Most Important GPU Specs - Comparing Mismatched GPU's?

[JimmyRecard00]

Hello

So I'm still trying to get a handle on balancing different types of specs in different arrangements...

eg. which would be better...

6600 XT 8GB see specs here


Base Clock
1968 MHz

Game Clock
2359 MHz

Boost Clock
2589 MHz

Memory Clock
2000 MHz

16 Gbps effective


Memory Size
8 GB

Memory Type
GDDR6

Memory Bus
128 bit

Bandwidth
256.0 GB/s




or


RTX 3060 12GB see specs here

Base Clock
1320 MHz

Boost Clock
1777 MHz

Memory Clock
1875 MHz

15 Gbps effective



Memory Size
12 GB

Memory Type
GDDR6

Memory Bus
192 bit

Bandwidth
360.0 GB/s


Is the Memory Bus Width the most important for RAM specs? Is Memory Clock not soo important?

if a card had either half the core clock or half the ram of an otherwise identical clock... but had double the Memory Bus -would that cancel out the difference? How do these things balance?

And how much and which factors for rendering/cuda etc matter the most... are they no where near as important as the clock speed and ram specs?

again see those two links for the details in two different boxes
Theoretical Performance and Render config


To my undeducated eye - the 6600 XT looks like it might be better simply because the much higher base and boost clocks - however it has 4GB less RAM than the 3060?


Another example....

These below some have double the CUDA but nearly half the GPU Core Boost and also very different Memory Bus Width.... what order would you rate these card and if the best one of these was 10/10 what would you rate the others too?

Does double Cuda always trump half the GPU clock? And the Memory Bus?

RTX 3060 Ti	$700	4864	80	1410 MHz	1665 MHz	1750 MHz	GA104	17400M	8 GB, GDDR6, 256-bit
RX 6700 XT	$700 MSRP: $480	2560	64	2424 MHz	2581 MHz	2000 MHz	Navi 22	17200M	12 GB, GDDR6, 192-bit
PowerColor RX 6700 XT	$750	2560	64	2424 MHz	2649 MHz	2000 MHz	Navi 22	17200M	12 GB, GDDR6, 192-bit
RTX 2080 Ti	$1000	4352	88	1350 MHz	1545 MHz	1750 MHz	TU102	18600M	11 GB, GDDR6, 352-bit

It's this and similar kinds of conundrums I'm trying to get a handle on - what are some simliar pairings/mismatchings that I'm likely to come across looking at cards like this?

Does power usage matter at all... or just matters what it does with it?

Is there anything else in particular that is really important to look for when buying a GPU for pretty good quality (not average not insane ultra)?


Cheers

 

[Kaamos_Llama]

Hello

So I'm still trying to get a handle on balancing different types of specs in different arrangements...

eg. which would be better...

Is the Memory Bus Width the most important for RAM specs? Is Memory Clock not soo important?

if a card had either half the core clock or half the ram of an otherwise identical clock... but had double the Memory Bus -would that cancel out the difference? How do these things balance?

And how much and which factors for rendering/cuda etc matter the most... are they no where near as important as the clock speed and ram specs?

again see those two links for the details in two different boxes
Theoretical Performance and Render config

Nvidia uses CUDA cores, and AMD uses stream processors, they have different designs entirely and performance cannot be compared by measuring clock speeds and core counts. It can only be compared by measuring actual in game performance. Which is why I have been posting benchmarks and tier lists for you to look at so you can judge which cards are better on average.

To my undeducated eye - the 6600 XT looks like it might be better simply because the much higher base and boost clocks - however it has 4GB less RAM than the 3060?

VRAM amount is not a good measure of the strength of a card at a certain point. Some cards have more VRAM then they could ever use.

For example, look at the Radeon VII in this table, it has 16GB of a more advanced form of VRAM (HBM) and 3840 stream processors and yet, performs around the same as the Radeon 5700XT which has 2560 stream processors and 8GB of technically slower VRAM, because of other architecture advances.

AMD vs Nvidia is even less comparable in this way because the designs are completely different

Another example....

These below some have double the CUDA but nearly half the GPU Core Boost and also very different Memory Bus Width.... what order would you rate these card and if the best one of these was 10/10 what would you rate the others too?

Does double Cuda always trump half the GPU clock? And the Memory Bus?

See above

Does power usage matter at all... or just matters what it does with it?

Is there anything else in particular that is really important to look for when buying a GPU for pretty good quality (not average not insane ultra)?


Cheers

For the purposes here, power usage only matters for your electricity bill, and for the size of PSU you need to buy.

So, finally please check out Toms tier list to see that the AMD 6600XT is very slightly faster than the RTX 3060 on average over Toms suite of games they tested.

There may be specific games in which the 3060 wins, but by the time this level of card is using textures and settings that would push over 8GB of VRAM in a game you would need to lower them to improve performance anyway.

I'm not sure at this point why you are blanking my posts, but these have been good questions for beginners, that someone on a more intermediate level like myself can answer for others who might come across the posts with similar questions.