse, and monitor? Also, due to Android running off an SD card, it feels slow, but it actually can game quite nicely. What is it with many computers being able to feel slow but still game surprisingly well?
-
Hey there So, my last post was about my mostly bought used - upgrade of my system everything arrived, looking good (backplate of the scythe mugen is missing but after a google search it seems scythe cant produce a fitting am4 backplate anyway so i did some stuff myself to mount the cooler) Specifically the gpu: i got a RTX 3060 Ti Zotac Twin Edge Not the OC But i think( dont have the box) it is an LHR card. I dont care about mining so doesnt matter i only want to enable resizeable bar and obviously you need to update the vbios for that. and if you select the lhr version in zotacs download center, theres just no bios update. only for the non-lhr there is one and its not working with my card, tried it. gpu-z shows all indications that im able to use rebar...(Asus x470 prime pro, bios version 6063; ryzen 5 5600x, gpt boot drive etc.) Am i doing something wrong? or is there no rebar for lhr cards? Thanks in advance for your answers
-
To my understanding, a graphics card that display artifacts can either have a dying GPU or faulty VRAM, correct? In the first case, it's just a matter of time for the card to die altogether, with the GPU losing contact with the die. But if it's the VRAM, can it be used as a secondary graphics card for a SLI setup, or would the artifacts be displayed on the screen? Someone told me that the GTX 1000 series was projected to display a black screen if any sector of the VRAM is faulty, and that's why you don't often find artifacting GTX 1000 GPUs. Does it work similarly on another generation of cards, so I can distiguish between good or bad SLI candidates?
-
Hi everyone, this is a continuation from another thread where I figured out that the problem isn't windows related but GPU driver, here's the first thread: https://linustechtips.com/topic/1518959-bios-load-fast-rtx-4090-driver-load-is-very-slow-black-screen-then-w10-appears-before-password/#comment-16034102 One day I noticed that my pc started to boot slower than usual, before it was 40 seconds (and windows is loaded) but now, it takes up to 1 minute and 7 seconds to load windows, after loading bios it stays black, so I thought maybe it's a gpu driver issue. With DDU I removed it (to upgrade to a newer one) and windows loaded after only 30 seconds (gpu driverless), so defenately a gpu driver issue. When I installed the newer driver it was back to 1m and 7 seconds (before it was only 40 seconds). I always download gpu drivers manually, not with geforce experience. Here's my pc config: CPU: Amd 3950x Mobo: Asus Tuf x570 gaming GPU: MSI 4090 Gaming Trio NVME: Samsung Pro 980 2TO PSU: MPG A1000G Sound card: FocUSrite Scarlett Solo 3Rd Gen EDIT: Fast Startup is disabled, I removed all the usb connected except the mouse and keyboard, I even removed some DP and HDMI from the gpu, nothing changed. Any idea what it could be? Does anyone saw something similar? and please, I know that a fresh windows install could correct this, but I would like to avoid deleting/reinstalling all my programs and android parameters, I would like to troubleshoot and understand the problem. Thanks everyone Final EDIT: I found the solution!
-
I am planning to build a pc but I am at loss to choose between RTX 4070 gaming x trio 12 GB or RX 6950 XT gaming x trio 16 GB. enlighten me please ! i need the PC for video editing streaming and also gaming most of them are AAA games
-
Hi, I'm new to this forum. Not sure if this is the right area to ask this question. I'm using an ai benchmarker from this website, https://ai-benchmark.com/ranking_deeplearning.html. It was the only ai benchmarker I could find that I could understand. For some reason the results are in AMD's favor. I always thought ai wasn't good enough on AMD, but it's beating out the 4090! Am I using this benchmark wrong or has AMD really gotten good recently? Also if anyone knows another benchmark I could use, that would be wonderful.
-
Hello community, I have an Intel based gaming PC from 2020 with an RTX 3080 which is working great. For my brother I have build an AMD gaming PC in 2021. In first place he used an "Asus TUF GAMING OC Radeon RX 6900 XT 16 GB" which worked with Games well. But this GPU doesn't worked well with his special applications like ARCHICAD or Space Engine. Bad performance or sometimes black screen in the application. According to some forums this applications are more compatible with NVIDIA GPU's. So we bought some weeks ago an "Asus TUF GAMING GeForce RTX 4070 Ti". But the problem is with the NVIDIA, that the clock speed is always 825 MHz. I found out with the tool "GPU-Z", that bus interface is always 1.1. Power supply has 850 Watt, so there should be enough power. What I already tried: Installed new Windows 11 on additional SSD drive. Installed the newest drivers and Nvidia driver via Nvidia Experience. With the ASUS tool Armoury Crate, checked and installed drivers and also ASUS GPU Firmware (I think this is for the LED effects AURA). Installed the newest BIOS Version and reset settings to default. Checked power supply connections, two separate cables are connected to the NVIDIA power adapter. At Corsair 850 W power supply there is a switch for SINGLE and MULTIPLE. -> I set it to SINGLE. Tipp from ASUS support: Turn OFF in Windows Power Management "Turn Link State Power Management". Checked with tool "GPU-Z": PCI-Express x16 Port always use bus interface 1.1 instead of 4.0. According to ASUS Support, there could be a technical problem at the new GPU. So I send it back and I will buy a new RTX 4070 Ti. But before I send it back, I installed the RTX 4070 Ti at my Intel based PC and it performed normal. So I don't think that it will work with a new RTX 4070 Ti... Also I tested my RTX 3080 at the AMD PC and it's working good. PCI bus interface is 4.0 according to "GPU-Z" but according to UserBenchmark GPU performance is much to low... My question is, what can I try next when I get the new RTX 4070 Ti and still has bad performance? This is the PC setup: OS - Windows 11 x64 (Installed 2 weeks ago on a additional SSD for testing) OS is fully activated. Hardware is from December 2021. GPU is new. Motherboard: Asus TUF GAMING X570-PRO (WI-FI) ATX AM4 (BIOS Version: 4602) CPU: AMD Ryzen 7 5800X 3.8 GHz 8-Core CPU Cooler: Noctua NH-D15 82.5 CFM CPU Memory: Corsair Vengeance LPX 64 GB (4 x 16 GB) DDR4-3000 CL16 Storage: Samsung 980 Pro 2 TB M.2-2280 PCIe 4.0 X4 NVME Video Card: Asus TUF GAMING GeForce RTX 4070 Ti 12 GB Power Supply: Corsair HX850 Platinum 850 W 80+ Platinum Certified Fully Modular ATX Case: Fractal Design Meshify 2 ATX Mid Tower I'm a bit desperate, because I think there could be a hardware problem in a different component... Because the RTX 4070 worked at my Intel PC and my RTX 3080 had also bad performance at the AMD PC... But maybe there are some BIOS settings or Tools that I can test, if it's helping or to get information what component is not working. And when it's a hardware problem the question is if I can get a replacement from the manufacturer under warranty. Thank you for helping me. Greetings, Alex