ไม่สามารถเล่นวิดีโอนี้
ขออภัยในความไม่สะดวก
MICRON RESPONDS! What max temps should you run your G6X GPU at? RTX 3090 / 3080 Ti / 3080 / 3070 Ti
ฝัง
- เผยแพร่เมื่อ 15 ส.ค. 2024
- In this video we go over GDDR6X (or G6X for short) Memory Junction Temperature Reliability and Functionality as stated by Micron.
Note I mention Igor's Lab from here and there - I actually cut out a couple of minutes of it because most people know it anyway - but you can read their article about G6X temps here - www.igorslab.d...
Chapters:
0:00 Intro
1:05 Micron G6X Website
1:57 Memory Junction Temp - Reliability Definition
2:59 Memory Junction Temp - Functionality Definition
3:45 Memory cross section diagram
4:56 G6X data sheet - Operational Temperature - Functionality
5:58 G6X data sheet - Absolute Max Temp - Reliability
6:45 Conclusion
Sources -
www.micron.com...
My Videos testing RTX 3090 / RTX 3080 G6X temps -
Don't Buy a RTX 3090 for mining (or gaming)
• Don't Buy a RTX 3090 f...
Undervolting Guide (for any GPU)
• Undervolting and Under...
Testing RTX 3090 - Undervolting, underclocking and lowering GDDR6X Temps (Cyberpunk 2077)
• Testing RTX 3090 - Und...
Testing RTX 3090 vram in MS Flight Sim 2020 (4k/Ultra)
• Testing RTX 3090 vram ...
Testing RTX 3090 vram temps in massively modded Skyrim - Toasty or Toasted?
• Testing RTX 3090 vram ...
Testing RTX 3090: How low will the vram temps go with this $8 heatsink and fan mod?
• Testing RTX 3090: How ...
RTX 3090 (and 3080) Owners! CHANGE Your Stock Thermal Pads! (Guide, Tips, Results) w/ Dr. JimRPG
• RTX 3090 (and 3080) Ow...
Livestream - Palit RTX 3080 Gaming Pro Thermal Pad Replacement LIVE (semi fail)
• Livestream - Palit RTX...
WATCH THIS! Before Changing RTX 3080 and 3090 Thermal Pads. (Especially Palit Owners)
• WATCH THIS! Before Cha...
MSI RTX 3090 Ventus 3X Thermal Pad Replacement - Full and Unedited, (Guide/Results Video Check Desc)
• MSI RTX 3090 Ventus 3X...
JimRPG Hangout - Discord Server Invite Link
/ discord
Hey everyone, looks like there's a bit of confusion a few people asking me what's going on at 2:50 and the 92 degrees temps on screen.
1. I cut a minute's worth of video at the start where I went over Igor's Labs article (link in the description) where he says 120 degrees could degrade the memory. Here's the full quote -
"In response to questions among colleagues, for example from the R&D departments, it was unanimously agreed that the maximum temperature Ttot before the start of a possible destruction of the chip should be 120 °C and that Tjunction should probably be set at 105 °C or for the GDDR6X even at 110 °C is specified as the maximum value."
I figured the video still made sense, but I guess in hindsight just throwing out 120 degrees is very confusing. Sorry about that!
2. The 92 degrees vram shown on the screen is an older video where I tested out Skyrim trying to max out the temps there and I got to 92 degrees using 18gb vram. This shows how close you can get to the 95-105 degrees operational temperature just in gaming.
Hope that clears it up - ask away if you're unsure.
----------------------------------------------------
Either way here's a summary -
-> 95-105 degrees celsius - might have problems with your games/software, not a good idea to run it this hot.
-> 120 degrees celsius - memory will fail, shorten lifespan, die (but your GPU probably won't let you throttle past 104-110 degrees celsius anyway)
By “might have problems”, do you mean just while applications are being used, or that you are permanently hurting performance?
@@collinsutherland311 Both - you may get glitches in gaming, or you could get slower framerates because your memory has to throttle to stay cooler.
@@JimRPG my 3080ti max ive seen from it is 92 would you say this is ok to run at whilst gaming on consistent basis?
@@BoringBryan memory it gets up to 110c while mining.
Usually during editing YT videos, it is 64~70 but I've seen it go over 100c
Thanks bud.. now .. where did i keep that..
Yo zotac HOLO AMP OCed 3070ti , lets burn baby burn some games and may be the mobo.... 😁🤣🤣🤣.
Joke apart dude , where were you all these time. I get always garbage on TH-cam search on specific topics, you are like the tenth on the search result videos. 🤣🤣🤣
Nice video, thanks! It answers a very important question. One correction though: the case temperature will always be lower than the junction temperature because the junction is closer to the heat source while, due to heat exchange with air / pads & cooling plate, the case will be a good bit cooler than the silicon chip as there is a thermal gradient across the plastic case. In transistors / FET design, this difference is actually very well discussed (and critical) if you want to learn more about it.
Great review! Straight to the point with references from the manufacturer.
Thanks
Dude, you're awesome. Thanks for the info, and you've somewhat helped to sooth my anxiety over my own card. It does make me wonder what hardware failures will look like in a few years time.
When doing VRAM repair, the temperature used on the heat gun goes over 200c (close to 300c) to melt the solders for replacement work. Interesting that the VRAM modules are not damaged by that considering the max storage temperature is only 125c
Basically, Micron say to "not worry too much" when it operates at 0 to 95 C for their junction temperatures which is in line with your previous experiments.
There is a gap between 95 and 105 degrees stated by Micron; I'd like to assume it is the theoretical "safe" operating temperatures in which memory would experience some sort of degradation. I'll cross-check with Hynix and Samsung documents if possible. So far, Micron is the only company that had released fact sheet as such.
EDIT: I've made a stupid statement. Only Mircron had only made GDDR6X modules and for specific use in Nvidia cards. As such, any other data outside from Micron in terms of the temperatures would be not perfectly applicable. We really have no data on its durability per se... We're fiddling with cutting edge tech right here. At least with GDDR6, there are reports of memory failure in what temperatures and what their duration are (despite being anecdotal in nature). This is why purchasing these graphics cards with GDDR6X makes me a little bit uneasy... It has not been "tested" yet in real-life application, particularly in its reliability in comparison to GDDR6 or, say, GDDR5.
Yes its possible, though below 95 degrees is much better for the memory ;) I agree their wide margin is lame.
I wouldn't assume anything over 95 is safe even if Micron said point blank that it was. It's not to hard to keep them cooler than that I'd tell everyone to just keep em cooler. I have a mid case with 2 fans for intake and a single 120 exhaust. I average 84c on the memory while mining Ravencoin. So it's definitely possible to keep them cool, so I'd just do it.
