Did I just seriously lose the silicone lottery?

[MVPernula]

Okay so just got my GTX 1080 MSI Armor OC and Firestrike keeps crashing on below 2000MHz. I'm uncertain about how good/bad that is. Thoughts?

 

Right now I'm testing:
+180 Core Clock
+500 Memory Clock

100% Fan mode

121 Power limit.

 

I was really looking forward to getting it up on 2GHz.. Seems I wont be going there.

[DevilishBooster]

What are the temps hitting?

[MVPernula]

Just now, DevilishBooster said:

What are the temps hitting?

74 Maximum.

[App4that]

Lower the memory clock to stock and try again.

 

[Vikesh]

i have heard that 2000+ mhz are possible with FE cards 3rd party are stuck slightly below that but its just a rumor

[Vellinious]

1 minute ago, MVPernula said:

74 Maximum.

That's why you're not hitting higher clocks.  Need to keep the temps lower.

Get GPUz sensors tab open, and have it running during your testing as well.  Pay particular attention to the PerfCap Reason line.

[MVPernula]

Just now, App4that said:

Lower the memory clock to stock and try again.

 

Will do. Testing what I said on my first post now and it seems stable so far, butt I'll try going up to 2GHz with base mem clock after to test that as well.

[App4that]

Just now, MVPernula said:

Will do. Testing what I said on my first post now and it seems stable so far, butt I'll try going up to 2GHz with base mem clock after to test that as well.

The core clock influences performance more than the memory, always start with the core and then see what the memory will do. Otherwise you might lose out of core speed.

[MVPernula]

Just now, Vellinious said:

That's why you're not hitting higher clocks.  Need to keep the temps lower.

Still hitting that but it seems alright atm. Noted however.

[i_build_nanosuits]

6 minutes ago, MVPernula said:

Okay so just got my GTX 1080 MSI Armor OC and Firestrike keeps crashing on below 2000MHz. I'm uncertain about how good/bad that is. Thoughts?

 

Right now I'm testing:
+180 Core Clock
+500 Memory Clock

100% Fan mode

121 Power limit.

 

I was really looking forward to getting it up on 2GHz.. Seems I wont be going there.

thoughts?! you bought the cheapest end model from MSI...of course you got a bad silicon...

they have the Gaming, gaming X and Gaming Z models above yours...and then the REAL good ones are probably all being set aside waiting to receive the MSI lightning board and cooling solution.

[Wander Away]

1 minute ago, Vellinious said:

That's why you're not hitting higher clocks.  Need to keep the temps lower.

Get GPUz sensors tab open, and have it running during your testing as well.  Pay particular attention to the PerfCap Reason line.

 

I'm pretty sure 74C is pretty healthy for a GPU

[MVPernula]

Just now, Wander Away said:

I'm pretty sure 74C is pretty healthy for a GPU

Just my thought.

[i_build_nanosuits]

Just now, Wander Away said:

I'm pretty sure 74C is pretty healthy for a GPU

it is...anything bellow 80c is perfectly fine.

[MVPernula]

8 minutes ago, DevilishBooster said:

What are the temps hitting?

 

4 minutes ago, App4that said:

Lower the memory clock to stock and try again.

 

 

2 minutes ago, Vellinious said:

That's why you're not hitting higher clocks.  Need to keep the temps lower.

Get GPUz sensors tab open, and have it running during your testing as well.  Pay particular attention to the PerfCap Reason line.

 

Just now, i_build_nanosuits said:

thoughts?! you bought the cheapest end model from MSI...of course you got a bad silicon...

they have the Gaming, gaming X and Gaming Z models above yours...and then the REAL good ones are probably all being set aside waiting to receive the MSI lightning board and cooling solution.

Finished Fire Strike now with http://www.3dmark.com/3dm/13458357 as my result with the settings I posted above. Will mix around a bit more though.

[Vellinious]

1 minute ago, Wander Away said:

I'm pretty sure 74C is pretty healthy for a GPU

Not if you're overclocking.  The cooler you keep them, the higher they'll bost.  The thermal layers in Maxwell are even more defined with Pascal.  

[i_build_nanosuits]

Just now, MVPernula said:

 

 

 

Finished Fire Strike now with http://www.3dmark.com/3dm/13458357 as my result with the settings I posted above. Will mix around a bit more though.

