OC Scanner Issue

[kylejs]

I have an RTX 2080 and when i do the oc scanner, it takes about 5 minutes to do the first 3 points, and then it runs until 25 minutes. I turn it off,  because by then the temps are pretty high. Should I let it keep going? I couple times it has worked normally, and it worked perfectly for a friend.

[Jurrunio]

how hot?

[kylejs]

56 minutes ago, Jurrunio said:

how hot?

83 degrees, max it can go.

[Jurrunio]

1 minute ago, kylejs said:

83 degrees, max it can go.

which model out of so many with 2080? If it can hit 83C with OC Scanner, then so it will with manual overclock which is a sign that this card is not suited for such thing without lowering power limit first.

[kylejs]

 

On 4/1/2020 at 10:32 PM, Jurrunio said:

which model out of so many with 2080? If it can hit 83C with OC Scanner, then so it will with manual overclock which is a sign that this card is not suited for such thing without lowering power limit first.

I have the msi rtx 2080 gaming x trio and a ryzen 7 2700, another thing I noticed is that with a 83 degree limit, its gone up to 86 during the process of OC scanner. Also, it says the boost clock for the card is 1860, but i noticed in game in fortnite with ultra settings the clock was 1920 mhz. Im wondering if my card is damaged? or if something else is wrong

[Jurrunio]

31 minutes ago, kylejs said:

 

I have the msi rtx 2080 gaming x trio and a ryzen 7 2700, another thing I noticed is that with a 83 degree limit, its gone up to 86 during the process of OC scanner. Also, it says the boost clock for the card is 1860, but i noticed in game in fortnite with ultra settings the clock was 1920 mhz. Im wondering if my card is damaged? or if something else is wrong

Something definitely wrong with the thermal paste or mounting pressure if case airflow doesn't suck, you could verify by taking off the side panel while running OC scanner

 

and that clock speed is normal behaviour, GPU boost 4.0 on your card does that.

[kylejs]

Just now, Jurrunio said:

Something definitely wrong with the thermal paste or mounting pressure if case airflow doesn't suck, you could verify by taking off the side panel while running OC scanner

 

and that clock speed is normal behaviour, GPU boost 4.0 on your card does that.

if that boost behaviour is normal, how come it says Boost: 1860 MHz on the msi website? And I have good airflow, my idl temps are around 30 c, and while around 50-65 while gaming.

[Jurrunio]

5 minutes ago, kylejs said:

if that boost behaviour is normal, how come it says Boost: 1860 MHz on the msi website?

It's like an extra level of boost from GPU boost 4.0 which is a way that helps squeeze out extra performance at lower temperatures (as transistors work better when they are cooler). You cannot understand this the same way as boost on intel CPUs, though it is similar to Ryzen's boost behavior but Nvidia chooses not to specify the max boost so customers dont complain "why my GPU doesnt hit this"

[kylejs]

1 minute ago, Jurrunio said:

It's like an extra level of boost from GPU boost 4.0 which is a way that helps squeeze out extra performance at lower temperatures (as transistors work better when they are cooler). You cannot understand this the same way as boost on intel CPUs, though it is similar to Ryzen's boost behavior but Nvidia chooses not to specify the max boost so customers dont complain "why my GPU doesnt hit this"

Oh! Thats actually really cool. But one big thing that is worrying me, is this: https://www.gpucheck.com/gpu/nvidia-geforce-rtx-2080/amd-ryzen-7-2700/  if you scroll down to all the benchmarks, and set the settings to low, it says the average fps for fortnite battle royale is around 350 fps. with a 1% low of 314. I have never hit 314 fps in game. The only time close is in the lobby, loading screen, or when the match first starts and i look up into the sky.

[Jurrunio]

2 minutes ago, kylejs said:

Oh! Thats actually really cool. But one big thing that is worrying me, is this: https://www.gpucheck.com/gpu/nvidia-geforce-rtx-2080/amd-ryzen-7-2700/  if you scroll down to all the benchmarks, and set the settings to low, it says the average fps for fortnite battle royale is around 350 fps. with a 1% low of 314. I have never hit 314 fps in game. The only time close is in the lobby, loading screen, or when the match first starts and i look up into the sky.

That's why such "benchmark site" should not be used. Instead of testing games with actual hardware, what they do is extract benchmark data of individual components in an environment that hopefully doesnt bottleneck them (so 2080 tested with an 9900k, 2700 tested with super aggressive memory settings and 2080ti etc) and then spit out numbers extrapolated from that. They are meaningless.

