New benchmarks faggots Post all GoW4 benchmarks you can find ITT

>>56953748
That scaling is sexy as fuck.

>>56953735
Why is Intel sucking so bad here compared to AMD on multicore scaling? AMD has near 100% scaling where as Intel is around 75% on 4c8t and down to shit scaling on 8c16t

What the fuck is going on? AMD optimized CPU calls?

>>56954041
...What? Both sets of numbers are showing that the game is very well-threaded and each core doesn't have to work as hard the more you have. The hex core AMD chips being completely tapped out and the octa core one barely coping isn't a positive sign for those chips. The load is so much lower on the Intel chips because the cores are that much stronger.

>>56954041
Personally I'd say it's because the four or six core i7s have a lot more breathing room than the FX CPUs.

Basically these charts show that the the FX CPUs are only saved by DX12 and the i3 6100 is a clear but passable minimum.

>>56953643
Where is the 480?

No Linux port,
Bye
y
e

>>56954300
Unfortunately gameGPU doesn't seem to have one. I believe that at one time they did, but maybe it's undergoing RMA or something.

>>56954300
Same as GTX 1060 6GB. Both being stock settings. That test was done on i7-5960x.

Given the scaling here >>56953748 on AMD, I'm 95% sure the RX 480 would lose badly on 8350 vs the GTX 1060.

<no sli support

are developers even trying anymore?

>>56953643
gamegpu has to be one the worst most useless benchmark sites on the web. They don't even use a 390 for testing.

>>56954563
that would imply they ever tried to begin with.

SLI/Crossfire has always been a meme concept and an afterthought for devs.

People who invest in dual GPU setups just get punished for it with belated or non existent support for AAA game titles

>>56954571
390 is same as 290X. So its worthless to do it again.

>>56954535
Only a 480 reference would be outperformed by the 1060. Sapphire version kills it.

>>56954535
Can't prove or disprove that statement without real world benchmarks.

DX12 driver overhead is small for both AMD and NVIDIA, so the 480 and 1060 are probably about equal.

>>56954571
Rehash m8

>>56954631
Only a 1060 reference would be outperformed by the 480. [Inset oc card here] version kills it.

>>56954713
Nice, except the 480 overclocks better.

>>56953724
Fx6300 and 8350 aren't doing bad at all. I'm impressed.

An OC6300 would kick ass

File: no_games_station.png (75KB, 1089x525px) Image search: [Google]

75KB, 1089x525px

>>56954745
>An OC6300 would kick ass
aka a 6350, too much trouble to OC manually

>>56954631
Both are stock. Both can overclock. I have a Windforce 1060 3GB. Reference is 1506 Mhz, my OC is 1737Mhz. Thats 15%. Saphire Nitro+ user overclock gets about 1360 Mhz. With reference being 1266, thats 10% OC.

>>56954669
Sorry about that. I thought I had the image uploaded. Image is already uploaded here at >>56953759

>Source
>http://www.pcworld.com/article/3128346/software-games/tested-gears-of-wars-4-pc-benchmarks-yield-glorious-graphics-options-galore.html

>>56954797
>3GB

>>56954793
Some 6300s could hit 5ghz, but even at 4.5ghz it might be on par with that haswell i5, if not slightly ahead.

For an i3 competitor it's performing far beyond its price point

>>56954797
The GCN arch scales better with clockspeed. The fucking amazing overclocker is the 470 though, considering ref clocks are so low. Of course the silicon lottery applies, but if you get a factory OC'd card, you're assured to get a better chip.

>>56954846
>Some 6300s could hit 5ghz, but even at 4.5ghz
Extremely hot, no thanks. ~4.2GHz is the sweet spot.

File: Power-Load1-670x217.jpg (34KB, 670x217px) Image search: [Google]

34KB, 670x217px

>>56954910
I don't think it'll be that bad, check out the power consumption difference.

http://play3r.net/reviews/amd-piledriver-fx-4350-fx6350-fx-8350-shootout-review/10/

http://www.pcper.com/reviews/Processors/AMD-FX-8350-and-FX-6300-Processor-Review-Vishera-Breaks-Cover/Power-Overclocking-

File: 1475643180340.gif (2MB, 400x291px) Image search: [Google]

2MB, 400x291px

GO AMD GO

>>56954733
>480 overclocks better.
Nice, except that's not true.

>>56954797
Made a bit of mistake somehow.

