Asus ROG Strix RX480 OC vs MSI Gaming GTX 1060 OC [HardOCP]

[Hail The Brain Slug]

I can see that you didn't read the entire article. They also tested the overclock with stock voltage, and they got 1360.. Obviously you have a golden sample, as I haven't seen any reviewer hit the speeds that you have, with or without additional voltage.

And as it's been mentioned, benchmarks and actual games are two different species.

I tried, but I was reading it on my phone and must have missed that part. My point was more that you can easily make any component look like its a power hog if you just crank voltage to maximum.

You can't really compare a video card with maxed out voltage to one with locked voltage control and consider it a fair comparison. They clearly hit the voltage wall on theirs and only gained 50 MHz for a ton more watts.

Not to mention the power usage on the RX480 is very peaky. At the power usage conclusion, they don't say if the power figures are measured as an average under load or at the peak/maximum draw. I'd be interested to see the difference between those two figures for both cards.

 

[Bacon1]

Gears 4 is a pure DX12 title.

So is Gears of War Ultimate Edition right?

An "official" UE4 roadmap, on t-rello.com?

T-rello.com (why is this censored?) is where a lot of companies post their public roadmaps if they do sprints / scrum. Many use it internally as well.

https://forums.unrealengine.com/showthread.php?4418-UE4-Roadmap

Which links to:

http://unrealengine.com/roadmap

Which in turn redirects too.... the link he gave you on trello.

I never originally said that people who buy 1060/480 GPUs overclock them heavily. I said that people that buy premium versions of these GPUs are more likely to engage in overclocking

Ahh ok, so only people buying these two specific cards OC them heavily. Do you have stats for that to show?

 

[Carfax83]

I like how the OP managed to turn the review into a negative for the RX480

I don't think I was being negative towards the RX480. I've said publicly that it's a great GPU in fact. My point is that Pascal is just a superior architecture, as it does more with less. A lot of people on these forums think that performance per watt is some obscure engineering term that has no relevance to end consumers. But it does, because it has a major impact on the end performance of a product..

 

[Carfax83]

I tried, but I was reading it on my phone and must have missed that part. My point was more that you can easily make any component look like its a power hog if you just crank voltage to maximum.

You can't really compare a video card with maxed out voltage to one with locked voltage control and consider it a fair comparison. They clearly hit the voltage wall on theirs and only gained 50 MHz for a ton more watts.

Perhaps it's not completely fair as you say, but I can understand HardOCP's motive for attempting to get the maximum amount of overclock possible from each card. Regardless, the point remains as salient as ever. The stock clocked Asus ROG RX480 still consumes more power than the overclocked GTX 1060 whilst delivering less performance.. That's to those that don't think performance per watt matters

 

[Carfax83]

So is Gears of War Ultimate Edition right?

Gears of War Ultimate Edition is based on UE3, and I think was ported over to DX12 by The Coalition only. GoW4 on the other hand received much more extensive assistance from Microsoft and Epic when it came to DX12 implementation.

Which in turn redirects too.... the link he gave you on trello.

I never heard of that website before. Still, like I said the cards aren't exactly up to date. Gears of War 4 came out in October, and the last card to be updated for DX12 occurred in January.. That said, I fully expect the DX12 implementation to be an ongoing thing..

When SM 6.0 is released, DX12 will be updated again as will any engine that supports it. Progress is unceasing after all.

Ahh ok, so only people buying these two specific cards OC them heavily. Do you have stats for that to show?

You know full well what I mean. Premium models that are geared towards overclocking (better cooling, power delivery etcetera) are usually bought by hardware enthusiasts that do overclocking. There is nothing controversial or unreasonable about this statement, and it's self explanatory. AiB manufacturers make these cards for all tiers, from the low end to the ultra high end, which busts your theory that hardware enthusiasts don't buy low or midrange cards..

