MSI GeForce RTX 3060 Gaming X Trio

[Deleted member 50521]

@W1zzard Still no plan to test mining performance? This is especially relevant now for 3060

 

[W1zzard]

@W1zzard
Can you test on AMD SAM ?

It only makes a small difference, but yes, next test system is Ryzen + PCIe 4 + SAM

@W1zzard Still no plan to test mining performance? This is especially relevant now for 3060

No plans. I tested it briefly, mining rate dropped to half the initial rate within 1 minute or so, as mentioned in the conclusion

 

[thebluebumblebee]

RTX 3050 Ti

 

[N3M3515]

Very underwhelming, i thought it would at least be on par with the 2070 Super.
This is a $250 MSRP gpu.

 

[birdie]

• Barely faster than the RTX 2060 Super
• Barely more power efficient than the RTX 2060 Super
• Overpriced as F
• Halved mining rate in order to sell worthless crap/e-waste called CMP

I'm disappointed. My 1660 Ti will remain with me until NVIDIA releases a whole better midrange card at a sane price.

 

[BSim500]

W1zzard, is the "power consumption for 60Hz VSync" figure correct?-

RTX 3060Ti = 73w
RTX 3070 = 73w
RTX 2070S = 90w
RTX 2070 = 114w
RTX 3060 = 183w

Aside from being out of whack with all the others, it actually seems higher at "60Hz" than under normal load. Or is the chart a mis-labelled "Furmark" chart (it's in the same place as older GPU reviews)?

Edit: Same iffy power figures with the other EVGA, Palit & Zotac reviews.

 

[Shatun_Bear]

'Starting at £450...'. What a waste of time this line-up is. £550 will be the average price!

 

[r9]

Kicking myself for missing out on 1080ti for $250.

 

[mechtech]

W1zz, did you have a chance to test out the anti-miner driver to see if it delivers??

Very underwhelming, i thought it would at least be on par with the 2070 Super.
This is a $250 MSRP gpu.

More RTX, moar shaders, more silicon more $$$

 

[cst1992]

@W1zzard Power consumption figures seem odd.
Gaming: 177W
Maximum: 176W
V-Sync 60Hz: 183W

Something's fishy here. I'm thinking you need to recheck this.

 

[ppn]

Performance gain generation to generation:

GTX 760 -> 960 = +90%
GTX 960 -> 1060 = +72%
GTX 1060-> 2060 = +59%
RTX 2060 ->3060 = +19%

Why no reviewer is calling 3060 for what it really is, a total shitshow even at (non existent) MSRP?!

LOL take it easy, GTX 760 to GTX 960 is 10%, or call it single digit, you meant 9% right?

Yeah after the colossal failure comes the 1060. so have patience for 4060.

70-tier is always 50%.
60-tier is a little dodgy sometimes matches one generation back like 960=680 instead of 780, sometimes 1060=980.

 

[Shatun_Bear]

LOL take it easy, GTX 760 to GTX 960 is 10%, or call it single digit, you meant 9% right?

Yeah after the colossal failure comes the 1060. so have patience for 4060.

70-tier is always 50%.
60-tier is a little dodgy sometimes matches one generation back like 960=680 instead of 780, sometimes 1060=980.

The generationl jump between the 2060 to this is bad enough but when you add the price on top, you have one almighty stinker.

 

[tfdsaf]

Performance gain generation to generation:

GTX 760 -> 960 = +90%
GTX 960 -> 1060 = +72%
GTX 1060-> 2060 = +59%
RTX 2060 ->3060 = +19%

Why no reviewer is calling 3060 for what it really is, a total shitshow even at (non existent) MSRP?!

Probably because they are scared they will get black listed by Nvidia and won't get cards anymore. That doesn't excuse them not calling out Nvidia on this one, but they are more news mouths for Nvidia, rather than critics and reviewers.

This is an utter crappy card and considering the msrp is really around $450, its one expensive turd.

 

[B-Real]

So this card is exactly a 2070 w/o RT and between a 2070 and 2070S with RT. For $330 that's a pity for me.

 

[ODOGG26]

Why is this "Highly Recommended". This card is total garbage. Seriously trash. Very disappointed.

 

[geon2k2]

It is already out of stock in major retailers around my location and the price was equivalent of 800 to 950 EUR, depending on the model
Lol, there is no need for scalpers anymore legit retails can scalp as well.