 

others simply use user reports even, which is even worse

 

2nd gen Ryzen just wont go so high in frame rates, that I can be sure. Also high end custom models of a 2080 should do even more than 1920MHz that you now get, though 86C in OC Scanner is definitely limiting it (expecting around 70C). I would expect a Ventus XS (not even the large fan Ventus) to run that hot, not a Trio.

[kylejs]

1 minute ago, Jurrunio said:

That's why such "benchmark site" should not be used. Instead of testing games with actual hardware, what they do is extract benchmark data of individual components in an environment that hopefully doesnt bottleneck them (so 2080 tested with an 9900k, 2700 tested with super aggressive memory settings and 2080ti etc) and then spit out numbers extrapolated from that. They are meaningless.

 

others simply use user reports even, which is even worse

 

2nd gen Ryzen just wont go so high in frame rates, that I can be sure. Also high end custom models of a 2080 should do even more than 1920MHz that you now get, though 86C in OC Scanner is definitely limiting it (expecting around 70C). I would expect a Ventus XS (not even the large fan Ventus) to run that hot, not a Trio.

Well i have a custom fan curve, but that makes it the fans go harder. I tried to use the auto in afterburner and it just seemed too hot. The thing is, the day before that, it worked a few times and got an average overclock of 65, but because it went so hot, i got worried and tried again the next day. I tried it about 3 or 4 times the next day, letting it run in 20 minute intervals, and it never finished, and it got to around 84 max. Could that have damaged it? Because on the msi website it says that oc scanner is completely safe and does not void warranty. And whats weirder is my friend with a dell optiplex with a new power supply and a gtx 1650, his oc scanner finished in 14 minutes, and only got to around 70 max.

[Jurrunio]

2 minutes ago, kylejs said:

Well i have a custom fan curve, but that makes it the fans go harder. I tried to use the auto in afterburner and it just seemed too hot. The thing is, the day before that, it worked a few times and got an average overclock of 65, but because it went so hot, i got worried and tried again the next day. I tried it about 3 or 4 times the next day, letting it run in 20 minute intervals, and it never finished, and it got to around 84 max. Could that have damaged it? Because on the msi website it says that oc scanner is completely safe and does not void warranty. And whats weirder is my friend with a dell optiplex with a new power supply and a gtx 1650, his oc scanner finished in 14 minutes, and only got to around 70 max.

That's why I recommend checking the mounting pressure (i.e. whether the heatsink is loosely attached to the card) and thermal paste (may require RMA if the disgusting warranty void stickers are in the way)

 

From my experience OC scanner is not that stressful (I've seen games push my 1070 harder than it) so it definitely shouldnt run that hot.

[kylejs]

Just now, Jurrunio said:

That's why I recommend checking the mounting pressure (i.e. whether the heatsink is loosely attached to the card) and thermal paste (may require RMA if the disgusting warranty void stickers are in the way)

 

From my experience OC scanner is not that stressful (I've seen games push my 1070 harder than it) so it definitely shouldnt run that hot.

what would the mounting pressure of the heatsink do? i got this card christmas before last I think, but I do remember having it in a cabinet under my desk, and sometimes the  door was closed, and it was at 81 for around a hour of playing, but then i just left it open from then on. Now i have moved and it is on my desk, and I just dont get how it could get that hot. I wish I could take it somewhere if it weren't for this virus.

[Jurrunio]

1 minute ago, kylejs said:

what would the mounting pressure of the heatsink do?

The tighter you push your hand into something, the better you feel its temperature. This is roughly the same principle, that it increases heat conductivity (since air is really bad at that, also helps thin out thermal paste)

 

this is what happens when a heatsink is badly attached (different card but concept is the same)

 

[kylejs]

18 minutes ago, Jurrunio said:

The tighter you push your hand into something, the better you feel its temperature. This is roughly the same principle, that it increases heat conductivity (since air is really bad at that, also helps thin out thermal paste)

 

this is what happens when a heatsink is badly attached (different card but concept is the same)

 

Wow! That's amazing. Ill definitely try that later when its all cooled down. Thanks! Do you think that could actually be the issue?

[Jurrunio]

4 minutes ago, kylejs said:

Thanks! Do you think that could actually be the issue?

All I know is the Trio cooler should do much better.

[kylejs]

2 minutes ago, Jurrunio said:

All I know is the Trio cooler should do much better.

alright. And should I keep my fan curve or just use the auto in afterburner(2nd pic)

[Jurrunio]

1 hour ago, kylejs said:

alright. And should I keep my fan curve or just use the auto in afterburner(2nd pic)

Your fan curve looks ok, though I myself would use a faster refresh rate for fan speeds (2 seconds)

 

Custom fan curve or just factory defaults is up to you.