Gaming Performance Tested On 'Worn Out' RTX 2080 Ti Mining Card

[K E N N Y]

Multiple RTX 2080 TI graphic cards were tested for GPU performance drops after 1.5 years of hard GPU mining. These cards were ran hard for a year and a half and were tested for gaming performance drops against a brand new 2080 TI. YouTube channel Testing Games posted a new video to look into the matter. On average, the used mining 2080 Ti was about 10% slower than the brand new 2080 Ti. One outlier was Forza Horizon 4, which showed the mining card as 20% slower than the new card. One extra note is, the GPU thermal pads or paste were not replaced on the mining cards. This is could change the results drastically. 

 

 

 

Summary

Interesting reduction of GPU performance after 1.5y of mining.  

 

Quotes

Quote

The main culprits for the reduced performance are GPU clock speed and temperatures. On average, the heavily used RTX 2080 Ti mining card was 16C hotter than the brand new RTX 2080 Ti. This caused over a 100MHz drop in boost frequency for the mining card, creating the performance losses. This is totally normal as Nvidia's GPU Boost 4.0 algorithm (equipped on Turing and Ampere-based cards) is tuned to be very sensitive to GPU temperature.

 

My thoughts

This news is more definitely interesting. Can put thought into buying used GPU's or even the importance of card maintenance, thermal pads, paste or even fan health.  

 

Sources

 https://www.tomshardware.com/news/mining-card-rtx-2080-ti-tested-gaming

 

[Levent]

20 hours ago, K E N N Y said:

One extra note is, the GPU thermal pads or paste were not replaced on the mining cards. This is could change the results drastically. 

And that alone invalidates that claim, not even worth clicking.

20 hours ago, K E N N Y said:

showed the mining card as 20% slower than the new card

[PlayStation 2]

Surely you'd clean up a mining card with new thermal pads and paste after using it for a year and a half as a mining card, no?

[LAwLz]

4 hours ago, Levent said:

And that alone invalidates that claim.

And it's some unknown YouTube that might as well be lying to get views. 

 

But yeah... A cards cooling getting slightly worn out and therefore not boosting as high isn't exactly big news. 

You might actually experience the same performance drop from just the cooling getting clogged with dust. 

I don't really think mining on the cards will cause any special performance drop you wouldn't see from let's say a card used for gaming. 

[Downkey]

think reapplying the thermal paste would be the solution...

[Blademaster91]

Sure replacing the thermal paste and thermal pads fixes the performance issue, but the average gamer buying a used GPU isn't going to do that, also the only GPU brand I can think of that doesn't void your warranty for taking apart a GPU is EVGA.

[Levent]

29 minutes ago, Blademaster91 said:

Sure replacing the thermal paste and thermal pads fixes the performance issue, but the average gamer buying a used GPU isn't going to do that, also the only GPU brand I can think of that doesn't void your warranty for taking apart a GPU is EVGA.

Sure, it makes more sense to buy a brand new graphics card to replace a one that is not performing as it did on day one because of $5 worth of thermal paste and pads.

 

29 minutes ago, Blademaster91 said:

average gamer buying a used GPU isn't going to do that, also the only GPU brand I can think of that doesn't void your warranty for taking apart a GPU is EVGA.

Then "average gamer" will go on and create a thread in a forum like this one asking why their system is getitng 20% less FPS in a game compared to a youtube video.

[Moonzy]

interesting, definitely not something I've experienced on my GPU that i've mined on since 2017

 

what can cause a 16c delta though? my gpu temps barely rose 5c compared to new and my annual (or so) cleaning brings it back down there

 

edit: ah, this is bullshit

Quote

Testing Games didn't benchmark the used mining card with replacement thermal pads and a fresh new application of thermal paste. Oh, and cleaning — dust buildup in the heatsink fins can also greatly hinder cooling performance. 

another misleading media to paint miners in the black light, very typical

 

i can agree on paste and pad degradation and the average user wont know to replace it, but dusting? please

[DildorTheDecent]

Tom's Hardware really has hit the shitter if they're posting garbage like this.

 

A dust clogged, dried out card runs hotter and doesn't boost as much as a fresh one? What an eye opener! /s

[wanderingfool2]

21 minutes ago, Moonzy said:

interesting, definitely not something I've experienced on my GPU that i've mined on since 2017

 

what can cause a 16c delta though? my gpu temps barely rose 5c compared to new and my annual (or so) cleaning brings it back down there

 

edit: ah, this is bullshit

another misleading media to paint miners in the black light, very typical

 

i can agree on paste and pad degradation and the average user wont know to replace it, but dusting? please

The video itself isn't well done (in that they don't provide any summary information)...but at the same time, they didn't provide any information about the condition of the video card...unless I am missing something, which if so feel free to correct me as I honestly could have (See below)

8 minutes ago, DildorTheDecent said:

Tom's Hardware really has hit the shitter if they're posting garbage like this.

