• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

ASUS Strix RTX 3080 Running Below Advertised Boost Clock

Associate
Joined
30 Aug 2012
Posts
8
Hi!

I recently purchased an ASUS Strix RTX 3080 V2 LHR from Overclockers, replacing my MSI Gaming Trio X 3080 which I sold on to fund it. Unfortunately, I've been left with a bitter taste in my mouth as it's seemingly underperforming. For reference, the advertised boost clock of the card is 1905MHz (1935MHz with GPU Tweak OC Mode).

Running Time Spy, my graphics score for the new card is in an average range of 17300-17400 out of the box...from what I can gather it's not far off ballpark for a 3080, but still pretty low for a card like the Strix in my opinion. My previous MSI card for reference, reached 17700-17800 out of the box. In real world terms, sure, it won't make a huge difference, but it's still a disappointing result.

ASUS Strix Time Spy
MSI Trio Time Spy

What I've noticed however, whilst the benchmark is running, the card will boost to 1935-1950MHz at the very start, but it will start to tumble pretty quickly as soon as it bounces off the power limit. It then spends the rest of the benchmark running in a range between 1830-1890MHz. In terms of temperature, I've not seen it break 72 degrees.

I figured I'd try some games to see if the same thing happens. In Control, once again, the clock speed drops to and settles around 1860-1890MHz with only the occasional increase back above 1900MHz when load reduces. Same thing with RDR2 and GTA5.

Things I've tried so far:

- Clean install of several different driver versions
- Reseated the graphics card
- I am using 3 separate, stock PCI-E power cables to a Corsair HX1200i
- Checked input voltages in GPU-Z (all ~12.1v minimum under load)
- Checked input power in GPU-Z (~110W on each of the 3 inputs, and ~40W from the PCI slot at max TDP)
- Uninstalled Armoury Crate, MSI Afterburner (just in case)
- Changing to High Performance power plan (nets me a slight increase to 17400-17500, nothing earth shattering)
- Running all case and GPU fans at 100% (nets only very small gains)

The rest of my system is:

Intel i7 10700K (stock, no OC)
ASUS Maximus XII Hero
32GB Corsair Vengeance 3600MHz RAM (XMP is on)
Corsair HX1200i
Corsair 5000D Airflow Case

My question is, should the card be able to sustain at least the advertised boost clock under load, or am I mistaken and just unlucky? I have pinged Overclockers support about it, but I figured I would try and gather some opinions while I wait for a response.
 
Soldato
Joined
22 May 2010
Posts
8,060
I think this card has a Dual bios switch, have you checked to see if its switched to Quiet Mode or Performance Mode? Seems quite hot running for a big card like the Strix, my guess is its in Quiet mode and heating up because the fans aren't spinning as fast as they should be therefore lower overall performance after certain load times.
 
Associate
Joined
30 Aug 2012
Posts
8
No, it's already switched to Performance Mode. :(

There's little more I can do for airflow as I've already recently switched over to an airflow centric case. But I don't think 71 is too hot by any means, especially for a 3080. The temperature is much lower than my previous 3080, which would happily go up to 78 without a custom fan curve to keep it in check.

Just to check, I've just done another Time Spy benchmark, but with all fans in my case and the GPU running at 100%. Temperature maxed out at 65 degrees, and I still only got 17384 with the associated dips in clock speed (1830-1890MHz).
 
Permabanned
Joined
31 Aug 2013
Posts
3,364
Location
Scotland
You could try an undervolt for which CP2077 is a good test. I run my original 3080 TUF OC at 900mV / 1920MHz, which can hit 1935MHz at times. I think the max temp I've seen form it, in a Fractal Torrent case, was 62C with all fans left at default.
 
Associate
Joined
30 Aug 2012
Posts
8
Yeah, undervolting did wonders for my previous 3080 in terms of temperature, although I ultimately ran without it as I didn't want to deal with potential instability at the time.

I just wanted to establish whether the new card I have is working properly out of the box before proceeding with stuff like undervolting. It certainly doesn't seem to want to sustain the advertised boost clock for very long, resulting in low benchmark scores, which is why I'm concerned.
 
Associate
Joined
5 Mar 2017
Posts
1,815
Location
Cambridge
Managed to get rid of a troublesome RX6800 XT and got a MSI 3080 Suprim X, 800ish pounds, open box. All adhesives in, not a finger print on the card. Who knows.
The card boost to advertised speed and some. No coil whine and dead quiet, even using performance BIOS. This card simply killed watercooler for me for as long as I use it. Memory Hotspot maximum 82C, GPU after 3D Mark Stress Test Port Royal, 20 times, 66C, HotSpot mid 70s.
The lack of the passthrough on the backplate, and the card being vertically mounted, the thermals in the O11 EVO with front mesh panel are optimal. NVME and Chipset considerably cooler than ever before. Room temperature similar.
As you were using a 3080 before, I wouldn’t even speculate about power delivery from your PSU.
I assume drivers have been reinstalled, just in case?
For temperature, I would try open side panel and maybe add a fan under the GPU, as your case, despite good airflow, the lack of bottom mounted fans was a reason I decided to sell it. I’m always after the very best thermals, and trying to avoid watercooling.
 
Associate
Joined
30 Aug 2012
Posts
8
Yep, tried a few different drivers using DDU in between each install.

