Question are bottleneck calculators reliable?

Mar 6, 2020
47
2
35
0
just ran this test

and it's showing that the radeon hd 6970 which is a hot runnin power hungry monster from 2010 with 2gb ddr5 is much more powerful than a gtx 1050 with 3 gbIS this accurate?

I need to replace my hd 6970 for my 4400 i5 and I want something good, but affordable without pushing bottlenecking too far. I'm looking at 1650 super, 1060 and other stuff in that range if anyone has any ideas.

i would upgrade a much better card if i found a good deal so i could use it in a later build, as long as the bottlenecking the CPU is not harmful?
 

USAFRet

Titan
Moderator
Mar 16, 2013
135,003
6,447
165,690
20,800
"are bottleneck calculators reliable?"

Short answer - No.
Long answer - Nooooooooooooooooooooooooooooo
Actual answer - They are pure snake oil. No connection with any reality. You can't "calculate a bottleneck like that. It depends on the particular application or game, and your settings.

They are total bunk.


Q. Which of these is correct, 1.87%, 3.42%, or 18%?
A. All of them. Or none of them.
Same exact parts (mine), over time.
 
Reactions: sizzling
Mar 6, 2020
47
2
35
0
should i just say F it yolo and buy the best card I can buy and not worry about bottlenecks? cuz 1060 6 GB looks perfect at 170 bucks, but at 200 I can get into the 8GB range of cards
 

InvalidError

Titan
Moderator
should i just say F it yolo and buy the best card I can buy and not worry about bottlenecks? cuz 1060 6 GB looks perfect at 170 bucks, but at 200 I can get into the 8GB range of cards
Depends on what the 8GB choices are since there are many slower GPUs with more VRAM. With a faster lower-VRAM GPU, you always have the option of lowering details until the game can fit reasonably comfortably in whatever VRAM you have and get pretty good frame rates. With a slower higher-VRAM GPU, you may end up lowering details enough to get adequate performance that the extra VRAM may end up mostly useless anyway. For example, in most cases where the 8GB RX5500 leads its 4GB counterpart by substantial margins, the 8GB model is already barely adequate and many people would end up lowering details anyway.

As a budget/casual PC gamer myself, I'd be far more interested in seeing GPUs benchmarked at whatever details preset they hit 60fps 1% lows at to avoid over-emphasizing extra VRAM at Ultra/Highest details that don't make sense when they bring the GPUs below 60fps average, never mind 1% lows.
 

ASK THE COMMUNITY

TRENDING THREADS