Hardware enthusiasts come in all shapes and sizes, and income levels

 

[casiofx]

I don't think I was being negative towards the RX480. I've said publicly that it's a great GPU in fact. My point is that Pascal is just a superior architecture, as it does more with less. A lot of people on these forums think that performance per watt is some obscure engineering term that has no relevance to end consumers. But it does, because it has a major impact on the end performance of a product..

It does have impact like you said. People with crap power supply can run gtx1060 without much problems

 

[YBS1]

During the run I wasn't afraid of the card though. More worried about the motherboard going *POP* with how high the peak power draw was. I wonder if there's an easy way to check and see if, in cases of extreme draw like this, the MB power draw is still limited and all the extra power is drawn from the PSU directly or not.

That's exactly what I was afraid of in my benchmarking runs. I've never seen a full on confirmation that the card will not draw more than 75 watts from the motherboard, though I know an early driver addressed this issue by choosing to "pull more power from the six pin", it was not specifically stated the pci-e rail was hard limited to just 75w under any condition AFAIK. Not sure just how much abuse my mini-itx board could take.

 

[Dice144]

Love my mini 1060 6GB. Replaced my coil whine 290x which allowed me to move to a smaller case, as the old one was water cooled with a NZXT kraken x41. Currently running it at +115 core +50 memory, boosts to ~1999 when playing Skyrim SE. Honestly haven't tried to go higher but temps stay under 60 C when gaming in my ice box. Room temperature around 60 F.

 

[guskline]

YBS1, how do you like the RX 480 in your 6700k rig? From your sig, I can see you have the $$ to spend on very good quality parts.

When I built my third rig below, also 6700k, I already owned a GTX 1080 and a GTX 980TI SC so perhaps the attraction of the newly released Polaris chip was enough for me to choose an RX480 vs a GTX1060.

I think the moral is that AMD has really closed the gap in mid range cards at the RX480/GTX1060 price range.

I have not tested a GTX 1060 6g but have every reason to believe it is an excellent midrange card.

Some of us "explore" other gpu avenues because we can afford too; thus my 2 RX480s in CF!
From a pure price/performance standput, a single GTX 1070 is most like better. However, since I started out with a single RX480 8 g reference card (bought at $239) and since a few months later I snagged a second identical card, lightly used for the same price, I thought I would give it a go to see how it runs. Especially since AMD Raja hinted that it might surpass a GTX1080.

WELL, it does surpass a GTX 1080 slightly in the graphics score number of Firestrike normal but starts to tail off in Extreme and Ultra. Sure 2 RX480s cost me @$485 vs $699 for my GTX1080.

Seriously, though, these are different gpus for different resolutions.

 

[caswow]

lol the msi is the biggest power hog of all the rx480s.

 

[USER8000]

Finally a review comparison that pits a highly overclocked RX480 vs a highly overclocked GTX 1060..

RX480 is set to 1410 and 8.8ghz on the VRAM, and the GTX 1060 is at 2.16ghz and 9.2ghz..

This review shows the potency of the Pascal architecture, and its power efficiency. RX480 when overclocked uses over 100w more than the GTX 1060, whilst providing less performance in the majority of the titles tested..

I don't think I was being negative towards the RX480. I've said publicly that it's a great GPU in fact. My point is that Pascal is just a superior architecture, as it does more with less. A lot of people on these forums think that performance per watt is some obscure engineering term that has no relevance to end consumers. But it does, because it has a major impact on the end performance of a product..

You are trying to make the HardOCP review look negative,since they are other reviews which show much less of a difference between the various GTX1060 and RX480 and I don't remember you making a new thread about it.

Or should we wheel out the ones which show them much closer together??

Like this one:

https://www.computerbase.de/2016-07/powercolor-radeon-rx-480-red-devil-test/5/

Oh,wait should we all start making new threads for that.

Plus you are another enthusiast who looks at absolute numbers - they are using a highish end setup which will consume more power than most desktops a £250 card will be put into.

LMAO,380W at the wall if you actually overclocked is probably closer to a 300W to 350W internal load on the PSU with an overclocked Core i7 running at 4.7GHZ!

Basically any reasonable 400W to 500W PSU is fine.

