Nvidia recently scored a big win with its RTX 4070 Ti graphics card, which has me pretty optimistic about the future of Nvidia’s midrange offerings to come, but a new spec leak from a fairly reliable Twitter leaker has me seriously questioning what the company is thinking.
For starters, it’s important to qualify spec leaks like this since Nvidia hasn’t announced anything yet, and any “leaks” online need to be taken with a grain of salt. But kopite7kimi has been fairly on point in the past, so these specs can’t be written off entirely either.
RTX 4060 still uses PG190. AD107-400-A13072FP328G GDDR6 18Gbps115W24M L2I will try to remain neutral about any leaks in the future. 😁😁😁February 13, 2023
For starters, the RTX 4060 looks like it’ll be using the AD107 GPU, which is a step down from the AD106 that we would expect to see in the RTX 4060, since the RTX 3060 used the GA106 GPU. Even the RTX 3050 8GB used a cut down GA106 GPU, so the AD107 GPU would appear to be a regression here.
What’s more, the purported RTX 4060 specs (as well as some of our back-of-the-napkin calculations) are nearly identical to the RTX 4060 Mobile specs for which we have official numbers for comparison.
As you can see, the biggest difference in the leaked specs from the confirmed RTX 4060 Mobile is the large downgrade in the number of streaming multiprocessors going off the CUDA core count, from 3,584 in the RTX 3060 and 3,072 in the RTX 4060 — a 14.28% decrease. Without the decrease, this falls light-years short of the best graphics card in Nvidia’s lineup, the RTX 4090, and comes in at just north half that of the RTX 4070 Ti, so the decrease in core count here is going to limit the RTX 4060’s potential.
There is also the matter of the slightly higher effective memory clock over the RTX 4060 Mobile, which gives the RTX 4060 desktop a 12.5% increase in memory bandwidth, but otherwise it’s more or less the same on the memory front as its mobile variant.
A key unknown right now is what the final base and boost clock speeds will be for the RTX 4060, but given that the Nvidia Lovelace architecture is a 4nm process versus Nvidia Ampere’s 8nm process, we expect base clock speeds north of 2,000MHz, with the boost clock possibly upwards of 2,500MHz if the roughly 56% faster base clock and 48% faster boost clock gen-on-gen pattern for the RTX 4090 and RTX 4080 hold for the RTX 4060.
Two reasons these specs have me worried
First, let me start with this: the decrease in streaming multiprocessors and CUDA cores isn’t that big of a deal. With 128 CUDA cores per SM, the leaked specs mean a decrease from 28 SMs in the RTX 3060 to 24 SMs in the RTX 4060. A 14.28% drop in tensor cores and ray tracing cores isn’t great, but also consider that these are fourth-gen tensor and third-gen ray tracing cores. They are simply much, much better than Ampere’s third- and second-gen cores, respectively, so they will actually perform better despite there being fewer of them.
What does worry me though is the memory. 8GB VRAM at this stage is rather paltry, and while this technically should be a high-end 1080p graphics card, given how well the RTX 4070 Ti handles 4K and how well the RTX 3060 Ti manages with 1440p gaming, we would hope that the RTX 4060 would be a solid candidate for the best 1440p graphics card. Given its specs though, I’m doubtful.
The problem here is that 8GB VRAM is just fine for 1080p, since the size of texture files that can quickly fill up VRAM are much smaller than they are at 1440p or 4K. And while the RTX 3060 Ti also had 8GB GDDR6 VRAM, it also had a much wider memory bus (256-bit), giving it an effective memory bandwidth of 448.0 GB/s.
That’s more than enough to get 1440p textures efficiently loaded up and processed despite the smaller VRAM pool, while the RTX 3060 had 50% more VRAM (12GB) and a larger memory bus (192-bit), so it clocked in with a 360.0GB/s memory bandwidth, also giving it some decent-ish 1440p performance. The RTX 4060, meanwhile, looks like it will have a much lower memory bandwidth, so larger textures at 1440p are going to bottleneck much faster, limiting 1440p performance and pretty much restricting this card to 1080p gaming.
There’s also the matter of it using the same base GPU as the RTX 4060 Mobile. We haven’t been able to test the RTX 4060 Mobile for ourselves yet, but normally the mobile chips are about a tier to a tier-and-a-half lower than their desktop counterparts in terms of the GPU used (so the RTX 4090 desktop’s AD102 is a tier up from the RTX 4090 Mobile’s AD103), while also being cut down versions of those chips.
The RTX 4060 Mobile is definitely a cut down variant of the AD107, but if the RTX 4060 desktop card is also an AD107, then you definitely can’t expect too much more performance out of the desktop version than you’d get with the mobile GPU. That’s not great for desktop gaming, even at 1080p.
Still, there’s one big reason these specs might be good news for gamers
While all of this might be bad news for gamers in terms of performance, there is a sliver of hope here, and that’s price. If Nvidia was going to try to make a RTX 4060 that is aggressively priced (I’m thinking less than $300/£300), then the trade off here wouldn’t just be justified, it might very well be what gamers are calling for.
Most gamers are still gaming at 1080p, according to the Steam Hardware Survey with a growing number upgrading to 1440p at most. While the memory constraints don’t bode well for 1440p gaming, getting an affordable graphics card into the hands of gamers where they’re currently at would definitely be something to celebrate.
A lot of gamers are also still running GTX era GPUs, so the performance upgrade with an RTX 4060 will still be substantial enough that I’m sure a lot of gamers aren’t going to care about how much better it could be with an AD106 GPU, and in the end, that’s really all that should matter — especially if Nvidia can get the pricing on this card right.