Be honest what is your gpu running out on full load...does having a box full of fans really drop it more then 2c?
Open Test bench's are nice as a baseline to take case variation out as a factor for tech tubers or flat comparisons across a distance between yourself and whoever you are shareing information with. However That doesn't mean that open test bench > than a case ...IF <-- case is setup to direct air where it needs to go and exhaust it at a rate where it has time to cool the target hardware.
Gamers Nexus does a good job considering the amount of hardware they get to check out, a very good job even vs the amount of staff they are using.
The room they use is referenced as having a controlled 21C ambient Temp.
My setup while being 3 C higher on ambient is consistently 6-8 C lower avg during gaming loads than the benchmarks they use on that test bench.
But then my case also has the GPU /CPU seperated from eachother . 1 can't leak/bleed heat to impact the other. I also have my power supply seperated and out from under my Graphics card. Past 2 years or so Ive primarly ran 60 series cards seeing how much I could push what was the "budget option" . Sadly now it would seem no card made after pascal is available for moderate budgets ;/.
Anyway while I only test in C long enough to make sure Im where Id like to be vs popular benchmarks as Fahrenheit tho much frowned upon by us "educated nerds" is more easy to relate too . as 48c vs 51 C ..thats only 3 degrees! no biggie ..sure and 1 C is the same /test error. This is a mistake in my opinion as I can tell physically the difference if my house is 70F vs 78F . and 3c depending where u are on the scale is near 8 degree's F . 50C seems fine in general till you realize its over 130F ..literally scorching hot.
F gives a more relatable measurement for some of us Americans anyway and a much more granular sensitive dare I say it ..reading for anyone . But then I don't like having alot of Decimals on my monitors aside from Voltage sensors.
TLD I run 240m cheapest "budget" AIO available last year on 2700 non X with blck OC and a llc bug on x470 gigabyte boards boosting to 43xx (no pbo on that series non x chips) avg 48c gaming 497 cpuz sT score 1980something CBr15 with a static oc at 4.35ghz I wanna say id have to look . Still think its #1 2700 (x or non x) without exotic cooling on Firestrike coupled with 1660TI lol.
Currently running 3600x 51 c gaming avg (bf5 full map rotation for reference) Turns out the Little pump/block sitting on top of the processor gets hot , oddly the rear of the motherboard gets hot for that matter. Fans can fix this if directed and channeled where they need to be while exhausting air at a rate that doesn't mess up the idea in bringing air into /across targets. My Gpu has its own fans obviously I don't care about noise as this is easily mitigated as a case Modder . without harming the look even tho it takes some work. I had 5 fans back in 1995 playing Quake 2 trying to maintain 300 frames cuz i liked the # lol on open GL and Nvidia's first TnT card .
I now run 4 fans push/pull on that cheap 240m aio for cpu ...Pulling air in , the cold side of the aio i let have more time and it gets blown by fan #5 in cpu area across the block/Ram dimms/VRMs and exhausted out the back by fan #6. Then another fan(#7) keeps my front IO heat from interfering in my plans #8 and 9 and both being intakes for the GPU with #10 being the Gpu's exhaust assistant ... #11 pushing cold air across the back of the Mobo where it exists the rear of the case.
Could I live with my 1660TI last year avging more than 48C in 22/23c ambient room? sure just didn't want too n i liked being # 1 on air on a few sites . N battling the Eastern Euro questionable truth using zombies on userbenchmark list haha.
Would I be ok with the hot spot on my 2060Super FE being more than 60C? no ..im not even ok with it being over 58c but I wanna stay on air for GPu as most ppl can't afford water blocks or want to install them . N while I can n have in the past Its not practical to most gamers that Im trying to assist or even build for ..I don't bother.
TLDR ...Do you hit smaller and smaller returns with a higher # of fans to where its not useful ? Yes however most ppl who parrot that have never actually hit that point even if they believe they did . Truth is untill the heat generated by the Fan's center mechanism starts to out weigh the additional control of the air said fan provides added on top of the rest of the fans in the system that also generate heat during operation .. u aren't there. Years of tinkering with the case I currently use which U could purchase but Ive modded it so much hard to recognize it other than exterior shape.
BFV cheap and well enough known game and a good stress on both components even in the menu along with a good memory stresser I use it alot but , I hit 146avg fps with 2700non x and 1660TI dual fan card took awhile but did it vs 101 or whatever most tech benchers would suggest at 1080p ultra (minus undergrowth/terrain / Rtx off) 175 with 3600x /2060super.
My system is low end vs this years components that are now available and was even at launch just mid tier items , and or super cheap.
But it was doable at a reasonable price even counting the fans . Don't get caught up watching tech tubers who while knowledgeable don't tend to suggest "Open air setup > any case on air" I think for good reason as in My testing if effort is applied in the setup ... Open air isn't better even with a massive fan blowing on it unless u go sub ambient on that open bench its not gonna beat my case as u won't be able to control if the air supplied has enough time to cool the components .
To OP .. I dunno man id try and add more fans haha lord knows next upgrade I do likely 5800x or 5600 (rather go back to 16 threads tho to test on next cpu cooler i try out ) and 3060 TI / 3070 i'll run more tests see if a new area due to GPU's design and yet un purchased undecided aio coolers behavior , requires yet another fan or just a change in direction.