This is what HardOCP had to say:

ASUS has done a tremendously good job at engineering a robust video card that can take the AMD Radeon RX 480 GPU to new heights. The ASUS ROG STRIX RX 480 O8G GAMING is capable of a high voltage setting, and it works, with the combined excellent cooling of DirectCU III and the Power Limit slider we are able to experience a high consistent clock speed overclocked. The DirectCU III cooling solution works very well, keeping the GPU at cool temperatures with no noise. Even overclocked, where we manually increased the fan speed the noise level was more than tolerable and the temps were excellent. You could probably lower the fan speed from what we had set with the overclock achieved.

As we mentioned, this is the highest overclock we’ve achieved yet with any AMD Radeon RX 480 GPU. The overclock has improved performance over the card’s default out-of-box clock speeds as well as a reference AMD Radeon RX 480 by a great deal. The overclock provides a real, tangible benefit to the gameplay experience. With the overclock the ASUS ROG STRIX RX 480 O8G GAMING video card is very good competition to highly overclockable GeForce GTX 1060 video cards as we have shown.

There is an MSRP of $299 on the ASUS ROG STRIX RX 480 O8G GAMING video card, however, you do get a lot of potential and customized hardware with that price. Thankfully however there are price savings and rebates currently making this video card much less expensive. It can be had at $259.99 after $20 MIR at both Amazon and Newegg.

At $260 this video card competes well with the likes of highly overclockable GeForce GTX 1060 video cards like we have compared to in this evaluation. At $260 you know you’ll be getting a video card that has what it takes to push the AMD Radeon RX 480 GPU to its limits in performance.



If you are looking for one of the best AMD Radeon RX 480 GPU based video cards out there this holiday season, the ASUS ROG STRIX RX 480 O8G GAMING should be on your short list.

HardOCP are very positive about the card overall.

 

[USER8000]

I also decided to look at some other reviews of the card,and it looks like the OP waited for the one which has the worst power consumption overall too.

So a 34W to 52W difference over the GTX1060,and yes the GTX1060 is more efficient,but not to the extent the OP is framing the very positive HardOCP review.

Edit to post.

Also,apparently some have forgotten this card:

It is the first AIB card using a single six pin power connector and is one of the cheaper RX480 cards. It uses a different PCB and cooler compared to the reference cards.

It will be limited to 150W at most for the card.

All of the reviews indicate an RX480 in a desktop with a Core i7 as a total system will draw around 215W to 269W at the wall.

Sure the GTX1060 is better still,but its not like any of these cards are exactly taxing a PSU anyway.

I also think not all of it is the GPU design too - AMD is using Global Foundries who are behind Samsung using the same technology. Just look at how the GP107 can be had in decent volume as a full part - look at the RX460,it was the volume Polaris 11 part and it needed shaders to be disabled.

 

[happy medium]

max overclock vs max overclock the rx 480 dumps 109 watts more heat in your computer case and your room, than the faster gtx1060.
That's what this review tells us.

It seems when you overclock a gtx1060 you use less extra power than a overclocked rx480 does. Are some of you guys saying that HardOCP is making up power figures?

I would suggest anyone that has their own power tester (not gpuz) and wants to show us they are using less power with their 480 @ 1410 MHz link some proof.

 

[Hail The Brain Slug]

max overclock vs max overclock the rx 480 dumps 109 watts more heat in your computer case and your room, than the faster gtx1060.
That's what this review tells us.

It seems when you overclock a gtx1060 you use less extra power than a overclocked rx480 does. Are some of you guys saying that HardOCP is making up power figures?

I would suggest anyone that has their own power tester (not gpuz) and wants to show us they are using less power with their 480 @ 1410 MHz link some proof.

After my little experiment, I'm still not happy with GPU tweak, so I started testing 1410 at 1.175 through Wattman and its looking promising.

I have a killawatt and would do a little power testing but I just won't have time today. Maybe over the weekend.

 

[happy medium]

After my little experiment, I'm still not happy with GPU tweak, so I started testing 1410 at 1.175 through Wattman and its looking promising.

