Nvidia recently scored a huge win with its RTX XNUMX Ti graphics card, which makes me pretty optimistic about the future of Nvidia's mid-range offerings to come, but a new specs leak from a fairly reliable Twitter leaker has me makes you seriously question what the company thinks. .
Rating spec leaks like this is essential to get started, as Nvidia hasn't announced anything yet, and any online "leaks" should be taken with a grain of salt. But kopite7kimi (opens in a new tab) has been pretty relevant before, so those specs can't be completely undone either.
RTX 4060 always uses PG190. AD107-13072-A328FP6G GDDR18 115Gbps24W2M L13I will endeavor to remain neutral to any future leaks. 😁😁😁February 2023, XNUMX
Read more
For starters, the RTX 4060 appears to be using the AD107 GPU, which is a step up from the AD106 that we'd expected to see in the RTX 4060, since the RTX 106 used the GA106 GPU. Even the 107GB RTX XNUMX used a stripped down GAXNUMX GPU, so the ADXNUMX GPU feels like a throwback here.
Also, the RTX 4060's purported specs (as some of our back-of-napkin calculations are) are virtually identical to the RTX 4060 Mobile specs for which we have official figures for comparison purposes.
Slide to scroll horizontally RTX Spec Comparison: 4060 Mobile / 4060 / 4060 Component RTX 4060 Mobile (Official) RTX XNUMX Desktop (Rumored) RTX XNUMX Desktop Speed XNUMX Gbps XNUMX Gbps XNUMX Gbps Memory Bus Width XNUMX bits one hundred twenty eight bits one hundred ninety two bits Memory Bandwidth two hundred fifty six GB/s two hundred eighty eight GB/s three hundred sixty GB/sTDP one hundred fifteen W one hundred fifteen W one hundred seventy WAs you can see, the biggest difference in the leaked specs from the confirmed RTX 4060 Mobile is the sharp drop in the number of streaming multiprocessors springing from the CUDA core count, from 4060 on the RTX 4060. one thousand seventy-two on the RTX XNUMX, a decrease of fourteen with twenty-eight%. Without the drop, it's light-years ahead of the best graphics card in Nvidia's lineup, the RTX XNUMX, and just north of the middle of the RTX XNUMX Ti, so the drop in number of cores here will limit the potential of the RTX XNUMX.
There's also the matter of the slightly higher effective memory clock compared to the RTX 4060 Mobile, which gives the RTX 4060 desktop a twelve.5% increase in memory bandwidth, but is otherwise roughly same on the memory front as its mobile variation. .
One unknown key right now is what the final base and boost clock speeds will be for the RTX 4060, but since the Nvidia Lovelace architecture is a 2000nm process vs. Nvidia Ampere's 4060nm process, we expect speeds. base clock north of XNUMXMHz, with the boost clock likely to be over XNUMX MHz if the base clock is about XNUMX% faster and the gen model boost clock is XNUMX% faster. XNUMX% faster for RTX XNUMX and RTX XNUMX are valid for RTX XNUMX.
Two reasons why these specs concern me
First, let me start with this: the throttling of multiprocessors and CUDA cores isn't too bad. With 4060 CUDA cores per SM, the leaked specs mean a drop from XNUMX SM on the RTX XNUMX to XNUMX SM on the RTX XNUMX. A XNUMX% drop in Tensor Cores and Ray Tracing Cores isn't great, but Also consider that these are XNUMXth generation Tensor and XNUMXrd generation ray tracing cores. They're just so much better than Ampere's XNUMXrd and XNUMXnd generation cores, respectively, so they're going to perform better even if there's less of it.
What worries me is the memory. eight GB of VRAM at this point is pretty inconsequential, and while technically it would have to be a high-end 1080p graphics card, given how well the RTX 4 Ti handles and how well the RTX 1440 Ti handles With 4060p gaming, we'd expect the RTX 1440 to be a strong contender for the best XNUMXp graphics card. Given its peculiarities, I have my doubts.
The downside here is that 1080 GB of VRAM is fine for 1440p, because the size of the texture files that can quickly clutter up the VRAM is much smaller than in 4p or 6K. And while the RTX XNUMX Ti also had XNUMX GB of GDDRXNUMX VRAM, it also had a much larger memory bus (XNUMX bits), giving it an effective memory bandwidth of XNUMX with zero GB/s.
This is more than enough for 1440p textures to load and render efficiently despite the smaller VRAM pool, whereas the RTX 0 had fifty percent more VRAM (1440 GB) and a bus largest memory (one hundred and ninety-two bits), so it registered three hundred and sixty.4060 GB. /s of memory bandwidth, which also gives you passable 1440p performance. The RTX 1440, on the other hand, seems to have much lower memory bandwidth, so larger textures at 1080p will crash much faster, limiting performance to XNUMXp and limiting this card to XNUMXp gaming. .
There's also the downside of using the exact same base GPU as the RTX 4060 Mobile. We haven't been able to test the RTX 4060 Mobile for ourselves yet, but overall mobile chips are between a level and a level and a half below their desktop counterparts in terms of GPU used (so the RTX 102, the The desktop AD103 is a step up from the RTX XNUMX Mobile ADXNUMX, while also being stripped-down versions of those chips.
The RTX 4060 Mobile is clearly a scaled-down variation of the AD107, but if your RTX 4060 desktop card is also an AD107, you certainly can't expect much more performance from the desktop version than you'd get on mobile. GPU. It's not great for desktop gaming, even at 1080p.
Still, there's a huge reason these specs could be good news for gamers.
While all of this may be bad news for players in terms of performance, there is a silver lining here, and that is cost. If Nvidia were trying to make a violently expensive RTX 4060 (less than $XNUMX / $XNUMX, I think), then the tradeoff here would not only be justified, it might as well be what gamers call it.
The majority of gamers are still playing at 1080p, according to the Steam Hardware Survey (opens in a new tab) with an increasing number of gamers switching to a maximum of 1440p. While memory constraints don't bode well for 1440p gaming, putting an affordable graphics card in gamers' hands where they are today would certainly be something to celebrate.
Many gamers are still using GTX-era GPUs as well, so the performance improvement with an RTX 4060 will still be significant enough to be sure many gamers won't care how bad it could be better with an AD106 GPU, And in the end, that's really all that should matter, especially if Nvidia can get the right price on this card.