@@whateveryousaygroomer131 in summer, were we had 34°C i was able to run the card at max capacity at 94°-96°C. worst days.... were temp hit 42°C i had to pull back the memory clock. Right now... its 26°C and i have card at full cap again with 90° BUT fans are at 85%
Well im glad my evga ftw 3 3090 has pretty decent memory cooling with adequate thermal pads on the backplate and even a heatpipe in the backplate. Doesn’t hurt that I also was able to get a 10 year warranty for 60 extra dollars. So if g6x turns out to be junk I have 10 years versus the default 3 year warranty. Not sure if other manufactures added more thermal pads to more recently manufactured cards versus those available at launch, but comparing temps to people who got the ftw 3 3090 at launch and me who got it in November it really does seem like they added some more thermal pads/ other modifications.
Manufacturers really couldnt add good thermal pads to such an expensive card...?:(
I read that connecting the memory to the heatsink actually increases the overall core temp
Still running strong with my 3090s at around 100 C.... Thanks for the update and research.
Same here,smoky strong
im at 96c and 94c, 2 cards 3090 rtx, its been mining 8 days, 117,5mh each card, big fan behind though
Mine gets easly to 100c, when overclocking the memory clock for mining, Im pretty concern about the longevity of my card, should I just bring the performance down on mh/s and keep it below 90 degres ?
How's your GPU today? Still working after 2 years?
Well, this makes me feel much better!
I did a dumb thing and overpaid for a 3090, thinking that I could let it mine in the background to have it help subsidize what I paid for it. But I don't buy new hardware all that often, so my hope was to have it last a good 4-6 years like my venerable GTX1080 has. So when mining I saw my temps hitting 102*c I freaked the hell out and started turning things waaaay down which killed any hope of profits I was hoping for, as well as efficiency.
But then I noticed that in gaming, what the card was made to do, pretty easily brought VRAM temps well into the mid 80*c temp range, which was warmer than I thought it should be. So I moved a front intake fan to the side panel to blow across the top of the card, and flipped my CPU fans from blowing up, to blowing down across the back of the card (the main heat sink wasn't even getting warm, all the heat was on the back plate)... and that cooled things down, but not by much, and when mining it still needed to be up near 80-90% fan speed which is crazy loud on my GPU.
In a final act of 'maybe it will work', I took a custom zalman cooler plate off the front of an old 570, and set that down upside-down on the back plate. A little extra thermal mass, some fins to help guide the airflow, and it cooled things down considerably, but still pretty hot an loud trying to keep temps under 90*c
But now that I know damage isn't until hitting closer to 120*c, I think I can calm down a little. Still want to shoot for under 100*c, but I can run stock while mining and get that sweet sweet 105MH/s with my 3090 now. Setting my fans back to 'auto' I am now keeping the vram under 98*c, and fan speeds to a much more human-friendly 50-60% which sounds significantly less like a jet engine. I'll probably still turn the mining off when I am actively using it just to make the thing shut up, but for the 80-90% of the day that I'm not, it can just mine away and help pay for itself.
End of the day, kinda wish I had just paid the scalper prices earlier... A year ago this thing could have been making me $8-10/day and not only paid for itself, but the whole computer! These days I'll be lucky to hit $5/day, and with ETH going away some time this summer/fall it will mostly pay for itself over the course of a year... but may fall a little bit short.
I'll have to add a few bits of something to better attach that plate to the back. It dropped temps a solid 10*c just resting on top. With an actual connection I bet it could allow things to run even quieter while mining at stock. Plus, I don't want a kiddo or critter to bump my case and have a metal plate bounce around and short something... that would be pretty awful!
Yeah a 3090 is really hot, and i swapped it out for a 3070 which is much quieter in my gaming rig. I think you're mostly ok if you're in the 80-90c range. and 90c-100c is still ok, just its less desirable. if the temps eventually degrade to over 100c you may want to look into changing out the thermal pads on the memory as that's probably why temps are increasing. there's a video i did on that - its in popular uploads.
I will add that - I think fans should be orientated so the case is in neutral air pressure. So 3 fans intake, 3 fans outake. That's just me though but whatever works best for you and agreed you can add some more metallic mass on the back of the GPU card to reduce temps.
Thanks for the info. Just got a 3090 and freaked out when I seen the memory was at 105c. So apparently that’s normal though plus I got a 2 year warranty so I’ll stop worrying about it now.
So glad to see your video, I was worried because my 3070 ti fe was hitting 81c(playing everything on max settings and around 90fps stable) I got worried because my last zotac 3060 twin edge didn't even got to 70c when I had it, it was usually always at 61c in any game at max settings around70fps.
DDR6X is hot - so turn up your GPU fans :)
Resume:
105Cº is the max operative temperature, and it will suffer damage at 125Cº
Yep!
got my hotspot up to 106c for 5 to 10 mins and didn't crash card still alive!
I think there is a big difference, when mining and hitting those temps 24/7 compared to gaming.
Excellent 3090 Series Jim and great follow up vids. Thank you good Sir.
Glad you enjoyed it!
My MSI Suprim X 3090 runs at around 82-88 and can often spike to 92c when gaming, to me even 88c is very hot and not everyone can open their GPU's and replace the pads to help with the cooling. I think Micron needs to step up and have a more manageable temperature requirements.
My 3080 suprim-x mining eth running +1200 on memory goes to 98-102 C° at memory junction, Just got today a 3080 gainward phoenix GS LHR and it goes max 80 C° while mining same settings,wow... need to change thermal pads I guess.
@@Darkroby91 thanks to LHR it's just putting less stress on memory
A somewhat decent fix without opening your gpu is get some heat sinks and put then on the back plate and then get a decent fan and put that on top. Not ideal since it makes your bulky gpu even bulkier but it does drop your temp a good 5-10 degrees.
Yes I have. It can take slight overclock and work just fine for casual gaming.
Good information thanks as always.
I have a Zotac Trinity OC 3090 and I use a Noctua 140mm 3000rpm fan on backplate and Ac running 24/7 to keep the Memory Temp at 90c while mining. Dropped the idea to change pads as Zotac said it will void warranty
Which way do you have the fan blowing towards the card or pulling air away from the card
@@JadeFicca Towards the backplate. Exactly behind memory modules.
Well thats not good, my gigabyte 3080 was being throttled when the ram hit 110c. If Micron say 95 then they shouldnt be allowing it to hit 110.
Micron said 95 Tc but HWinfo64 its showing you Tj that is around 15ºC hotter than Tc. So its all correct. 95ºC Tc = 110ºC Tj and is where throttle start
I have an RTX 3080, 12 gb, and I was just looking at my hardware temps log, and it seems when I game, the "memory Junction temp" tops at 80c, and never goes over. This is how its supposed to work I guess?