I have a killawatt and would do a little power testing but I just won't have time today. Maybe over the weekend.

That's great, thanks and you seem to have the best, least power consuming, rx 480 on the internet.
I have read at least 10 reviews comparing 480's and 1060's. Generally the 480 uses about 40 watts more at stock than a gtx1060 while gaming.

I think we would all agree that when you overclock a polaris AMD card you will expect to use much more power than the equivalent overclocked Pascal Nvidia card this round.

 

[AtenRa]

From my XFX RX 480 GTR Black

BF4 at 2560x1440 custom settings (most of them at Ultra)

System specs
Core i7 3770K @ 4.4GHz
4x 4GB DRR3 2133MHz
Corsair i80 Water Cooler
Vertex 4 250GB SSD + Kingston V400 120GB SSD
Enermax 850W Revolution 85+

Win 10 64bit
Crimson 16.12.2

With the RX 480 at default, peak power consumption for the entire system at the wall was 345W.
Maximum card power draw according to GPUz was 165.3W at 1338MHz.

 

[happy medium]

TPU did a review with the same ASus card. 180 watts peak stock for the card.
They test the card itself , not the entire system.
https://www.techpowerup.com/reviews/ASUS/RX_480_STRIX_OC/21.html

HardOCP states that their system power increased by 70 watts.
Quote:
"The ASUS STRIX RX 480 O8G GAMING by default idled at 66W and while gaming pulled a total system Wattage of 310W. When we overclocked the video card the gaming Wattage did shoot up high to 380W"

180 + 70 = 250 watts for a rx480 when max overclocked.

 

[AtenRa]

This ASUS RX 480 [H] used is definitely not very efficient,

From the [H] review the card at 1310MHz and only up to 1.15v used 208W, when my XFX at 1338MHz with up to 1.2V only used 165,3W

 

[Mopetar]

As with previous similar discussions you can't make direct comparisons about the architectures since they're made on separate processes which will impact power draw and OC potential as much as the design of the chip itself.

The 1060 clearly has better OC potential and is less wasteful at the plug, but that doesn't mean Polaris itself is bad, inefficient, or a poor OC, just that the 480 isn't as good. Unless you get AMD to fab Polaris at TSMC or NV to fab at GF you can't compare the two architectures, just the cards themselves.

 

[tential]

max overclock vs max overclock the rx 480 dumps 109 watts more heat in your computer case and your room, than the faster gtx1060.
That's what this review tells us.

It seems when you overclock a gtx1060 you use less extra power than a overclocked rx480 does. Are some of you guys saying that HardOCP is making up power figures?

I would suggest anyone that has their own power tester (not gpuz) and wants to show us they are using less power with their 480 @ 1410 MHz link some proof.

Is it really relevant the exact wattage? Why argue over that.
Amd dumps more beat into your room. That's it. Period.
I was actually wondering if the 480 oc is even reasonable

 

[AtenRa]

Is it really relevant the exact wattage? Why argue over that.
Amd dumps more beat into your room. That's it. Period.
I was actually wondering if the 480 oc is even reasonable

I dont know if my XFX RX 480 GTR Black is golden sumple or not but at 1405MHz on the Core and 2200MHz for the memory + 50% power my card used 184.2W. That is even lower than the ASUS on default.

Edit: This is auto OC, i haven't even touched the voltages yet.


Same system, same game, same settings as above.

 

[Hail The Brain Slug]

I dont know if my XFX RX 480 GTR Black is golden sumple or not but at 1405MHz on the Core and 2200MHz for the memory + 50% power my card used 184.2W. That is even lower than the ASUS on default.

Edit: This is auto OC, i haven't even touched the voltages yet.


Same system, same game, same settings as above.

What game is that?

 

[AtenRa]

What game is that?

BF4 Multi 64player

 

[happy medium]

seems gpuz is wrong.

 

[96Firebird]

GPU 'only' power =/= card power. GPU-Z even updated that note for Polaris cards to distinguish the difference.