The PSU is actually a brand new replacement from Corsair as the previous one's 24-pin socket was keyed incorrectly unbeknownst to me. Cable still fit and it worked fine for a year, but I discovered that the holes were the wrong shape when I went to tidy up some cable management a few weeks back! I'm having a run of bad luck with PC hardware it seems...

But the new PSU seems to be working fine (according to GPU-Z/HWInfo/iCUE at least), with good voltages being delivered to all inputs on the card (12.1v minimum under load) and the power looks to be well balanced. In fact the voltages are higher and more stable now as I was getting droop to 11.8v with my previous PSU.

I still can't imagine temperature being the issue either. Maxing the fans brings it down into the 60s but I'm still seeing drops below advertised boost and little to no gains in benchmarks. ~70+ degrees should be fine for pretty much any GPU anyway from what I can gather. My MSI Trio would run up to 78 degrees in my old cramped 500D case (and it constantly bounced off its power limit too) but it still didn't drop below its advertised boost like this current one as far as I'm aware. It'd usually stay above 1900MHz.
 
Associate
Joined
5 Mar 2017
Posts
1,815
Location
Cambridge
The main offender for some thermal throttling would be Memory Junction nowadays. Easier to keep the core and GPU hotspot under control than the memory’s temperature.
In very few occasions, if the delta between Core Temperature and GPU hotspot is too high (20+ Celsius), that may cause some issue. Most cases is just uneven plate pressure, and very rare bad thermal paste application.
With my previous 6800XT repasting I managed to drop the delta to 10-12C, but the issue with the card was that it would boost too high and crash with anything with Ray Tracing. Later had the same issue with non-Ray Tracing applications. Clearly some issue with the card.
I would try with HWInfo64 and check the main things, like GPU temperature, Hotspot and Memory Junction.
Also you can check if the card is being limited by something. They’ll list voltage, thermal, etc.
 
Associate
Joined
30 Aug 2012
Posts
8
It's always power limited according to Perfcap reason in GPU-Z, which was my experience with my previous 3080 too.

Hotspot temperature is the highest of the 3, running ~16 degrees above GPU temperature on average, which is within the normal delta range, from what I've read. Pulling some points from my GPU-Z log file during two Time Spy benchmarks:

At stock, Performance mode fan speed:

72 GPU = 88 hotspot = 76 memory.

With fans running at 100%:

64 GPU = 80 hotspot = 68 memory.

Clock speed drops as low as 1830 in both cases. :(
 
Soldato
Joined
15 Oct 2019
Posts
8,449
Location
Uk
Since they have so many GA102 models in use now it's probably just a case of the silicon not being as good on the standard 3080s now with the better binned dies going to the higher speced models.
 
Soldato
Joined
9 Mar 2015
Posts
4,407
Location
Earth
Your card is technically performing to Spec. The base clock of the card is 1440 MHz. The boost clock as others said is when there is either thermal or power headroom. I expect the voltage out the box its consuming is pretty high, possibly due to not being a great chip, therefore smashing into the power limit quickly (hence what your seeing in GPU-Z).

In actual gameplay and use, the slight knock in clock speed or numbers vs your MSI card will be unnoticeable. Would defo however play with undervolting and overclocking at same time.
 
Associate
Joined
30 Aug 2012
Posts
8
Ugh that's frustrating. Seems a bit pointless buying a stupidly expensive card like the Strix if there's a risk of it rarely sustaining the high boost it advertises. One of the reasons I swapped to it is because it's supposed to perform better than most, if not all other models. Comparing my 3D Mark results to other models seems to suggest mine is the complete opposite...

I guess I figured the advertised boost was supposed to be a minimum and it'd boost past that if it had the overhead, as was the case with my MSI. I mean if the card was so bad that it ran exactly at base clock, they would just shrug their shoulders and brush me off? Surely that can't be the acceptable. The base clock figure is almost never mentioned on product pages either, adding fuel to my assumption.

But yeah, you're right, it doesn't seem to make much difference in performance...and it looks much nicer than the MSI, at least. :p
 
Last edited:
Associate
Joined
15 Sep 2009
Posts
360
Location
London
Run GPU-Z whilst you're gaming and watch the PerfCap Reason under Sensors - it should tell you why it's downclocking (power/thermals).

*Edit* I just noticed you tried that - if it's power then you need to either increase the power limit or undervolt it - it's already doing the best it can at stock.
 
Associate
Joined
30 Aug 2012
Posts
8
370W out of the box I believe. I haven't tried undervolting yet as I'm still trying to determine if the card is working as advertised out of the box. I personally don't think it is, but that seems to be open to interpretation.

But yes, during a typical benchmark run or intensive game, it starts around 1935-1950MHz, but within seconds it hits the power limit, drops and then fluctuates between 1830-1890MHz. It doesn't want to sustain at or above the advertised boost clock for very long at all.

My previous card only had 320W to play with and it too was always riding the power limit, but didn't drop as much, despite having a lower advertised boost clock.
 
Associate
Joined
19 Sep 2009
Posts
2,298
Location
Riedquat system
My 3080 FE seems pretty inefficient compared to some on here meaning at stock it will drop to consistently due to hammering the power limit. Even undervolted seems worse than most who have posted on the forums here.

Quake 2 RTX at 4K drops to 1815MHz with power slider max (~370W) and like 1770MHz at stock power (320W)

With a 937mV undervolt I can get about 1900MHz out of it but still drops occasionally in games with high power usage.
 
Top Bottom