Thanks for the update on this important issue! What are the temps that you're flashing on screen while playing the games at 2:47 and 3:09 ? Were these the current in-game temps? If yes, was this before or after you changed your VRAM thermal pads? Also, ( 5:09 ) I think that T case temperatures can never be the same or lower than T junction because one is measured on the silicon, and the other on the case which is further away from the heat source. T junction temps will always be higher, no? I always try to keep the VRAM T junction temps on my 3090's at or below 90*C. I had to crank up my case and graphics cards' fans above the factory defaults to do so. I would much rather try changing the thermal pads, but I'm not to keen on losing the warranty on GPUs that cost $3000 in the current market.
Hi, do you mind if I pin your comment?
1. The temps for Skyrim and MS Flight Sim were on my MSI RTX 3090 Ventus 3X before thermal pad change. It was 92 deg on Skyrim (18gb vram) and MS Flight Sim was 88 deg (13gb vram). Yep they are in game temps. Video for before and after temps is here -
RTX 3090 (and 3080) Owners! CHANGE Your Stock Thermal Pads! (Guide, Tips, Results) w/ Dr. JimRPG
th-cam.com/video/ixjkY1Xpg1o/w-d-xo.html
2. Yes actually you're most likely right about the T-case temps being lower than T-junction. Not to blame Gamers Nexus or anything 😄 but I was watching his video here about the PS5 Teardown where the G6 memory runs hot but he says its not impossible that T-case for mem module could be warmer, he mentions typically silicon would be warmer (makes more sense like that) (th-cam.com/video/MmggkW6usmQ/w-d-xo.html) While recording I was thinking the plastic material of the memory module could absorb some of the heat without dissipating it, but i probably shouldn't have said that given I can't prove it. In any case I'm not an expert or anything so take what you can out of my material and apply it to your own use case. Also T-case was used in the past because there were no APIs on Nvidia's side to read the sensors on their G6 memory (AMD does), and so users would have to measure T-case and use equations (in that thermal application document from Micron I mentioned) to estimate T-Junction. But now you don't need T-Case you have T-Junction in HWinfo64 and GPU-Z anyway.
Also I totally understand about not changing thermal pads, completely different situation if cards were $200-300. My advice is not to change them unless you're past 95 deg and you have problems, errors, artifacting etc. I changed my Palit card and it was 84 degrees and I thought I'd be a little clever and get better temps but I actually made it worse, and had to change it 9 times to get it back to 84 degrees which was where I left it.
@@JimRPG Thanks for clarifying! No, I don't mind the pin at all. 🙂
You had to change the pads 9 times!? That doesn't sound very reassuring. 😬 I'll definitely pass for now, haha.
@@bgtubber Might have to unpin this to pin something else might pin it back later. 😄
@@JimRPG Go for it.
I mine 24/7 on a liquid-cooled 3090 FE. Sits very comfortably under 37°C on the GPU and 60-62°C on the VRAM thanks to EKWB's Quantum Vector FE RTX 3090 waterblock and the Active Backplate. Before liquid cooling though VRAM temps were above 105°C. Yikes!
Very clear explanation
Thanks!
I'm SO, so so so glad I saw this. My 3090 Xtreme memory junction hits around 92c (core around 72c, hotspot is ~10-11c delta) when gaming in PSO2 NGS at 1440p max settings, all the shinies on, at 165+ FPS. This makes me feel a bit better, though it's higher than I'd like. I've never taken apart a card before, but I might consider re-padding the GPU (with a guide or two to learn from!). I assume all 30-series cards are like this?
3070 ti and up are like this as cards 3070 ti and higher use g6x. 3070 and lower use regular gddr6 which doesn’t seem to run as hot.
It's pretty crazy that your 3080 card even gets above 100°C imo for its memory junction temp. I've been checking my memory junction temp ever since HWiNFO implemented it and I've never went above 84°C. I do run 1440p and I've tried games which are very demanding, so my gpu definitely was maxed out. But a gap over 20°C is crazy.
It was my own fault. The Palit 3080 was originally 84 degrees on the mem, but I thought I'd try to be a bit clever and make it better, but I made it worse. The tolerance for the heatsink was super tight.. about 0.1mm, and the pads had to be 0.75mm and no one sold that thickness. Of course I had no idea what was going on, so I tried with thick and thin pads, in the end after about 9 goes I realised I had to get exactly 0.75mm by squishing the 1mm pads and I got it back to 84 degrees. But 100 degrees was when the memory wasn't touching the heatsink.
Thats what I was thinking I'm hitting around 65-70c max on my 3080, over 100c is nuts.
@@madac219 65-70°C seems unrealisticly low for the junction temp. Keep in mind we are talking temps for the memory and not the gpu core. 65-70°C seems more like the gpu core temp.
@@JimRPG Yeah, that makes more sense. Should've maybe mentioned that or maybe I missed it.
mining and gaming are two very different beasts
Left my RTX 3080 Ti to mine Ethereum, memory temp was maxing out at 107C, and the card died on me. So im still concerned with what's safe and what's not.
I'd still keep it under 95 deg to reduce errors, but its still safe between 95-105 - its just not a really good idea.
@Jesus is Lord Thanks man, I’ll give it a try!
Brilliant explaination man.
Hi Jim would be glad if you could reply to this.
So basically i bought RTX 3080 few months back and had no issues with the temps whatsoever but as you know recently the Hwinfo64 started showing the Memory junction temperatures.
And it freaked me out watching it go above 90 degrees
so i put FPS limiter on my games and managed to pull the temps down in 80s
Now my question is that the maximum temperature it hits now is 90 degrees other than that it stays in 80s i wanted to ask that is it okay to play games if the maximum it goes is 90 degrees and will it affect the card or the performance in the long run if i manage to keep the temps in 80s while the max it goes is 90 degrees.
Thank you
Also is it possible that Nvidia knew that these memory junction temps would freak out the customers so they intentionally decided to hide these from us throughout these years ?
Because Nvidia says that they expect the temps to be cross 100 degrees but Hwinfo64 says that the memory will start thermal throttling if it goes 110 degrees
I really don't know what to believe out of all this.
Hey Arceuss, ive based my recommendations off the Micron comments on their website, as well as listening to Gamers Nexus, Igor's Lab and other tech outlets. Micron says errors may occur between 95-105 deg. Memory will fail, die past 120 deg. As a safe bet you want to try to operate below 95 deg anyway... once you're getting errors, its probably not good for the memory. 90-95 is hot, but the memory will continue to run as normal. In terms of longevity, the cooler you can run it the better. In your case you shouldn't have to put a FPS limiter on your card, because games in general should operate below that. If its one or two games that do, id probably just put up with it - then I suppose for those games you can put a FPS limiter. If its for every game, perhaps consider changing thermal pads but you may want to test it in a range of games before opening your card. Changing thermal pads should really be when you know this card is way too hot, the thermal pads inside have basically failed. I have changed a card hoping to get lower temps (it was getting 84deg), but in reality there was nothing wrong and I got no gains and ended up with 84deg.