@App4that is right though...you should try to let go the memory overclocking for now...first find a stable core clock and then start messing with the VRAM oc.

[Wander Away]

Just now, Vellinious said:

Not if you're overclocking.  The cooler you keep them, the higher they'll bost.  The thermal layers in Maxwell are even more defined with Pascal.  

 

If it's just a problem with the temperature though, firestrike wouldn't crash, the clock would just be pulled back. 

[MVPernula]

Just now, i_build_nanosuits said:

@App4that is right though...you should try to let go the memory overclocking for now...first find a stable core clock and then start messing with the VRAM oc.

Yes I'm doing that now! Trying on +203 core clock with memory on stock.

[Mk7-Golf-R]

2 minutes ago, i_build_nanosuits said:

thoughts?! you bought the cheapest end model from MSI...of course you got a bad silicon...

they have the Gaming, gaming X and Gaming Z models above yours...and then the REAL good ones are probably all being set aside waiting to receive the MSI lightning board and cooling solution.

I have 2 armor 1080s in sli. Both of mine peak at 2100mhz about like all the other cards.

 

I managed to get the #1 spot on 4k Valley benchmark.

 

 

The armor isn't bad cards, I've had 1070s and 1080s in them. 

 

 

 

[App4that]

3 minutes ago, MVPernula said:

 

 

 

Finished Fire Strike now with http://www.3dmark.com/3dm/13458357 as my result with the settings I posted above. Will mix around a bit more though.

Keep pushing, you're just over a 980ti. http://www.3dmark.com/fs/9006144

[themctipers]

Meanwhile I can't get a boost from 990 -> 1000mhz on my 380 without instability. 

 

And my 3770k reaches 105c under load and takes 1.365v for 4.5ghz (80 idle at this speed) 

[MVPernula]

Just now, App4that said:

Keep pushing, you're just over a 980ti. http://www.3dmark.com/fs/9006144

Can't seem to hit that 2GHz sweetspot, even with memory clock on stock.

1 minute ago, rrubberr said:

Unless it's an AMD card, in which case anything below 100 is okay, but the cutoff is 105.

 

Not necessarily true, but did the original poster ever mention overvolting, or did he just raise the power limit and expect it to work? 

The 1080 seems to have a voltage lock, can't budge it no matter the settings.

1 minute ago, Mk7-Golf-R said:

I have 2 armor 1080s in sli. Both of mine peak at 2100mhz about like all the other cards.

 

I managed to get the #1 spot on 4k Valley benchmark.

 

 

The armor isn't bad cards, I've had 1070s and 1080s in them. 

 

 

 

This I like!

[TheRandomness]

1 minute ago, themctipers said:

Meanwhile I can't get a boost from 990 -> 1000mhz on my 380 without instability. 

 

And my 3770k reaches 105c under load and takes 1.365v for 4.5ghz (80 idle at this speed) 

Then there's something wrong. 80 idle means something's wrong.

[othertomperson]

12 minutes ago, Vellinious said:

That's why you're not hitting higher clocks.  Need to keep the temps lower.

Get GPUz sensors tab open, and have it running during your testing as well.  Pay particular attention to the PerfCap Reason line.

It doesn't throttle that aggressively! It's still worth setting the temperature limit in Afterburner manually.

 

Also put the voltage and power sliders up a bit. It won't use the maximum power or voltage unless it needs to anyway.

 

I wouldn't have expected you to need +180 just to reach 2000 either. Remember you're offsetting GPUboost, not stock. +90 gets me ~2050. If I run at stock it naturally wants to be ~1911 MHz.

[i_build_nanosuits]

7 minutes ago, rrubberr said:

Unless it's an AMD card, in which case anything below 100 is okay, but the cutoff is 105.

AMD can claim all they want but a piece of silicon is a piece of silicon and running it at 90c is bad...AMD say 100c is fine because they fucked up and pushed cards that do run at 100c with the fans at 100%.
AMD make power hungry cards and they tend to run hot...and also tend to fail prematurely you have no idea how many ''did my GPU just died on me'' threads we had this past year concerning 290 and 290X GPU,s just tapping out...100c is not fine...it's really not fine at all unless you plan to run the card for 2 years tops and don't mind having the card failing on you.