AMD's Crimson ReLive Edition driver update brings a host of gaming improvements and new features

[midian182]

It’s been a busy week for AMD. Tuesday brought reports that Intel has signed a deal with the company to license its GPU tech, and today sees the release of a major update for its Radeon Crimson drivers: Crimson ReLive Edition (click to download).

The update, which AMD is calling its biggest software release ever, will improve gaming, stability, and VR experiences for owners of Radeon and Radeon Pro graphics cards, as well as introducing a number of new features.

When it comes to gaming, AMD shows how Crimson ReLive Edition can improve performance by between 4 and 8 percent using a GPU such as the RX 480. The chart below shows the difference when using the new update compared to June’s Crimson release.

Crimson ReLive brings support for both HDR 10 and Dolby Vision which, for owners of compatible displays, brings gaming and video content to another level with increased brightness, contrast, and colors. Moreover, AMD notes how gaming is stepping into the “Deep Pixel Era,” with 8K resolutions and the wide number of FreeSync Monitors available.

There’s also a handy new feature that automatically detects bad HDMI cables. Should one be detected, an algorithm will move through other resolutions and refresh rates until it can find a supported configuration, at which point it will notify the user.

Staying in the area of high resolutions, Crimson ReLive introduces new VP9 decode acceleration to enable fluid 4K 60Hz GPU-accelerated video streaming. There’s also FreeSync improvements with up to 24 percent lower click-to-respond time and a borderless fullscreen mode, along with support for DisplayPort HBR3 to allow single-link 4K/120 Hz, 5K/60 Hz and 8K/30 Hz rates.

One especially interesting new feature is Radeon Chill, which AMD claims can reduce a GPU’s power consumption, heat production, temperatures, and cooling noise without compromising a game’s performance. It does this by analyzing what’s happening on a screen; if things are fairly static, Chill will scale down frame rates to save power, increasing them only when the action starts up again. It can also save power by stopping games from running at excessively high frame rates that may not offer any benefits to the gamer, and has the potential to decrease the amount of input-to-response lag.

In what appears to be a response to Nvidia’s Share feature, Radeon ReLive Edition lets you capture, stream, and share your favorite gaming moments. The repositionable overlay toolbar allows instant replays from up to an hour earlier; 1-click record, stream, and screenshot functions; and, naturally, supports all the major streaming platforms.

AMD is also improving Virtual Reality experiences with its LiquidVR technology, including optimization for multi-GPU setups and rendering, along with what it calls “truly immersive audio.”

There’s good news for users of Radeon Pro GPUs. Crimson ReLive Edition brings YoY performance increases of up to 30 percent. AMD, meanwhile, says it is committed to quarterly updates for enterprise.

With this latest release, it looks as if AMD is going all-out in its battle against Nvidia. If you own an AMD GPU, make sure you download Crimson ReLive now to get the most out of your card.

Permalink to story.

https://www.techspot.com/news/67325-amd-crimson-relive-edition-driver-update-brings-host.html

 

[Cycloid Torus]

Not for older AMD cards... only HD 7700 and up & HD 8500 and up.

 

[Steve]

Not for older AMD cards... only HD 7700 and up & HD 8500 and up.

That's fair enough, I never really expected them to go back and support TeraScale. How many gamers are really using those cards and I can't imagine systems using such an old GPU could take advantage of the ReLive recording feature anyway.

 

[oldikes]

Not for older AMD cards... only HD 7700 and up & HD 8500 and up.

I mean the 7700 series is almost 5 years old now

 

[Puiu]

The benchmarks I've seen with the new driver seem to put the RX480 at the same performance as the GTX 1060 for DX11 and at around 6% better for DX12/Vulkan. The GTX 1060 had a big lead in the DX11 titles at launch (over 10%).

 

[NightAntilli]

The benchmarks I've seen with the new driver seem to put the RX480 at the same performance as the GTX 1060 for DX11 and at around 6% better for DX12/Vulkan. The GTX 1060 had a big lead in the DX11 titles at launch (over 10%).