@@JimRPG Hey Jim thanks a lot for you response.
The Micron says that below 95 degrees is all good but to be honest even at 85 degrees it feels very hot and as you said that the lower the temp the longer the card will survive as for that it's nearly impossible to keep the temps of VRAM in 70s if im were to game for hour or 2.
I'm thinking of putting on an aftermarket cooler on the Vram to see if it helps im not expecting a big difference but whatever helps at this point.
It's exhausting as i want to keep this card for 3 years at least.
T junction and T case are not similar at all, its not uncommon to see junction temperature be 25C above Tcase. I'm sure you've used HWInfo64 and seen how quickly the temperature drops from 90+ degrees to 70 degrees and below. Its almost instant. For example some TI parts have an operating ambient temperature of 125C but the junction temperature can be as high as 150C. If the operating range of the memory can be as high as 105C, then we can assume that the junction temperature can at least be 110C. GDDR6 max temp is rated at 95C but GDDR6X can be as high as 105C so junction temperature will be at least 110C. The biggest flaw of this video was assuming that Tj and Tc are the same thing.
Maybe, but still we don't monitor Tc, only Tj. So Tc is irrelevant.
Hi mate. I want to asked you what is should be the temperature on backplate on your 3090 when the temperature on VRAM is 100 degrees. Because i use simple mining and can’t see this temperature. Thanks.
I have MSI Ventus RTX 3060Ti 2x (OC) LHR and the temperature on idle is 44*C and on full load are 70-75*C is it normal temp range? The casing is NZXT H510 Flow, removing the side panel doesn't make any difference in GPU temp.
So ... My xfx oc 3080 ti mem goes to 104 is normal ? It stays there and everything works very good .
So my my max Vram temp of 86C which I saw on my Aorus Xtreme RTX 3080 is good?
It is undervolted to 1000Mv/2035MHz, Vram on stock. btw. I don't wanna change pads as long as it is under warranty.
Interestingly I saw 86Cmax on Borderlands 3 (non RTX game), RTX games i.e. Cyberpunk 2077 is 80-82C, Metro Exodus also 80-82C (both on max settings with Quality DLSS)
I have a 3090 Ventus, does replacing the pads void the warranty?
Wow, thanks for the research!
No problem!
Always best to go for the GPU with the best cooling solution, no matter what. With my 3080 Ti, I barely hitting 65 degrees GPU temp on heavy load (usualy just above 60), and have never topped 80 degrees on GPU memory junction. Stays cool and quiet no matter what I throw at it. No need for undervolting and mounting fans on top of it, that some do - I dont like hardware that I could cook on :P
Hey are you using GPUZ to measure the memory temperature junction is it listed as "MEMORY TEMPERATURE" (right below hotspot) on your GPUZ (if u are using that). I'm on LHR mode one @ 74% getting high 80s low 90s Mh/S and with fans at 100% my mem temp is around 96 - 100 c. any way i can reduce?
@@avar37 Hi! Sorry for late respons! I use HWiNFO64 for measurement. First, do you have good airflow in your case? If not, you should optimize your case for good airflow. A good airflow-focused pc case with good fans and the right configurations can improve both GPU and CPU temps a lot! I always use good airflow cases pared with good fans that push a lot of air at low RPMS. That way I get a quiet build with great airflow and great temps.
Great video. Thank you.
Glad you liked it!
Hi sorry im asking but is hot spot temp only the chip ( the paste on the chip) there is 3 temp in hwinfo..have a palit gamerock 3080 ti. While Gaming. The first is normal like 60-70 is the second junc temp its Also 60-70 and the last is hot spot temp and its like 100 c so about 30c over the 2 others. The hot spot temp goes up to 100 in 5 sec while Gaming. And the fans go crazy.. So i like to ask if hotspot temp is the chip only or phermal pads to? I really hope you can help me 🙏
I was so worried about my brand new 3070ti fe reaching hotspot temps of 94 but it’s just something that happens with gddr6x
Um. So is it +95 or +105? Which is it? It's a huge difference!
My 3080 FE regularly hits over 100 degrees, highest I have seen was 106. I have decent airflow and the side opened on my case…
I'm getting 108 when mining! I don't even know what I'm getting when gaming or video editing!
Same for me on my new 3090. Firestrike Extreme peaks the vram at 104 °C fresh out of the box. To say I'm worried is an understatement. I also HATE the power draw I've seen from the card -it's ridiculous, even tried undervolting but that barely helps. I also have the case side opened, 3x120mm fans at the bottom, 2x140 at the top, an additional 3x140 at the other side as intake to cool a 420 radiator. I doubt airflow gets much better than this. It's a Lian LI o11d xl case.
@@luckerNo01 I guess Nvidia GPUs just like to run hot similar to Intel CPUs.
As a watercooling hearing temps in triple digit make me go insane.
seems like originally they said its 95 °C, then later on changed it to 105 °C.
Definitely 95 below is better. A constant 105+ underload, even if its under its max permissible operating temperature, surely it will have a faster degradation over time.
Excellent
i did undervolt my 3070 ti rog strix and it reach 75c on hotspot now ... SOMEHOW still fells to much :(
3090 aurus extrem here. Did the thermal pad mod successfully and slaped 3 cm thik aluminium radiators on the back plate and a 120mm fan to top it all of. Results: from 92 max to 74 max in cyberpunk 2077 max rt 4k. 😁 even running the gddr 6x at 20gb/s. I am verry happy with the results
I have 3090 Aorus Xtreme and been thinking about doing this over the past week. Shame bscause xtreme has largest cooler of all 30's cards???
@@MinhNguyen-cl3iv its a good idea. You need 3 mm 2 mm 1 mm and 0.5 mm pads i got it from aliexpress. Same for the aluminium radiators from aliexpress. Slap a fan on the radiators and you are golden.
so when mining on the new nbminer im getting a max 94 to 96c , should i be ok with that tempature? 3070ti
hello there, thx for the video. May i ask. i have 80 ti's(msi ventus 3x) at MİNİNG. Atm they hit 110 C at hiveos. (memory temp, core temp 45 C) The question is should i worry about these tempretures and do something like undervolt or lowering overclocks or changing memory thermal pads like that(i am totaly noob at those and summer is coming) or i should not worry the memory temps BECAUSE gpu will already throttle ate 105 C so it can lower the memory temps. So it will not allow the higher temps. İf i have to intervane which solutions is the best for the lowering temps.
Hey there, I have a 3090 and a 3080. Yes I think 110c is too hot. While Micron states that 95-105C, you will get errors or lowered performance, its just not a great idea to have it above 95-100C.
When you set the most optimum settings for your graphics card in HIVEOS - it should be an underclock anyway. Check the settings and see if it says minus like (-200) for the clock speed. You also should turn the fan up to 90%. Reduce your memory overclock too which will lower the temps. It's possible these things will lower your hashrate, in that case, its about a balance between how much hashrate you want and what temps you feel you're comfortable with.
