Cute! :love:

Perfect for HTPC setups. Would be a nice upgrade for my 1060 to give DLSS capability and HDMI 2.1 to my eARC receiver and TV.

I am waiting for a low profile version.

If the cooler is anything like that on the previous Phoenix cards, the card will run pretty hot, especially considering that it is even more power-hungry than the last-gen highest-TDP Phoenix (RTX 2060, 160 W). Hopefully this cooler has a bigger fin stack or thicker heat pipes or something.

watzupkenI am waiting for a low profile version.

it already looks like it come from the Gremlin movie. :D

Nice looking card. Looking forward to seeing the "2025 In Stock Edition"...

Awww! Hehe, I just want to cuddle it, the lil' cutie! :D

RithsomIf the cooler is anything like that on the previous Phoenix cards, the card will run pretty hot, especially considering that it is even more power-hungry than the last-gen highest-TDP Phoenix (RTX 2060, 160 W). Hopefully this cooler has a bigger fin stack or thicker heat pipes or something.

The only Phoenix card I owned was a GTX 1050 Ti, and it runs warm for a 75W card with just a basic flower shaped aluminum only heatsink. I was hesitant about buying a GTX 1660 Ti Phoenix because I know it will run very hot and I don't think it came with any heatpipes, just a copper core. So I feel this card is going to run hot as well.

watzupkenI am waiting for a low profile version.

Not going to happen. The 3060 is a 170W TDP design. While it's possible for HHHL/LP GPUs to exceed 75W, it's difficult due to the sheer cooling needs of heat loads above that, plus the fact that in such a small area, even a 6-pin PCIe power connector will remove a significant amount of heatsink area. 100-120W is pretty much the maximum that can feasibly be kept cool in that form factor, and even that would be very loud. Of course there's also the issue that there is no way to physically fit the 3060 package + VRAM onto a low profile PCB. So while it would be really cool for LP GPUs in this performance class, it's still a few years away at the very least.

watzupkenThe only Phoenix card I owned was a GTX 1050 Ti, and it runs warm for a 75W card with just a basic flower shaped aluminum only heatsink. I was hesitant about buying a GTX 1660 Ti Phoenix because I know it will run very hot and I don't think it came with any heatpipes, just a copper core. So I feel this card is going to run hot as well.

This clearly has at least three heatpipes, so it's not that basic a cooler. It will obviously run hotter than something bigger, but that's expected for such a small design. Pretty much perfect for small cases though, and the large fan is a nice touch.

Sweet but if someone realised mali-g78 mp144 or great enchanted adreno 660 will be with same or better performance and much much less power consumption.

TumbleGeorgeSweet but if someone realised mali-g78 mp144 or great enchanted adreno 660 will be with same or better performance and much much less power consumption.

The internet has been saying a lot of shit about the 3060, but that's not a reason to tell lies:

dyonoctisThe internet has been saying a lot of shit about the 3060, but that's not a reason to tell lies:

Do you read all what I am write in comment. For enchanted variants with much more computational units?!?

TumbleGeorgeDo you read all what I am write in comment. For enchanted variants with much more computational units?!?

I assume you mean "enhanced"? Nonetheless, you can't assume performance and efficiency will scale linearly as a chip grows - physics generally tends to get in the way. A bigger chip has more internal losses just due to being bigger, needs higher voltages to run due to this, has more internal variance meaning it typically needs bigger voltage margins for operation, needs a wider memory interface to keep its cores fed, which consumes significant power, etc. Besides, Mali G78 scales to a maximum of 24 cores - it doesn't have the internal connectivity needed to connect more cores than that, and there's no reason to expect its internal architecture to be geared towards effective work distribution across a wider design than 24 cores. In other words, a 144-core design would need significant changes, and in all likelihood sacrifice significant efficiency to get there. Smaller designs can always be more efficient than larger designs. A theoretical 144-core Mali G78-like GPU would also likely be absolutely freaking massive.

Asus have changed the cooling solutions on their Phonenix ITX line of cards, these new ones are better ib both cooling performance and decibels. However, how they pit against MSI`s aero itx versions, I have no information. as reviews on ITX cards are very rare and hard to find.

TurmaniaAsus have changed the cooling solutions on their Phonenix ITX line of cards, these new ones are better ib both cooling performance and decibels. However, how they pit against MSI`s aero itx versions, I have no information. as reviews on ITX cards are very rare and hard to find.

Yeah, sadly they tend not to get seeded to reviewers much - they'd much rather push out their large, flashy high-end SKUs for reviews. Hopefully someone can do an ITX 3060 roundup whenever(/if) supply evens out.

If the cooling solution hasn't evolved for the Phoenix since the 1660, it will be a piece of hot crap, really:

The cooling on this looks reasonable. I see three heatpipes that should be enough for boost clocks and the horizontal heatsink means that at least half of the airflow is exhausted out of the back which is great for cramped ITX builds and halves the recyling of hot air by the cards cooling fan.

watzupkenI am waiting for a low profile version.

You'll be waiting for a while. Even the reference-clocked cards have a TDP of 170W; That's not going to be practical to cool as it would need a bank of screaming 60mm server fans to stand any chance.