If you're talking about the HardwareCanucks benchmarks, that was before the ReLive Drivers. The ReLive drivers boosts the performance of the RX 480 with another 2% - 3% on average if we can believe TechPowerUp. In some games even 7%...;
https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/6.html

 

[Bluescreendeath]

The benchmarks I've seen with the new driver seem to put the RX480 at the same performance as the GTX 1060 for DX11 and at around 6% better for DX12/Vulkan. The GTX 1060 had a big lead in the DX11 titles at launch (over 10%).

The benchmarks I've seen from late October still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card.

https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

 

[Puiu]

If you're talking about the HardwareCanucks benchmarks, that was before the ReLive Drivers. The ReLive drivers boosts the performance of the RX 480 with another 2% - 3% on average if we can believe TechPowerUp. In some games even 7%...;

oh yeah, you are right, I didn't notice. thanks for pointing it out.

The benchmarks I've seen from late October still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card.

https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

this is before the ReLive (5th of december): http://www.hardwarecanucks.com/foru.../73945-gtx-1060-vs-rx-480-updated-review.html

this is for ReLife (tnx NightAntilli): https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/

TL;DR the cards trade blows in different games, but the RX 480 is now doing better.

 

[Bluescreendeath]

The benchmarks I've seen with the new driver seem to put the RX480 at the same performance as the GTX 1060 for DX11 and at around 6% better for DX12/Vulkan. The GTX 1060 had a big lead in the DX11 titles at launch (over 10%).

The benchmarks I've seen from late October (Techspot) to early December (Techpowerup) still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and slightly beating/equaling the RX480 in other DX12 games such as Ashes of Singularity. The 1060 loses in some DX12 games such as Hitman or Deus Ex, and loses in Doom when running Vulkan. So even in DX12 game benchmarks, the RX480 is not automatically superior against the GTX1060 since it wins some, loses some, and ties some. Let's not forget that you need Windows 10 to run DX12, and half of gamers are still using Windows 7 according to steam surveys. Thus, the GTX1060 is still the superior card in the majority of situations.


Benchmarks:
https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

https://www.techpowerup.com/reviews/ASUS/GTX_1050_Ti_Strix_OC/29.html

 

[Puiu]

The benchmarks I've seen from late October (Techspot) to early December (Techpowerup) still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and slightly beating/equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card in the vast majority of situations - especially since you need Windows 10 to run DX12, and half of gamers are still using Windows 7 according to steam surveys.

Benchmarks:
https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

https://www.techpowerup.com/reviews/ASUS/GTX_1050_Ti_Strix_OC/29.html

I gave you the links. I don't think you understood what "trade blows" meant. Some games run better on AMD hardware some on Nvidia. You have a huge number of tests in the hardwarecanucks benchmark. Deus Ex MD, BF1, Witcher 3 are all big games where AMD wins (aside the ones you mentioned). GTA 5, Overwatch work best on Nvidia hardware. And it also shows the performance of the factory OC oem cards (Witcher 3 is a good example where the OEM matters).

the links I gave you are the latests benchmarks. the game specific driver updates are very important for both companies.

just buy which one is cheaper and if the cards are the same price I recommend to just go for the AMD one since it has more VRAM, better DX12/Vulkan performance, supports cheaper monitors with Freesync and no gameworks mambo jambo. most people don't upgrade the GPU after just a year or 2.