For example for my 3090 i opted for 115MH/s at 92deg for memory temp. If I increased the memory clock speed Im sure I could get the optimum 120MH/s, but my temps would also go up. So I settled for the above.
It is ultimately up to you - run it at 105C/110C if you really want. At 120C is when the mem chips fail, but id rather be a little safer for me. Hope that helps.
Thanks for the video Jim. You've made an error here, though. The junction temperatures and case temperatures are not the same thing, and counter to your assumption that the case should be hotter than the junction, the reverse is true. Think of it like a pie a few minutes out of the oven... Where is it hottest? At its core, not the pastry, since the pastry is the surface that directly interfaces with the cooling. The junction temperature is going to always be a great deal hotter than the casing temperature, which is why GDDR6X doesn't throttle the card till 110 degrees TJ. At that point the TC is going to be less than 105 degrees and well within spec. TC and TJ are not interchangeable numbers and are never the same, basically, and the reverse is true to your assumption as to which one is hottest.
The truth is you can operate these cards at 110 degrees 24/7 for years. You cannot damage the memory at these temperatures, and far too many people are worrying unduly about this when damage protection is already in place.
Hey dan, thanks for the comment. I don't recall exactly what I said it was 3 or 4 months ago, but you're correct the junction temp or basically the memory temp is the hottest part. I think I did say in one of my videos that the temps on the case could be hot (but I'm pretty sure I was talking about the case of the memory chip, not the case of the PC itself), but after saying that I think I said it's probably unlikely that's the case. Anyways I probably confused you, and confused myself, sometimes its like that for talking head videos where i go unscripted. Apologies for that.
Anyways - all i do now is look at the memory junction temp and follow that. Personally I'd rather go with what Micron has said below 95-105 deg for memory to operate within spec. and 110 deg i'd just rather not - but that's up to everyone to decide on their own. I use my 3090 for mining and I can get 115mh/s at 95 deg on the memory temps. It could probably go 122-123mh/s but its running well right now, so not too bothered honestly. Sometimes pushing it, the card can also stop mining cause it can't sustain the higher clocks.
@@JimRPG Yeah, not talking about the PC case. The casing of the chip. Always much cooler than the junction, and the 95-105 operating temp stated in the GDDR6X spec is for the casing, not the junction. That's why these cards do not throttle till 110c at the junction. In the last third of the video you state that the junction and casing temperatures are interchangeable and that the casing is likely to be hotter. Both of these assumptions are incorrect, so just wanted to pick you up on that. As you say, people are free to run their cards how they please, but I would challenge anyone to damage the card with either mining or gaming. The 110c throttle at the junction is miles away from where you might do any damage to the card. 👍
@@danhowdle9508 The absolute limit of Tc is 125 Celsius, doe it also mean that the absolute limit of Tj is above 125 Celsius since Tj is always bigger than Tc?
@@danhowdle9508 Hey Dan, you sound knowledgeable, I'm trying to understand if I have a problem or not. I have a 3090 Ventus, almost a year now. According to GPUZ when gaming at max my GPU temp is around 73, my memory 80 but I also have a hotspot that is always 103/104 at maximum. Is that different from my memory, and do I need to worry about it? My computer never crashes or behaves funnily, and my performance seems great. I run everything at stock. I'm very reluctant to open my card, but keep watching videos about how bad triple digit scores are for memory, and so I'm wondering if I need to do something. I don't know what a hot spot is though as nobody talks about that only vram temps and according to GPUZ my vram temps are good, around 80.
@@kevinbrook7033 You're specially in luck, Kevin as I have a 3090 Ventus too. No, in essence is the answer. Those temperatures are nothing at all to worry about. The ONLY reason to open your card up and replace the thermal pads is if your card is hitting the 110c thermal throttle and therefore slowing down your performance. Remember, these cards are designed to throttle at 110c, and as I was explaining to Jim, 110c cannot harm your card. It's literally impossible. NVIDIA designed the throttle limit to be far, far below the point at which the card could be damaged, and you're not even getting close to the throttle. So, in essence, you have nothing to worry about. Like you say, you've not had any problems with the card, and you won't due to temperatures. Or, think about it this way, if NVIDIA has set the throttle limit at 110c and you could fesibly damage the card at that temperature, we would all know about it because all the tech channels on TH-cam, all the PC news sites, would be running huge stories on all these failing cards. You're fine.
So for RTX 3080 what is the operating temperature is it below 95 or below 105 ??
95-105 degrees is the 'range' because there is variability between good and bad silicon. But if you operate under 95 degrees you should be able to avoid most errors or slowdown/throttling.
120 degrees is where your memory may degrade or cut its lifespan short.
I have a 3060 ti and my drivers glitched the other day and I didn't notice the fans wen't really slow, the GPU got up to 96c and the hotspot was up to 106c, this was for maybe 5 or 10 minutes, the card throttled down to 1440mhz and the system didn't crash. I assume no damage was done by the info in this video. My AMD 480 gpu ran above 80c for like 6 years of average gaming so I think my card is fine.
Yeah I left my 3090 on in a 30 deg cel room for an entire day by accident at stock settings (not underclocked). I think temps were 104 or something... I was annoyed at myself more than anything, the card perf throttled itself so it was down to 100 mh/s iirc. I think its fine. But for regular continous operation, the card is always underclocked via Hive OS.
@@JimRPG i think it's impossible to cook the cards because the firmware won't let it die on you
You can run a GPU really, really hot in general and *most* of the time you'll be fine. It's just that above certain temps, the manufacturer can no longer guarantee that no damage or degradation will occur.
It's kind of like that even your beer can has a ''best by'' date. Doesn't mean you can't drink a properly cooled can of beer if it's months past its due time - however the manufacturer won't guarantee optimal quality anymore and if it did go bad... well... that's no longer on them ;p
My rx 480 has been doing 60 c core, 110c junction. Just cleaned an inch of dust out.
Just replaced my pads with 13.8 w/mk still hitting 100c under full load
I think, everything above 90C is not normal for longterm use.
Hahahaha you said "longterm use" - AMD & Nvidia want the hardware failure of our video cards because we will be forced to upgrade :-(
Yeah i knew it bro, very sad :'(.
My 3080 TUF have been just under 110C for over a year now mining ETH. Now I am planning to raplace cards thermal pads. Do you think that memory will degrade at that temperature?
you still have this card?
thank you for this
No problems!
