Nvidia Fires Shots at AMD
With the release of AMDs new 7mn node Navi architecture this Summer, many are hoping to see them pull a little closer to Nvidia in terms of performance. While, incidentally, keeping a (hopefully) more friendly price tag.
In a report via PCGamesN, however, Nvidia CEO Jensen Huang has fired some pretty hefty shots at AMD stating that a direct comparison between their 12nm design (seen in their 20XX and 16XX series) is “incomparable” to AMD’s 7nm design. This isn’t, for a matter of clarification, him saying that AMD is better either.
What Does He Mean?
Well, put simply, he’s talking about power and temperatures and, in fairness, he does have a point. With the release of the Radeon VII graphics card, while it was a very solid performer, it wasn’t comparable at all to Nvidia in terms of power consumption, noise levels or temperature control.
“If you take our Turing and you compare it against a 7nm GPU on energy efficiency, it’s incomparable. In fact, the world’s first 7nm GPU already exists, and it’s easy to go and pull that and compare the performance and energy efficiency against one of our current GPUs.” – Jensen Huang
While he doesn’t name the Radeon VII specifically, given that it’s the only 7nm GPU currently on the market, we all know what he’s referring to here.
Efficiency
Nvidia has already made it very clear that they have no immediate plans to jump into 7nm design just for the sake of it. A factor they have regularly backed up by claiming (in round about ways) that their 12nm design will always be better than anything AMD could do with 7nm.
While we do harbour hopes that the upcoming AMD Navi graphics card range will be decent, we do suspect that Jensen Huang might ultimately be proven right here.
What do you think? – Let us know in the comments!
Yes this is true. I can afford AMD products though so efficiency will not win me over. Coal burning power plants are probably more efficient than a solar grid, doesn’t mean the future is in fossil fuels (the clue is in the name lol). I remember intel making similar boasts, turns out their chip design is full of security holes. I wonder if Nvidia could suffer similar problems perhaps in CUDA processing tasks. I’m going to stick to AMD for the foreseeable future. I don’t buy into Nvidia’s hardware accelerated raytracing only being possible on their GPUs. It’s been demonstrated that a Vega 56/64 is capable of doing some decent RT. I’m also happy to support the company who’s knocking it out of the park with their CPU game atm.
Nvidia are scumbags, last one i brought was a geforce 2, since then its been Ati/amd and always will be there just way to much money for little extra performance… Amd is the smart choice for the smart person, oh and the drivers smash nvidia’s doors in.
Efficiency is easy to complain about when you completely disregard the fact that RvII has massive compute performance.
Compute hw is much more power hungry by itself.
Also, when one takes into account that despite being touted as 295w tdp gpu, RvII actually pulls less power in games (its comparable to 2080 in that regard, and it can be undervolted easily to surpass 2080 in efficiency).
NV can get away with seemingly better efficiency because they can afford to make separate GPUs… AMD on the other hand makes 1 gpu template and then makes both gaming and compute powerhouses from that.
Note that RVII also has INCREASED compute capabilities vs Vega64 (specifically in fp64).
I’d like to see NV having as much compute prowess and say they are better at efficiency.
Gcn is not a bad design… Its just more compute oriented.
This isn’t really accurate Deks. In terms of compute performance the RTX 2080 vs Radeon VII… the 2080 generally wins significantly in several compute tests or comes close in others. In a few the VII noticeably outperforms the 2080 reversing the trend in limited tests. Then the 2080 Ti exists… Thing is, the main addition to the Turing series was RTX and Tensor cores without shrinking and a bit of a boost in other performance. Significant hardware additions and despite this in both compute, power efficiency, and gaming performance the VII is totally struggling to keep up often either drastically underperforming, barely cutting even, and in some rare cases excelling noticeably. This is already bad but it gets worse when you put 7nm into perspective because that is a MASSIVE shrink vs 12nm and it can’t go much further with current technological limits atm. Nvidia is basically playing the market because they are confident they can still win readily at 12nm. If they shrank to 7nm, or at all really, AMD would no longer even be in the game as a competitor. They would be totally out without a massive architecture overhaul that is leaps and bounds better than their current. If AMD doesn’t make some major advances in the next 2-3 years the GPU market could be in serious trouble due to no longer having a competitor for Nvidia.