 

[ZackL04]

Those bar graphs some of the most misleading, purposely poorly scaled Ive ever seen. They are designed to show 75-100% increases but then labeled as 4-7% with an asterisk. Nice try

 

[NightAntilli]

Those bar graphs some of the most misleading, purposely poorly scaled Ive ever seen. They are designed to show 75-100% increases but then labeled as 4-7% with an asterisk. Nice try

All companies do this all the time.... Example from the opposite team;
http://i66.tinypic.com/oum8lk.jpg

 

[Puiu]

Those bar graphs some of the most misleading, purposely poorly scaled Ive ever seen. They are designed to show 75-100% increases but then labeled as 4-7% with an asterisk. Nice try

what are you talking about? I've seen graphs that don't even show the percentages on it. here you can't be misled since you have big numbers in front of you. (and the asterisk just says that the tests were done with the pre-release driver, not the final one)

these look to be presentation slides. if you've ever done something like this you'll know that you will always exaggerate the sizes when you have small percentages so that the differences can be seen clearly. it just doesn't look good otherwise.

 

[Cycloid Torus]

I mean the 7700 series is almost 5 years old now

Yup, most of what I have is over 5 years old... Office of the Budget and all that. Makes it hard to play Fallout 4.

 

[bandit8623]

I mean the 7700 series is almost 5 years old now

Yup, most of what I have is over 5 years old... Office of the Budget and all that. Makes it hard to play Fallout 4.

I have hd6970 and it easily plays all games. Just no driver support. Nvidia supports their cards way further back. Amd abandond their customers

 

[Bluescreendeath]

The benchmarks I've seen from late October (Techspot) to early December (Techpowerup) still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and slightly beating/equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card in the vast majority of situations - especially since you need Windows 10 to run DX12, and half of gamers are still using Windows 7 according to steam surveys.

Benchmarks:
https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

https://www.techpowerup.com/reviews/ASUS/GTX_1050_Ti_Strix_OC/29.html

I gave you the links. I don't think you understood what "trade blows" meant. Some games run better on AMD hardware some on Nvidia. You have a huge number of tests in the hardwarecanucks benchmark. Deus Ex MD, BF1, Witcher 3 are all big games where AMD wins (aside the ones you mentioned). GTA 5, Overwatch work best on Nvidia hardware. And it also shows the performance of the factory OC oem cards (Witcher 3 is a good example where the OEM matters).

the links I gave you are the latests benchmarks. the game specific driver updates are very important for both companies.

just buy which one is cheaper and if the cards are the same price I recommend to just go for the AMD one since it has more VRAM, better DX12/Vulkan performance, supports cheaper monitors with Freesync and no gameworks mambo jambo. most people don't upgrade the GPU after just a year or 2.

I think you have your benchmarks confused. The GTX1060 6GB wins by a sizeable margin over the RX480 8GB in Witcher 3 and wins by a narrow margin in Battlefield 1. And this is with mid-November drivers. https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1060_Mini_3_GB/25.html

Even after this AMD driver improvement, the GTX1060 6GB will still be winning in Witcher 3 and will probably lose by a small margin in Battlefield 1.

According to your Relife Techpowerup benchmarks, the fps gain in Witcher 3 is a negligible 0.5%, which is negligible because the Techpowerup benchmarks show the GTX1060 6GB with a ~10-12% fps advantage. The larger fps gain of 3-4% are in games that AMD RX400s were losing in larger margins, with the GTX1060s winning by much larger than 3-4% fps. https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/4.html

If they're the same price, people should go for the GTX1060 6GB over the RX480 8GB.
6GB vs 8GB is not much difference, and they have very different architectures and likely use memory different so it's probably not directly comparable. Nvidia's Pascal architecture is much more efficient, while AMD's RX400 is the equivalent of Maxwell architecture in terms of performance/watts. The GTX1060 6GB runs cooler, is less noisy, has lots of small size single-fan versions, requires a less powerful PSU, and has more overclocking headroom in terms of % clock speed increase and % fps gain. Most RX480s have very small overclocking headroom as the stock speed is already near the maximum stable clockspeed, and the RX480 power draw goes up exponentially when overclocking.