Very informative video... thankyou. Have the Aorus gigabyte 3090 watercooled egpu for my xps13 notebook. Works great and silent. My gpu temp never reaches 45degrees. Wish gigabyte put just as much effort into cooling the vram. Gaming is fine. But I mine at 85 mh/s just to keep the vram at 94 degrees. Want my gpu to last. Wish they allowed it to be undervolted... this would have been handy. Again not keen to replace thermal pads that would void warranty. Thought about quieter fans that have higher flow to noise ratio. Or perpaps the option to add copper finned heat sinks to the exposed alumimum heatsink above the vram. The gpu copper heatsink does have some heatpipes that extend towards the vram, but it doesn't look like it covers it entirely. Don't know what they have on the other side of the gpu board, I believe they have vram on both sides of the gpu. Not sure the cooling arrangements they have there. Again not worth smashing out an extra 30mh/s just for a few extra cents per 4 hours. Better off buying another smaller less powerful card to mine if that's what you want to do.
Unfortunately, all cooling for RTX 3090 (as opposed to RTX 3080 which is a different story altogether) I have seen so far is kinda ill designed. Memory chips are on *both* sides of the PCB while backplate is a steel or aluminum plate at best. Meaning, half of memory chips are not cooled properly. So far I haven't seen a single properly designed cooling (be it air or water) for an RTX 3090 with active cooling of both sides of the PCB. The Aorus eGPU is no exception, unfortunately, so it reaches same insane memory junction temps as any other flavour of RTX 3090 around (just run HWInfo under load). I don't know if all manufacturers are this lazy or what's the reason behind it??? It's like selling you a car with two tires top-hot-shot Pirelli and two made of chewing gum. This is just bad engineering (or a ridiculous, industry-wide scam).
RTX 3080 has only memory chips on one side of the PCB so a traditional, one-sided water-cooling with a decent, metal, passive backplate makes a lot of sense. I bet in a long run RTX 3080s will show much more resistance to a good OC and a far better longevity.
@@wiktornyckowski It appears its the Aorus egpu 3090 that's the problem... maybe such a small enclosure and not enough air flow to get to it. My fans tend not to even try to spin past the idle speed of 30% unless this thing is hot above 95 degrees. Keep wondering if they want this thing hot.
My new egpu ftw3 3090 ultra is a different story... I have this thing running at 122Mh/s and this thing is cool as 90 degrees. Haven't even played with the settings to try and undervolt. That's at stock settings and inside a case. (no open air mining). Drop the hashrate to 115Mh/s and its around 80 degrees. Feel like I hit the silicon lottery with this one. :)
My 3090 VENTUS 24 OC runs around 76-77 at around 77-80% fans. Junction is 90-93c. Hope it’s all right
great video friend, I play for a good part of the day and the other mines with my 3080ti, the maximum heat peaks in the memories is 82 C do you think it can harm?, another question leaving the card on all the time using it can cause me problems? I'm a gamer but I'm mining to pay for it, because in my country this card is worth more than 10 minimum salary, I wouldn't like a few months to see that my money went away with the card
Hi, no you're fine to mine - just use optimised settings for the card. If you're on Windows (I assume you are if you're playing games), and you're using MSI Afterburner - make sure you're using the power wattage and getting the same hashrate as - whattomine.com/gpus. For a 3080 Ti it'll be 280W for around 60mh/s (maybe its more).
Yep I know exactly what you mean - its a great way to pay back your GPU. I have been mining for the last 9 months, and paid back a couple of cards already so keep going (and maybe next gen you'll have mined enough to buy the next one!).
@@JimRPG ok thanks, for the answer you are very nice!
*I bought Msi Trident with 11tg gen i7 (which is bad and also creating a lot of heat), 3080Ti.
Just want to say to eveyone who plan to buy pc from Msi - make sure you DO NOT BUY THIS SET!!! Videos what they added on YT is not as real life use sadly :(
NOTE - check how Msi advertising their computers with any of 30 graphic crads
*Temps for me where going for CPU 88C and GPU 91C with only Cyberpunk standing on street doing nothing and not maxed graphic.
I checked with other games as some would say that Cyberpunk is AAA game etc... the same crap I heard about The Witcher 3 from MSI CS. I don't know why he didn't want to comment when I said Witcher 3 is a old game ;/
All games I tested where without max graphic and in 2k : Witcher 3 - stopped recordings, shakes frezzing etc. with 1080p works fine; Company of Heroes 2 - any place with subtitles had screen shakes; DOOM ETERNAL - all the same with super high temps, I can also add that was working far better on my old Msi pc Nightblade 3.
* There was few more games and issues but after long battle with store I had my pc changed. Working better without that many issues but I have to keep it at V-syng with min 60fps!!! If I use any more then heat and noise is just too much, I can't take it. There is no way for me to do any recordings or playing with top settings.
4k Euros MSi pc which should be for top gaming and up for me to be TOP GARBAGE!
I have 3080 ti msi gaming trio, GPU TEMP 74c/ GPU HOT SPOT TEMP 85 is good ? Tested on WARZONE VERY HIGH settings?
its hot, but safe. Your memory temps might get hot there too they can be 20 degrees hotter, so check those as well with GPU-Z.
I’ve got the fe 3080ti and after installing custom thermal pads, I never exceed 80c on memory temps. My gpu stays at 68-76c
Great job on those thermal pads!
What pads did you get exactly?
on my 3080ti founders edition
my temp is 60 deg and my
vram temp is 106 deg
at 100% fan speed
is this safe for long run ?
Update?
change them thermal pads. Just updated mine and was running around the 93-95 mining eth with 100% fan, now the mem temps stay below 82.
What?! I do not understand...
@ 2:50 - So if the memory goes to and above 92c, it is perma damaged?!?!?! but @ 3:25 - Op. Temp. is 0 to 95 c?
No the GPU may get damaged above 120 degrees according to Igor's Labs article.
the 92 degrees is the vram temp in skyrim that I hit when testing it. I added that clip to show how close it can get to 95 degrees.
@@JimRPG Ahh ok! Thank you so much for your reply! :)
I am "digging into" temps because I, hopefully, are getting an Asus Strix OC 3090 soon and really want to find out how much trouble it is to have a card like that and to keep it cooled when using it for work and gaming... not mining! Thank you for the video! :) super informative! Even if I do not understand everything! :)
@@chrisbenn The Strix is excellent and the reason there's no strix card thermal pad replacement is because its not needed. :)
@@JimRPG Ohh so i do not need to do that with the Strix! oh, thank you for that info! :) So I do not really need to do much with the Strix to keep it cool other than to have a good case with good airflow like the Corsair 5000d? I think I need a new PC case... I have a very old but good open coolermaster cases with mesh side panels but I think it is too small to fit the 3090 in it! So that is also why I am asking so many cooling questions! :D
MSI GeForce RTX 3070 Ti VENTUS 3X 8G OC running at 158W , VRAM temperature at 86C is it normal for this card ? Thanks
no you got a defect
I have two RTX3090s and don’t have temps anywhere near yours. You must have received a Monday model or your room ambient temperature is too high.
My 3090 doesn't run over 60 for mem temps. 70.max for chip.
