So a new generation looms and ATi branding is no more so we can't ask ourselves ATi or nVidia again but yet the question remains the same.
Red or Green?
Where do we stand, nVidia GeForce or AMD Radeon?
Or is there a third contender for the crown?
You guys know where I usually stand but I have to admit the Radeon HD7000 series is looking pretty nifty.
Both are shaping up to be absolute power houses so whichever way one goes power seems assured at least this time around.
The AMD 480 should have come with at least an 8pin power connector. Pulling over 75w avg from the PCIe interface is probably not a good idea for sustained gaming or overlocking. Considering that this card is marketed toward the mainstream who generally don't have the best motherboards on the market, it's a bit perplexing how this issue got overlooked.
"We continuously tune our GPUs in order to maximize their performance within their given power envelopes and the speed of the memory interface, which in this case is an unprecedented 8 Gbps for GDDR5. Recently, we identified select scenarios where the tuning of some RX 480 boards was not optimal. Fortunately, we can adjust the GPU's tuning via software in order to resolve this issue. We are already testing a driver that implements a fix, and we will provide an update to the community on our progress on Tuesday (July 5, 2016)."
so in other words make the card a bit weaker to get power usage down probably with throttling
In slightly less than 2 weeks, Nvidia is launching the 1060. Starts at 250$, Nvidia says it's on average 15% faster than the 480. It seems to be only a 6gb version, unless they announce a cheaper 3gb version later. But 250$ for a strong 6gb card is quite good. http://www.anandtech.com/show/10474/nvidia-announces-geforce-gtx-1060-july-19
Larger increase at lower resolutions as cpu bottlenecks reduced compared to high end. Nvidia has less of a cpu overhead. They also cannot do asynchronous compute but fake it.
Vega will destroy.
I have a 1080 but pls give Vega now. 4k no compromise
It's a mixed bag for nvidia users from different tests, but overall high gains for GCN users on every test.
I don't have the game, ive only tested the Demo which only has OpenGL and already running at around 100+FPS at Nightmare settings TSSAA at 1440p for me
Reading through Tiago Sousa's twitter (lead renderer programmer at id software) gives me a lot of hope about the future of gaming.
The effects of this Vulkan update, the general positive reaction from the community, the continued updates to the game shows that putting in the effort when possible leads to a product that resonates with the consumers.
Ashes of the Singularity has more or less become the default go to for DX12 benchmarking (yuck) and Rise of the Tomb Raider becoming better with updates, nothing compares to the efforts put forward by the guys at id.
FFS a basic bitch 480 with all its problems is smoking Doom. One of the cheaper cards is handling this game amazingly and I can't wait to see how the 1060 handles Vulkan.
It's also a big slap in the face to AMD for having such poor implementation of DX11 software.
Is it really slap in the face of AMD? They have a bit of overhead which reduces FPS in CPU-heavy games, but the 970 and 390 released at the same price (330$) and perform similarly in DX11 games. An average of multiple benchmarks had the 390 at 1% faster than the 970.
I mostly see this as vindication for an AMD employee who, about 2 years before the release of Mantle I think, said publicly that Directx could be improved. He described what AMD would eventually do with Mantle, and the worst part is that three reputed engine developers talked on twitter and in interviews, saying he was right: John Carmack, Repi from DICE, and some dude from Crytek (I mean, they do know how engines work).
Despite this, that AMD dude was made fun of by tech websites and forums, saying "lolololol AMD doesn't know anything about software and they dare insult Directx?".
Turns out, they were right and now Vulkan and DX12 are built on the foundations of Mantle.
Poor dude. He was right but probably talked about it earlier than he should have.
So, it seems that AMD cards are gaining around 11% in performance with Async in that specific benchmark, Nvidia 10x0 cards are gaining around 6%, and their 9x0 cards aren't gaining anything at all:
Since I don't know how to properly format it in the forum, the important numbers:
480 vs 1060
DX11 (35 games) 88% vs 100%
DX12 (5 games) 102% vs 100%
Doom OpenGL 86% vs 100%
Doom Vulkan 121% vs 100% (goddamn)
Average of every games (37 games) 91.5% vs 100%
So, the 480 is the better purchase for the new APIs, but the 1060 is the better purchase overall today.
So it seems Pascal will always benefit less from async compute compared to AMD not because it isn't capable of doing it, but more because of other architectural differences. Another way to look at it is: Pascal gets closer to it's theoretical peak performance already with Dx11 so it will benefit less from async compute in Dx12. It also seems performance increases from async compute for Nvidia Pascal will need to be more fine tuned to really benefit from it.
As long as I can continue to hit 1080p60 in new games with my 970 at medium+ settings I won't be upgrading any time soon so yeah. I already got a new monitor 2 years ago so that is not getting replaced for another 5+ years at this point in time.
Well, they dropped the GeForce GTX name for this one so its just nVidia Titan X
Status
Not open for further replies.
You have insufficient privileges to reply here.
Related Threads
?
?
?
?
?
Next Generation Emulation
2.2M posts
459.8K members
Since 2001
A forum community dedicated to all emulation enthusiasts. Come Join discussion on all platforms from Nintendo, Microsoft Xbox, Sony Playstation, to PC. Coding, tips, builds, specs, tricks and more.