 

A dust clogged, dried out card runs hotter and doesn't boost as much as a fresh one? What an eye opener! /s

What's interesting is that Tom's Hardware just magically comes to the conclusion that it was clogged in dust.  It would be safe to maybe assume that the pads/paste weren't switched, which I think the average gamer wouldn't know how to do.  The source video though isn't well done though, in that they don't even show pictures of the video card.  For all we know, they could have switched the thermal pads/paste (doubt it though).

 

Really though, there isn't enough information present to really conclude anything about the condition of the card.  There could be speculations based on the higher temps, but without either quotes from Testing Games or something such as images of the graphics card, it's a bit unreasonable to assume they didn't dust it.

 

With that said, the idea is actually pretty good.  It would be better if someone like LTT or Gamer's Nexus could do similar tests (before and after thermal paste replacement as well).  So it could be like a "straight from the shipper" vs "Paste replaced" vs "Stock".

 

39 minutes ago, Levent said:

Then "average gamer" will go on and create a thread in a forum like this one asking why their system is getitng 20% less FPS in a game compared to a youtube video

I doubt the average gamer would replace their thermal paste.  Even look at people like xQC and how terrible he was at assembling the PC, or for that matter look at a lot of the other gamer streamers who can't figure out computer equipment

[Moonzy]

8 minutes ago, wanderingfool2 said:

It would be better if someone like LTT or Gamer's Nexus could do similar tests (before and after thermal paste replacement as well).  So it could be like a "straight from the shipper" vs "Paste replaced" vs "Stock".

i forgot the conclusion since it's been so long ago

 

chips dont really degrade slowly like that

it's either working, or you have a dead GPU

[StDragon]

Are we sure it's thermal limited? Could the capacitors be worn out thus starving the GPU for power?

 

FWIW, I've never re-greased or re-applied new pads to my card. Is that something that's expected to be done every few years to prevent that 10% - 20% performance drop? That's really substantial! The only thing I've done was give it a good dusting by blowing all the crap out.

[Moonzy]

Just now, StDragon said:

Could the capacitors be worn out thus starving the GPU for power?

this would cause crashes, not throttling

derbauer de-soldered some caps from the back of a GPU and tested it

 

2 minutes ago, StDragon said:

FWIW, I've never regressed or re-applied new pads to my card. Is that something that's expected to be done every few years to prevent that 10% - 20% performance drop?

not to my knowledge '-'

i always keep track of my gpu temps and i never notice it drop, never did a paste swap on any of my GPU (other than 1, but that's while diagnosing another issue)

[StDragon]

6 minutes ago, Moonzy said:

this would cause crashes, not throttling

derbauer de-soldered some caps from the back of a GPU and tested it

 

He de-soldered caps to test a theory of signal filtering. In addition worn out capacitors aren't the same as failed (removed).

 

But yes, I've see MBs with failed caps where malfunctions occurred to the point of throwing kernel panics (BSODs). I guess it depends on how well the GPU can recover from being starved of power. 

[wanderingfool2]

25 minutes ago, Moonzy said:

i forgot the conclusion since it's been so long ago

 

chips dont really degrade slowly like that

it's either working, or you have a dead GPU

Yes, this one is a lot better video; thanks.  Still has issues in that they don't use the same model, but their conclusion is still pretty solid.  The Tom's Hardware though still bugs me in the sense that they didn't list their source for it being "dusty" and without changed pads...actually, the article itself really shouldn't have been written given there isn't any real substance to it.

 

With that said as well, it could change based on the type of person you bought the mining card from (someone who undervolts/doesn't undervolt) vs someone who has 1 GPU mining vs a whole bank of GPU's mining (in hot environments). My guess is that it's still the thermal paste though that got roasted (as opposed to dust)

[leadeater]

Where is the 1.5 year old only used for gaming control card? Where is the test of at least say a hundred cards so we can know the standard deviation, could just as easily be card to card difference an nothing else. Dumb test is dumb.

[StDragon]

1 minute ago, leadeater said:

Where is the 1.5 year old only used for gaming control card? Where is the test of at least say a hundred cards so we can know the standard deviation, could just as easily be card to card difference an nothing else. Dumb test is dumb.