 

[Bluescreendeath]

From the Techpowerup examination of the new Relife drivers, it seems the majority of recent games only gain ~1% fps or less. The majority of the larger fps gains seem to be in much older games (BF3, BF4), or games that came out as a buggy/unoptimized console port (eg. Batman games).

what are you talking about? I've seen graphs that don't even show the percentages on it. here you can't be misled since you have big numbers in front of you. (and the asterisk just says that the tests were done with the pre-release driver, not the final one) these look to be presentation slides. if you've ever done something like this you'll know that you will always exaggerate the sizes when you have small percentages so that the differences can be seen clearly. it just doesn't look good otherwise.

Just because these graphs aren't nearly as misleading as the terrible AMD Zen performance graphs that show double the bar height with no percentages, doesn't mean these graphs are not misleading. They're just "less" misleading than the blatantly misleading graphs without percentages. They are meant to be misleading because the graphs purposely distort the scale. So for folks who don't read the numbers or the fine print (a lot of folks), then they will be misled.

...it just doesn't look good otherwise.

Yes, that's why it is a marketing trick to mislead people into thinking there is a bigger difference than there actually is.

 

[Nobina]

The benchmarks I've seen from late October (Techspot) to early December (Techpowerup) still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and slightly beating/equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card in the vast majority of situations - especially since you need Windows 10 to run DX12, and half of gamers are still using Windows 7 according to steam surveys.

Benchmarks:
https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

https://www.techpowerup.com/reviews/ASUS/GTX_1050_Ti_Strix_OC/29.html

I gave you the links. I don't think you understood what "trade blows" meant. Some games run better on AMD hardware some on Nvidia. You have a huge number of tests in the hardwarecanucks benchmark. Deus Ex MD, BF1, Witcher 3 are all big games where AMD wins (aside the ones you mentioned). GTA 5, Overwatch work best on Nvidia hardware. And it also shows the performance of the factory OC oem cards (Witcher 3 is a good example where the OEM matters).

the links I gave you are the latests benchmarks. the game specific driver updates are very important for both companies.

just buy which one is cheaper and if the cards are the same price I recommend to just go for the AMD one since it has more VRAM, better DX12/Vulkan performance, supports cheaper monitors with Freesync and no gameworks mambo jambo. most people don't upgrade the GPU after just a year or 2.

I think you have your benchmarks confused. The GTX1060 6GB wins by a sizeable margin over the RX480 8GB in Witcher 3 and wins by a narrow margin in Battlefield 1. And this is with mid-November drivers. https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1060_Mini_3_GB/25.html

Even after this AMD driver improvement, the GTX1060 6GB will still be winning in Witcher 3 and will probably lose by a small margin in Battlefield 1.

According to your Relife Techpowerup benchmarks, the fps gain in Witcher 3 is a negligible 0.5%, which is negligible because the Techpowerup benchmarks show the GTX1060 6GB with a ~10-12% fps advantage. The larger fps gain of 3-4% are in games that AMD RX400s were losing in larger margins, with the GTX1060s winning by much larger than 3-4% fps. https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/4.html

If they're the same price, people should go for the GTX1060 6GB over the RX480 8GB.
6GB vs 8GB is not much difference, and they have very different architectures and likely use memory different so it's probably not directly comparable. Nvidia's Pascal architecture is much more efficient, while AMD's RX400 is the equivalent of Maxwell architecture in terms of performance/watts. The GTX1060 6GB runs cooler, is less noisy, has lots of small size single-fan versions, requires a less powerful PSU, and has more overclocking headroom in terms of % clock speed increase and % fps gain. Most RX480s have very small overclocking headroom as the stock speed is already near the maximum stable clockspeed, and the RX480 power draw goes up exponentially when overclocking.

I dislike that if I bought an RX480 I would need a new PSU so it could be said it's more expensive for me than Nvidias card. But Dx12 and W10 are the future and AMD is winning there since they actually improve performance over the years. As I've heard many times, Nvidia gimps their older GPUs to favor the new ones. I can't really say those are facts but I saw some pretty convincing stuff.