I have the evga ftw3 so thats weird younhave a hot card.
msi 3080 lhr gaming z trio mining at 96c, is it ok ?
I have a asus turbo rtx3080 and it runs at 87c when playing 4k Max settings is that okay ?
Sounds pretty good to me. Its under 95c
My 3080 runs at 104 junc temp when mining and about 90 when gaming
I should change the thermal pads right
mine when im mining is 105-110 junc temp
@@bobtbl2868 i added a 120mm fan on top of backplate, pushing air into gpu.. helped to get 98-100c
Also as a note i had removed side panel
Ugh, my 3090 hit 110c on the vram when I wasn't look when opening up to mine ravencoin after gaming. MSI afterburner didn't intiate my settings even though I clicked it and said apply. It only hit it for a minute but still makes me nervous now.
I had it on 104 degrees for 8 hours one time about six months ago (forgot to run optimised settings in MSI Afterburner)... and the card is still going strong. its ok. It'll be prolonged heat at 120 deg that causes damage.
Personally speaking its just not a great idea to have something running at 100+ for a long amount of time anyway, lower temps are always better. Plus saving electricity too.
And this apply for Mining???
yep for mining, you should be fine for gaming, but its always good to check in case something is wrong with the card.
Out of the box my Gigabyte Vision 3090 TJunction shoots to 105°C I lower power down to 85% to get it close to 100°. Looks like I may have to try new thermal pads
i just broke my rtx2080 ti for high temperature, i wanna buy a new 3000 gpu series but if i see this video im not happy about this, which is for you the best rtx3.... for the temperature'?
I have 3090, 3080 and 3070, and I just use the 3070 in my gaming machine. The other two are mining full time. However you did have a 2080ti, so 3070 is the same performance. It might even be a bit of a downgrade because of the 8GB of ram on 3070. So this is always a bit of a problem at the high end, because you have to pay top dollar and you only get a small increase in performance.
The 3070 is at least 15-20deg cooler because it only has GDDR6 ram. During mining the 3070 is around 50 deg for CPU Core, there's no data on ram temps, but it should be 20deg warmer (based on AMD GDDR6), so about 70 deg for memory.
The 3080 and 3090 can be 55 deg for CPU core and 80-100deg for memory.
As long as you know the temp limits (based on this video), you should be fine with your card - it doesn't matter which card you get - so pick the best value or the one that best suits your needs.
@@JimRPG thanks a lot for the answer ;)
sharp review 😁 unique info, thanks 😁
for everyone that is just gaming you could undervolt your card. I undervolted my 3080 FE andmem and junction temps stay under 85c while still hitting 1950 on the core clock
Damn nice, you seem to won the lottery. 3 Question though: What do you play? How good is your PC Case Airflow? And is it hot where you live?
@@peterpfankuchen
1. Elden ring, call of duty, overwatch, street fighter, total war and tarkov at 1440p on an ultra wide.
2. Corsair 4000d airflow with a 360 aio rad in the front and 3 fans at the top pulling heat out of the case.
3. Colorado. As of right now ambient temp in the house is 75 Fahrenheit and gpu isn't going over 72c. Hotspot temp is at 77c
My 3080ti FTW3 used to reach 90 C mem junction while mining ethereum and then I changed the thermal pads on the vram modules for 1.8mm copper shims+light thermal paste and now it mines at 76 C. You have to be careful the shims don't touch other components when closing the card up because of coppers conductivity.
@Bryce Czirr It gets 90 MH/s on ethereum and 270 on ergo
@Bryce Czirr along with my 3060ti which gets another 62.7 MH/s thats more than a 3090
@Bryce Czirr And anyone that mines crypto knows that multiple miners have unlocked LHR
@Bryce Czirr Including NBminer, Trex miner, and Gminer
@Bryce Czirr Like I said, 3080ti gets around 90 MH/s and 3080 is getting around 98 because the 3080ti software unlocks are brand new. It isn't fully unlocked yet but it will get better with updates.
Is playing games with 70c Temp daily bad for GPU ?
That’s the perfect temp for gpu core, but also check your memory temp if you have GDDR6X memory. Over 100 deg cel would be considered hot. It’s not going to ruin your gpu, but long term wise it’s always better with lower temps.
@@JimRPG Thankyou.
What about playing games at 80c frequently?
My EVGA 3080 Ti is hitting 82 degrees celcius. Is that within the operating temps for gaming?
Yeah that's fine for gaming even for GPU core. it can get pretty hot while gaming.
@@JimRPG Thanks for your thoughts!
The pinned video in the description for undervolting just says "video unavailable" when I click on it.
Thanks Benjamin - there were actually more broken links there as well - i think i fixed them. Here's the video you're after.
Undervolting Guide (for any GPU)
th-cam.com/video/3wAuM8VDnbI/w-d-xo.html
Also this is part 2 - Testing RTX 3090 - Undervolting, underclocking and lowering GDDR6X Temps (Cyberpunk 2077)
th-cam.com/video/KVC3r4CRIfU/w-d-xo.html
my 3090 ftw hit 110 yeah 1hour gaming but gpu temp sit at 66c so is time to change thermalpads! just order some rn xdd
Hi thanks for the thorough explanation. It has been bugging me for a while now that my VRAM temp reaches 98 degrees at 92 MH/s with my Gaming X Trio. Anyways I have lowered the power so that the temperature stays between 90-92.
No probs. i think that's a good idea. Having to deal with warranties down the track is a hassle, plus they might not make these 'full hash rate' cards any more!
@@JimRPG Definitely. I'm keeping warranty sticker on so I can't do any thermal mods. I wish there was a way to reduce the VRAM temperature without opening the GPU like the heatsink mod on 3090s, but oh well
@@rayjeremy878 You could always just put heatsinks and the fan on top without opening the card to change the internal pads. Looks like it could net you about 4 degrees less. This is the route I'm going to go.
@@TheGridMedia This is only going to work with 3090s or with 3080s with metal backplate. I have a plastic backplate 3080. Good luck!
i had max temps of 76core/85memory and the hotspot was arround 94° but today 11 months later im getting 104° hotspot 80 core and 90 memory and its insanely loud now. too bad i got one of these Palit gaming pros with custom 0.75mm pads. what clown had this idea to use custom sized thermal pads? its impossible to find any fitting replacement pads
Same that's what I got - I assume you watched my video trying to change the pads out of the Palit 3080? I almost tore all my hair out trying to fix it.