I agree. While the premise is sound, the entire test and review was half-assed. Waste of time for what otherwise could have been real insightful. 

 

As others have said, guess we'll have to wait for GN to provide more scientific analysis.

[Middcore]

53 minutes ago, wanderingfool2 said:

I doubt the average gamer would replace their thermal paste.  Even look at people like xQC and how terrible he was at assembling the PC, or for that matter look at a lot of the other gamer streamers who can't figure out computer equipment

 

There is a very popular Twitch streamer (250k+ followers) who I am going to leave nameless. One time while I was watching his stream he described how he didn't like the noise from the fans of his prebuilt gaming PC, so he called up customer service, and they told him "Well since it's liquid cooled you don't really need fans anyway, just stick a screwdriver in the fan blades on each one until they stop turning." So he did. 

 

I'd like to think he was just trolling but I can't bring myself to believe he was. Unfortunately he did not mention the name of the prebuilt system integrator who gave him this sage advice. 

[RejZoR]

Well, thermal pads or not as a factor, if card operates slower because thermal pads and thermal paste were so deteriorated, that's a legit and valid concern.

 

I mean, how many casual users ever replaced thermal paste on any graphic card? I've done it, but I generally avoid dealing with thermal pads because thickness is a legit issue and they aren't cheap if you look for good ones. Not to mention replacement of thermal pads is not a trivial task in a lot of cases because they come applied in different thickness and applying too thick or too thin thermal pads may actually do more harm if not applied and verified correctly. And many tear when taken apart and you can damage them when verifying.

 

So, even if electronic components weren't deteriorated, cooling system was and the card performed significantly worse as a result, replacing thermal paste and pads is a significant investment and operation on a card. Just dismissing this is just silly.

[SingleTMat]

I'd like to see them test again after the thermals have been cleaned up. Assuming they didn't take any damage over the time during mining, they should still perform the same as when they were new, shouldn't they?

[Middcore]

1 minute ago, SingleTMat said:

Assuming they didn't take any damage over the time during mining, they should still perform the same as when they were new, shouldn't they?

 

Well, Linus did the "Is performance degradation on GPU's real?" video a few years ago.

 

The short version is GPU's do "wear out" but more in a "suddenly just don't work anymore" kind of way than a "gradually work worse and worse" kind of way, 

[Blademaster91]

1 hour ago, leadeater said:

Where is the 1.5 year old only used for gaming control card? Where is the test of at least say a hundred cards so we can know the standard deviation, could just as easily be card to card difference an nothing else. Dumb test is dumb.

I agree there needs to be more than 2 cards tested, it could vary from someone that plays games every day compared to someone that doesn't, maybe even case airflow or high end AIB cards vs cheaper versions.

[SingleTMat]

9 minutes ago, Middcore said:

 

Well, Linus did the "Is performance degradation on GPU's real?" video a few years ago.

 

The short version is GPU's do "wear out" but more in a "suddenly just don't work anymore" kind of way than a "gradually work worse and worse" kind of way, 

I wonder what people would consider the best practice for making a card that is used heavily or used for mining to last the longest. Like, how often should you redo thermals and whatnot. Seems like it would be a hard thing to even test due to time limitations and product variations.

[MageTank]

I don't know if I would attribute mining to the cause of this potential degradation in performance. Speaking from my own experience, I had a 1080 Ti FTW3 Hybrid that saw very light usage (MMO gaming) with the added benefit of running cool thanks to the inclusion of the integrated AIO. I owned it for about 2 years before it started flaking out on me due to a cold solder joint and I refused to RMA it to EVGA as I had already RMA'd the first for my own screw-up with liquid metal and I didn't want to take advantage of their generosity a second time.

 

After leaving it sit on my shelf for another year and a half, I decided to go back and fix the card by manually reflowing the solder by hand. When I went to remove the AIO, the stock thermal paste had solidified into a chalk-like material and was no longer malleable at all. It wouldn't even leave any residue on your fingers when you touched it, no discoloration, it simply broke apart. This was a card that was thermally sound and never exceeded 40C operating temps during my gaming sessions yet a little over a year on the shelf and it would have likely been rendered thermally inoperable had I managed to fix it without repasting it and testing that paste.

 

My point is, depending on the quality of thermal paste used on these cards and something as mundane as their storage environment, you may still run into thermal issues regardless of how the card was originally used.

 

I will say, the card is back up and running, though it's a shadow of its former self, as I turned it into an external graphics monster for my brothers GPD Win Max.

[kakik09]

Wonder what kind of cooler was being used here. Not sure if the triple fan MSI card listed is the control or the actual model used for mining.