 

[JetFixxxer]

Can't overclock my cards since I installed these.

 

[Flebbert]

The benchmarks I've seen from late October (Techspot) to early December (Techpowerup) still shows the GTX1060 6GB winning over the RX480 8GBs in almost every DX11 game, and even beating the RX480 in DX12 games such as Rise of the Tomb Raider and Gears of War 4, and slightly beating/equaling the RX480 in other DX12 games such as Ashes of Singularity. It loses in some DX12 games such as Hitman, and loses in Doom when running Vulkan. The GTX1060 is still the superior card in the vast majority of situations - especially since you need Windows 10 to run DX12, and half of gamers are still using Windows 7 according to steam surveys.

Benchmarks:
https://www.techspot.com/review/1269-nvidia-geforce-gtx-1050/page2.html

https://www.techpowerup.com/reviews/ASUS/GTX_1050_Ti_Strix_OC/29.html

I gave you the links. I don't think you understood what "trade blows" meant. Some games run better on AMD hardware some on Nvidia. You have a huge number of tests in the hardwarecanucks benchmark. Deus Ex MD, BF1, Witcher 3 are all big games where AMD wins (aside the ones you mentioned). GTA 5, Overwatch work best on Nvidia hardware. And it also shows the performance of the factory OC oem cards (Witcher 3 is a good example where the OEM matters).

the links I gave you are the latests benchmarks. the game specific driver updates are very important for both companies.

just buy which one is cheaper and if the cards are the same price I recommend to just go for the AMD one since it has more VRAM, better DX12/Vulkan performance, supports cheaper monitors with Freesync and no gameworks mambo jambo. most people don't upgrade the GPU after just a year or 2.

I think you have your benchmarks confused. The GTX1060 6GB wins by a sizeable margin over the RX480 8GB in Witcher 3 and wins by a narrow margin in Battlefield 1. And this is with mid-November drivers. https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1060_Mini_3_GB/25.html

Even after this AMD driver improvement, the GTX1060 6GB will still be winning in Witcher 3 and will probably lose by a small margin in Battlefield 1.

According to your Relife Techpowerup benchmarks, the fps gain in Witcher 3 is a negligible 0.5%, which is negligible because the Techpowerup benchmarks show the GTX1060 6GB with a ~10-12% fps advantage. The larger fps gain of 3-4% are in games that AMD RX400s were losing in larger margins, with the GTX1060s winning by much larger than 3-4% fps. https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/4.html

If they're the same price, people should go for the GTX1060 6GB over the RX480 8GB.
6GB vs 8GB is not much difference, and they have very different architectures and likely use memory different so it's probably not directly comparable. Nvidia's Pascal architecture is much more efficient, while AMD's RX400 is the equivalent of Maxwell architecture in terms of performance/watts. The GTX1060 6GB runs cooler, is less noisy, has lots of small size single-fan versions, requires a less powerful PSU, and has more overclocking headroom in terms of % clock speed increase and % fps gain. Most RX480s have very small overclocking headroom as the stock speed is already near the maximum stable clockspeed, and the RX480 power draw goes up exponentially when overclocking.

I dislike that if I bought an RX480 I would need a new PSU so it could be said it's more expensive for me than Nvidias card. But Dx12 and W10 are the future and AMD is winning there since they actually improve performance over the years. As I've heard many times, Nvidia gimps their older GPUs to favor the new ones. I can't really say those are facts but I saw some pretty convincing stuff.

Why would you need a new PSU for the RX480 and not for the 1060?

 

[Bluescreendeath]

I dislike that if I bought an RX480 I would need a new PSU so it could be said it's more expensive for me than Nvidias card. But Dx12 and W10 are the future and AMD is winning there since they actually improve performance over the years. As I've heard many times, Nvidia gimps their older GPUs to favor the new ones. I can't really say those are facts but I saw some pretty convincing stuff.