The hightest-TDP low-profile cards to date have been 75W, and those were still double-wide. Perhaps a 3050 variant might be under 100W and make a low-profile variant feasible.

To me it looks like a halved/single fan design of their dual fan EVO serie.

For a small card it looks pretty nice imo.

S m o l

ValantarSmaller designs can always be more efficient than larger designs.

Prove that! Where is significant difference between rtx 3060 and rtx 3090 because of efficiency smaller than non efficiency larger design.

TumbleGeorgeProve that! Where is significant difference between rtx 3060 and rtx 3090 because of efficiency smaller than non efficiency larger design.

The 3060 is just a cut down 3090, and consume half as much power . nvidia is first developing a GPU arch with no breaks, where they try to get as much performance as they can from a node, and then they are cutting it down. Meanwhile Smartphone GPU are developed with maximum efficiency in mind, for an environment with lots of constraint. (really small size, can't use GDDR with a wide bus...)

What you are saying imply that both AMD and Nvidia have no idea on how to make GPUs, and ARM/QUALCOMM are light years ahead of them, yet they don't try to bank on that. They are both making SOC for laptop where they have much less constraint, yet we don't see them making chip offering 2060 level of performance even though that would be so easy to do in your opinion.

Look, samsung choosed to get in bed with AMD to get a boost in GPU power rather than boosting Mali, even though RDNA sucks if we follow your reasoning :
Samsung Exynos With AMD RDNA SoC Smashes A14 Bionic in Leaked GPU Benchmark | Tom's Hardware

TumbleGeorgeProve that! Where is significant difference between rtx 3060 and rtx 3090 because of efficiency smaller than non efficiency larger design.

Damn, I wish I placed a bet on getting this response - I was expecting this exact misunderstanding. The 3060 is a smaller implementation of the exact same architecture as the 3090. What I said was

ValantarSmaller designs can always be more efficient than larger designs

So, designs, not implementations. All currently known Ampere is the same design, the same overall architecture, with the same limits and possibilities, just implemented at different sizes. I could of course have said "architectures with more limited scaling potential" rather than "smaller designs", but I was expecting a tad more understanding given that this is a tech enthusiast forum. If, on the other hand, Nvidia made a special variant of Ampere that by design couldn't scale past 3584 shaders or 112 TMUs or 48 ROPs or whatever limitation they chose, they could indeed make such a design more efficient than the current one that can scale higher. How? By designing the internal interconnects, datapaths and layouts to be more optimal up to that size. Given the still large size of that chip, the difference might not be massive, as there's far more to gain in really small (APU/mobile type layouts) - the size differences are simply more meaningful between a 60mm² GPU and a 300mm² one than between a 300mm² one and a 500mm² one. Current PC GPU designs are made around easy scaling and implementation of a wide range of designs that need only minor driver and firmware tweaks between them; more specialized designs can always be more efficient - but the amount of work needed to make them work would also increase drastically.

B-RealIf the cooling solution hasn't evolved for the Phoenix since the 1660, it will be a piece of hot crap, really:

The 1660 is a significantly cheaper and lower end GPU with a much lower TDP (120W vs. 170W) than the 3060 so this won't have that kind of cooler.

dyonoctisAMD and Nvidia have no idea on how to make GPUs, and ARM/QUALCOMM are light years ahead of them

AMD and Nvidia have idea but there is ARM/QUALCOMM which is on light years ahead of them.
PS. GPU CU numbers scaling is not so bad like CPU number of cores scaling because operations for GPU are more simplify and is more easily to parallelized to many CU's.
The variety of tasks assigned to the CPU makes it difficult to scale and reduces efficiency, because the CPU also solves a lot of single-threaded tasks.
But my interest in this discussion is on GPU's not CPU's.

I'd buy that just because it's cute lol.

TumbleGeorgeAMD and Nvidia have idea but there is ARM/QUALCOMM which is on light years ahead of them.
PS. GPU CU numbers scaling is not so bad like CPU number of cores scaling because operations for GPU are more simplify and is more easily to parallelized to many CU's.
The variety of tasks assigned to the CPU makes it difficult to scale and reduces efficiency, because the CPU also solves a lot of single-threaded tasks.
But my interest in this discussion is on GPU's not CPU's.

I don't think the person you quoted mentioned CPUs whatsoever...

And, again, if ARM and Qualcomm could scale their GPUs up to much larger sizes without sacrificing efficiency, why haven't they done so? That would allow them entry into huge and very lucrative markets like consoles, gaming PCs, etc. Of course none of these come close to the sales volumes of smartphones, but smartphones also have near zero margins.

You're assuming they have some kind of magical technology that simply doesn't exist, as you're not taking into account the inherent efficiency that comes from designing for a small maximum size and overall limited layout. Smaller designs will always be more efficient than larger designs. Period. There's nothing saying that any current mobile GPU maker could match AMD or Nvidia at the 150-250W range, except maybe Apple. But given the drastic differences between mobile GPUs in power delivery, size and thus internal interconnects, VRAM interfaces and bus widths, thread/workload allocation, driver complexity, etc., etc., etc., there's no way of knowing until one of them tries.