I used 1mm pads squashed to 0.8mm (0.75mm if you can, but it gets hard to squash it further)... if you need to change them out - thats what you have to do. 104 hotspot is a bit high, so maybe try undervolting your card a bit more for the time being.
my suprim x 3090 runs since i have it over 103-104 Celisus. i send it back to msi after that 90 Celsius for 2 month and now back to 104 constant... Silent or gaming profil Realy Loud my 3080 TI Asus Tuf constant under 85 Celsius in Gaming and the hot spot is 7-10 Celsius higher then the Normal GPU Temp work fine by Asus i dont buy a MSI card again in my mind the RTX 4000 Serie is not Better we die 3000 Serie ( Temps)
2:50 you say it's 120 °C but on the white text you're saying 92 °C. Which one is it?
My bad - I cut out about a minutes worth of me talking about Igor's Lab article at the start of the video which talked about how 120 degrees was the limit where your memory would degrade. But if you listen back to 2:39, I'm referring to the Igor's Lab article about G6X temps. I cut that out of the video to get to the Micron section faster.
Here's the exact section from the article I'm referring to (www.igorslab.de/en/gddr6x-am-limit-ueber-100-grad-bei-der-geforce-rtx-3080-fe-im-chip-gemessen-2/) -
"In response to questions among colleagues, for example from the R&D departments, it was unanimously agreed that the maximum temperature Ttot before the start of a possible destruction of the chip should be 120 °C and that Tjunction should probably be set at 105 °C or for the GDDR6X even at 110 °C is specified as the maximum value."
The 92 degrees celsius is the temp from a video I did testing Skyrim temps and using as many mods as I could getting the vram usage up to 18gb vram.
@@JimRPG thanks
It's also bad cooling from nvidia side for there VRAM when they know it will hit such high temp. I have seen youtube vidoes where people customize there VRAM cooling bringing the temp down by huge margin without wasting a lot of money.
Definitely - even though it might be 'safe' at 100 deg, its always better to keep temps down so Nvidia just didn't care, but they were after 'speed' to sell more cards.
Well, my 3080 TUF had 112 degree on stock settings, not full 5 minutes load in furmark. In games it had similar resoults - all in good ventilated case and ambient of 20 degree C. ASUS Customer service sayed its OK (!!!!), i didint listened to them, sended GPU to RMA anyway, and now waiting for resoult. First it was in seller magazine for their own testing, they said they ran Furmark for 2 days (!!!!!) and it didnt overheated - definetly lie. They sent it over to ASUS now, this is final moment. If ASUS will do nothing, its gonna be definetly last of their products i bought, and not gonna be silent about it, if they fix it, i will try to sell it, as this GPU wasnt that hot since day 1, it started overheat after 6-8 months, and it look as poor termopad solution. I really wanted to replace termopads myself, but when i saw prices i abandoned this idea - i still didint "opened" card.
Have installed ekwb full cover water on 3090ftw3. Have memory Tj between ±60C and gpu less than 40C, -400@725mV +1300 pl 74% 124mh ±300W. Price for full cover ±250euro for each card + any water accessories
Nice, ate you running a custom you bios ?
I got lucky I got a 3090 ftw3 at 90c im getting 125mh/s consistent if im not using the computer and 119-124 if I'm using it.
Thats mad haha! nice one. I'm at about 115 mh/s in my rig and I don't really want to push it anymore haha.
90c is pretty good honestly.
@@JimRPG update it's been 2months now I'm still getting 125-126mh/s but at 95c maybe the thermal pads are degrading now than yet again the advice normal operating temp is 95-105c. Just want to know how your 3090 doing or any news on the topic. Thank you again for bringing this information forward!
ps. i just added a 3070 FE in the case I'm running that with the 3090 that might be a factor.
what rtx have best ram temps asus or msi?
Asus TUF and ROG Strix. But MSI is mostly manageable.
I mean with MSI you can underclock a bit or change out the thermal paste/pads, if you want to hit the same temps.
My 3080 ti runs at around 85-89c so that's fine I'm guessing?
Yeah that's pretty good for the memory. Don't touch it imo.
Meanwhile Zotac 3090 hit the 110c limit and throttle.
Yeah I have 3 of those and it's a constant heat battle
The stock cooling systems are insufficient.
How long have you been mining at 90 degrees for example..
A little over a year. Started Jan last year.
@@JimRPG so 90 seems average mining
Great video, I have a 3080 FE using it for gaming only. VRAM temps hangout between 89-93C when gaming (Metro exodus 1440p max settings/Warzone 1440p max settings. Would prefer not to open up the card and do the thermal pad mod given how hard it is to get cards currently. Plan is to monitor it and if it starts being 95C or above with any consistency at that point I’ll pursue the mod. Should I be concerned with those current temps with just gaming?
You can easily undervolt the card and not affect performance whist it does lower the temps quite a bit. This is an easier way to do it than replacing thermal pads.
@@jonboy2950 thanks! I’ll look into it
Looks like I'm going to need to replace some thermal pads on the FE 3080ti -_-',,
2.64-3.5 30% increased on heat
nice i wish to buy 3080 ti in my country but is price to high and no stock and no income. I am said because the advertising for these graphics cards are like there are enough graphics cards on the market. However, the manufacturer himself has never given information that he apologizes and that they cannot produce them because they are limited to external contractors, but it is really so difficult to communicate this to the public unless mining technology is more important than gamers or content creator. All they like to do is report new and better GPUs and high prices. THank you for this information @JimRpG
Hi Denis, GPU shortage is a big problem, but there's a shortage in many things including consoles, CPUs, power supplies even Teslas. For some people there is a choice, you can use the GPU to make some money back and mine with it. Obviously not every one can do this because mining is banned in some countries, but for some its possible. Otherwise to be perfectly honest, any GPU from the last 4 years is ok, so thats a RTX 20 series, or RX 5000 series, though I understand they're also expensive. Now that I've tried a 3080 and 3090, I don't think its an absolute must have over say a 2060/2070, if you're ok to play at medium/high graphics for the time being. Getting 20-30fps more on a top of the line GPU for $$$$ isn't worth it especially if you're not mining with it either. Hope you get a GPU soon.
@@JimRPG Thank you for thanks for the quick exhaustive answer right now i have gtx 1050 ti and i wish to buy i new GPU because of rendering some models. I have one more question it is batter to buy AMD or NVIDIA GPU for rendering in Blender.
Because AMD have BETTER vram 16GB - 6800Xt or 3080TI of 12GB What You recommend? Or maybe wait to lunch GPUs SUPER NVIDIA iI?
@@denisj32 I'm not an expert on Blender - but a quick look around, it looks like Nvidia is much better even if you have to go with the 10gb 3080 or 12gb 3080 ti, rather than the 16gb AMD cards. Best to double check that.
@@JimRPG Ok thank you
What about the Micron memory from 2 years ago on the 20 series? The degridation was horrible.
If we still have one of those cards that is still working, is it safe to overclock it?
Somehow my gut says no.
Thank you for the throught information! Just dont put those videos with confilcting info on the background the nextime... You almos gav me a heart attack when talking about permanent damage temperature with a video showing a giant red 88C on the screen 😂😂😂😂😂