Eh, it varies. About 2 years ago, I saw my GTX670 exceed the AMD HD7970 in performance in benchmarks of GTX900 cards and R9 300s thanks to new driver updates. When I first got my AMD HD7970, the GTX670 performed roughly between a 7950 and a 7970, learning more towards the weaker 7950.

 

[Bluescreendeath]

Why would you need a new PSU
The for the RX480 and not for the 1060?

Virtually all models of the GTX1060s use 30 watts less than the reference RX480s, and ~60+watts less than the factory overclocked RX480s.

http://www.hardocp.com/article/2016...0_g1_gaming_vs_msi_gtx_1060_x/16#.WEoiCLIrJpg

All GTX1060s consume a good deal less than 150watts (usually around 120watts at max load), so all models only need 1x PCIe 6pin. The RX480s that only need 1x PCIe 6pin are the reference versions that run hot, noisy, thermal throttle, and have minimal overclocking headroom - and they go right up to or slightly exceed the 150watt limit. Whereas the good models of the RX480 that don't run hot and noisy and thermal throttle are the aftermarket versions that have 1x PCIe 8pin or 2x PCIe 6pin because they consume a good deal over 150watts (some versions going over 200watts).

 

[Flebbert]

Why would you need a new PSU
The for the RX480 and not for the 1060?

Virtually all models of the GTX1060s use 30 watts less than the reference RX480s, and ~60+watts less than the factory overclocked RX480s.

http://www.hardocp.com/article/2016...0_g1_gaming_vs_msi_gtx_1060_x/16#.WEoiCLIrJpg

All GTX1060s consume a good deal less than 150watts (usually around 120watts at max load), so all models only need 1x PCIe 6pin. The RX480s that only need 1x PCIe 6pin are the reference versions that run hot, noisy, thermal throttle, and have minimal overclocking headroom - and they go right up to or slightly exceed the 150watt limit. Whereas the good models of the RX480 that don't run hot and noisy and thermal throttle are the aftermarket versions that have 1x PCIe 8pin or 2x PCIe 6pin because they consume a good deal over 150watts (some versions going over 200watts).

Sorry if I am miss reading but this reads as RX480 hating rather than anything else. Most of the 1060's require an 8pin connector and only the 3GB have the 6 pin power connector. Same as the 4GB RX480 have the 6 and the 8GB have the 8 pin.

If you have a look at TDP they are pretty much the same and real world power draw are within 10% of each other.

If you have to change the PSU for the RX480 you will have to change it for the 1060 aswell. Better yet replace it anyway if you are that close to your PSU limit.

FYI I have both cards and they perform almost identically.

 

[Bluescreendeath]

Sorry if I am miss reading but this reads as RX480 hating rather than anything else. Most of the 1060's require an 8pin connector and only the 3GB have the 6 pin power connector. Same as the 4GB RX480 have the 6 and the 8GB have the 8 pin.
If you have a look at TDP they are pretty much the same and real world power draw are within 10% of each other.
If you have to change the PSU for the RX480 you will have to change it for the 1060 aswell. Better yet replace it anyway if you are that close to your PSU limit.
FYI I have both cards and they perform almost identically.

You are incorrect. Most GTX1060s, even factory OCed ones, have a single 6 pin PCIe connector. The reference Founder's edition and every other non-overclocked model uses 1x 6 pin. The other factory overclocked ones like the EVGA Superclocked, Zotac Amp! Editiion, and Gigabyte Windorce all use 1x 6 pin.
http://www.techporn.ph/zotac-geforce-gtx-1060-amp-edition-review/
http://www.gigabyte.com/products/product-page.aspx?pid=6088#kf

The MSI Gaming X and Gigabyte G1s have the 8 pin, but those are the exceptions, not the rule. Even the MSI Gaming X with the 8 pin doesn't actually need the 8 pin unless you want to heavily overvolt to overclock - benchmarks (below) show it uses the same energy as a single 6 pin RX480 (which basically makes the 8 pin unnecessary since 1x 6 pin means the card gets up to 150watts). And for reference, the 1x 6pin factory OCed 1060s such as the EVGA SC version has the similar overclock and boost overclocks as the MSI Gaming X version.