 

[cst1992]

Why is this "Highly Recommended". This card is total garbage. Seriously trash. Very disappointed.

If you're specifically looking to buy a 3060, then this is the card you should get.
It's not a class-wide recommendation, rather a model-wide recommendation.

 

[Noztra]

Why is this "Highly Recommended". This card is total garbage. Seriously trash. Very disappointed.

Because its a Nvidia GPU. They always get “highly recommended” on TPU.

 

[cst1992]

I'm getting feelings of deja vu from this.
When I got the 3060Ti, I debated getting this or that.

That reminds me of May 2016 when the Moto G4 Plus was launched. The G4 was cheaper and didn't have the fingerprint scanner that the G4 Plus had, but the G4 was scheduled to launch a few months later.
Like the 3060, the G4 was also anticlimactic.

I got the G4 Plus then and the 3060Ti now. So it was the right choice both times, inspite of being the more expensive one.

Because its a Nvidia GPU. They always get “highly recommended” on TPU.

That's BS.

 

[squallheart]

@W1zzard

I would like to know if the results in the charts from the older cards (such as the 3080/3070, 5700XT) are taken from your old tests results or are they ran again when you conduct this test?

 

[B-Real]

If you're specifically looking to buy a 3060, then this is the card you should get.
It's not a class-wide recommendation, rather a model-wide recommendation.

I don't see the reason to buy the highest quality of coolers for a midrange card with 170W TDP.


HH, what happened to the cool performance per W graphs?

@W1zzard

I would like to know if the results in the charts from the older cards (such as the 3080/3070, 5700XT) are taken from your old tests results or are they ran again when you conduct this test?

If you check those original benchmarks and see the same results in different games as in this one, than it's the old.

 

[Crackong]

My thoughts:

When everything can be found in MSRP:
This product doesn't make sense when there is a 3060Ti with $399 MSRP.

When everything cannot be found in MSRP:
This product doesn't make sense higher than MSRP.

They clearly marked the "MSRP" based on the inflated pricing / card performance of 3060Ti
Stupid product in stupid config with stupid pricing.

 

[watzupken]

As always, MSI presents a hugely oversized PCB, that has no sense whatsoever.

I agree it is overkill for a card at this level, but for a higher end card, it is actually better. The reference design from Nvidia for the RTX 3080 for example is very cram due to a smaller than usual PCB. As a result, all the hot components, GPU, VRAM and VRM are clustered together which makes them run hotter than if you space them out with a bigger PCB. The edge temp shows that they are in the healthy temp range, but the junction temp prove otherwise.

"It looks like the RTX 3060 is only marginally faster at RT than Ampere, seems the new RT cores don't make that much of a difference."
W1zz, what kind of sorcery is that?

It depends on which card you are comparing the RTX 3060 against. While Nvidia claimed that the second gen RT cores are up to 2x faster, you need to remember that it has less RT cores as compared to the likes of the RTX 2070, i.e (28 vs 36 RT cores). Even the RTX 2060 has a higher RT core count at 30.

My thoughts:

When everything can be found in MSRP:
This product doesn't make sense when there is a 3060Ti with $399 MSRP.

When everything cannot be found in MSRP:
This product doesn't make sense higher than MSRP.

They clearly marked the "MSRP" based on the inflated pricing / card performance of 3060Ti
Stupid product in stupid config with stupid pricing.

When everything is at MSRP, I agree that the RTX 3060 Ti is probably the best in value. But that does not mean that there is no place for the RTX 3060. As you go down the product stack, usually the pricing becomes tighter, i.e. USD 399 vs 329. While in the dollar sense, its just 70 bucks, but in a percentage sense, it is quite a steep jump, and not everyone will want to spend that much % more since the RTX 3060 suits their needs anyway.

 

[W1zzard]

@W1zzard

I would like to know if the results in the charts from the older cards (such as the 3080/3070, 5700XT) are taken from your old tests results or are they ran again when you conduct this test?

They are not retested for each review, because I make changes to the test system only every few months. The results are still valid because I'm not patching the OS or games.

Right now I'm setting up my 2021 GPU review bench, which has new games, drivers, etc. and everything will be retested for that

HH, what happened to the cool performance per W graphs?

There's a single chart as "efficiency" now

 

[THU31]

Nice mining advertisement on techpowerup.