Compare that with the reference RX480 4GB/8GB with a TDP of 150watts, and had power issues that drew over 150watts when it was first released. After the power driver patch, it uses ~150watts, or basically the max from the 6 pin and the PCIe slot. Further compare that to the aftermarket factory OCed RX480s - ALL of them have 2x 6 pins or 1x 8 pins. I don't know of a single factory OCed RX480 with a single 6 pin power connector.

And benchmarks do not show they have the same power draw. If you look at my link from HardOCP, the factory OCed RX480 draws anywhere from 50 to 70 watts more than the factory OCed MSI Gaming X GTX1060. The reference RX480 8GB uses about the same amount of power as the OCed MSI Gaming X GTX1060. The reference stock GTX1060 uses 30-40 watts less than the Gaming X 1060 or the reference RX480 8GB.

So basically, the stock reference GTX1060 uses ~120watts. The RX480 8GB with the 1x 6pin and the factory OCed GTX1060 Gaming X both use ~150watts. The factory overclocked RX480 use ~200watts. So you're looking at 90-100watt difference between a reference 1060 6GB and an AIB factory OC RX480.

http://www.hardocp.com/article/2016...x_1060_founders_edition_review/9#.WEpIyrIrJpg
http://www.hardocp.com/article/2016...0_g1_gaming_vs_msi_gtx_1060_x/16#.WEpIn7IrJph

For manual overclocking/OCing in general, the stock speeds of the RX480 are already clocked near max stability, so you're going to exponentially increase the power consumption because you'll need to overvolt it. The 1060s, on the other hand can overclock very well on stock voltages, and thus doesn't significantly increase power consumption.

 

[Puiu]

I think you have your benchmarks confused. The GTX1060 6GB wins by a sizeable margin over the RX480 8GB in Witcher 3 and wins by a narrow margin in Battlefield 1. And this is with mid-November drivers. https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1060_Mini_3_GB/25.html

Even after this AMD driver improvement, the GTX1060 6GB will still be winning in Witcher 3 and will probably lose by a small margin in Battlefield 1.

According to your Relife Techpowerup benchmarks, the fps gain in Witcher 3 is a negligible 0.5%, which is negligible because the Techpowerup benchmarks show the GTX1060 6GB with a ~10-12% fps advantage. The larger fps gain of 3-4% are in games that AMD RX400s were losing in larger margins, with the GTX1060s winning by much larger than 3-4% fps. https://www.techpowerup.com/reviews/AMD/Radeon_Crimson_ReLive_Drivers/4.html

If they're the same price, people should go for the GTX1060 6GB over the RX480 8GB.
6GB vs 8GB is not much difference, and they have very different architectures and likely use memory different so it's probably not directly comparable. Nvidia's Pascal architecture is much more efficient, while AMD's RX400 is the equivalent of Maxwell architecture in terms of performance/watts. The GTX1060 6GB runs cooler, is less noisy, has lots of small size single-fan versions, requires a less powerful PSU, and has more overclocking headroom in terms of % clock speed increase and % fps gain. Most RX480s have very small overclocking headroom as the stock speed is already near the maximum stable clockspeed, and the RX480 power draw goes up exponentially when overclocking.

It seem that you like to ignore what people say and you didn't even check the links I gave you. The Witcher 3 tests you have were done with a reference RX 480. If you had checked what I gave you would see both the reference (which loses) and the Nitro+ oem card which wins.

I don't get it, do you have some sort of Nvidia fetish to actually be this aggressive in defending them? You have newer benchmarks done on the 5th of december using newer drivers with also both OEM and reference cards. Use those as a reference and add the 2% for the ReLife driver (from techpowerup's tests). Otherwise all of your arguments are thin and not grounded in reality.