1360/1266 = 7%

>>56954910
http://www.hardocp.com/article/2016/08/21/powercolor_red_devil_rx_470_overclocking_review/3

This one shows that it can go upto 1390 on lucky rolls. The stock clock for RX 470 is bit lower than 480 @ 1206, so this is about 15%.

Still pretty decent. You maybe right on saying GCN gains more real performance per % from OC than Nvidia, however it doesn't tell us much. Roughly speaking, I'd more go with the general rule of thumb, with direct OC percentage gives roughly the same performance boost for all vendors.

>>56954964
AYYMD HOUSEFIRES

This game has multi gpu support in the menu, but doesn't use it???

Must be driver updates coming then?

File: Power-Idle1-670x217.jpg (37KB, 670x217px) Image search: [Google]

37KB, 670x217px

>>56954964
Note those are system benchmarks.

Difference between 6350 @ stock at load and idle: ~130W
Difference between 6350 @ 5GHz at load and idle: ~175W
Difference between 6350 @ stock and 6350 @ 5GHz under load: ~60W
Difference between 6350 @ stock and 6350 @ 5GHz at idle: ~40W

Considering the 6350 has a 125W TDP (probably uses less at load, considering the 8350 has the same TDP), you're pumping an extra 30-40% heat into your case, which is a lot. My 6350 is already fucking hot at stock, considering modern CPUs have a 60W TDP.

But it's definitely doable.

>>56955056
Probably, or something.

>1060
>68 min 87 avg (or is it max?)

Regardless it feels good to finally have a GPU that isn't years behind. I know the 1060 is going to be made irrelevant in the next year or two but I'm going to enjoy playing games at 1080p at 60fps at high/max settings for awhile.

>>56955157
You're right. I forgot the 6350 was 125w, was thinking of the 6300 that was 95w tdp.

>>56955265
1060 should be good for ultra settings for the next 2 years and high settings for the next 2 after that.

Or at least that's how my 670 aged at least.

>>56955319
Yep, a mid range 7nm GPU will be your next upgrade

>>56953724
Whoa, what's with the divide between 2500K and 2600K?

>>56955748
4 c / 4 t vs. 4 c / 8 t
SMT has arrived

File: dx12_cpu_gears_of_war_19x10_medium_dx12-100647959-orig.png (13KB, 1097x697px) Image search: [Google]

13KB, 1097x697px

>>56955761
But PCWorld found no difference with hyperthreading.

>>56954461
nobody cares neckbeard
>>>/g//fglt/

>>56955789
By that (faulty) logic they also found no difference with more than two cores, but you can see that there actually is a slight difference between 2 c / 2 t and 2 c / 4t

>>56953643

Another Microsoft masterpiece that gets 0 multi-card scaling and gets half the FPS it should be getting at 1080p on every card.

>>56955789
Anyway what's the source of that image?

File: des_1920_12.png (146KB, 548x629px) Image search: [Google]

146KB, 548x629px

>>56955894
You're quite retarded m8, compare with DX:MD which looks worse IMO

>>56955985
>nvidia sponsored gears 4 looks nicer whilst also performing way better than the amd sponsored day sex which runs like shit on even the higher tier of hardware

who would have thought? its hitman all over again.

>>56955789
Thus we've reached a bottleneck somewhere.

File: qb_1920.png (151KB, 548x629px) Image search: [Google]

151KB, 548x629px

>>56956117
OK, here's another example. The point is that GoW4 gets great framerates.

File: lpcc.png (63KB, 999x757px) Image search: [Google]

63KB, 999x757px

async compute and tiled resources enabled/disabled

File: async.jpg (28KB, 840x389px) Image search: [Google]

28KB, 840x389px

>>56956223

>>56956178
another game shilled by amd because of muh dx12 which runs like absolute garbage.

>>56956223
Note that that bench is done on an outdated driver. AMD just launch 16.10.something which has support for GoW4.

Again we see that NVIDIA's new cards aren't really hurt by DX12, they just isn't helped. I'd like to see that bench repeated with the 970 and 780Ti though.

File: 1357917261093.jpg (102KB, 640x360px) Image search: [Google]

102KB, 640x360px

>>56956252
>game shilled by amd
No. That developer is all over NVIDIA's dick

>because of muh dx12
That was MS's doing because they wanted the game exclusive to Win10.

>>56956309
>remedy
>all over nvidia's dick

nigga never go full retard. amd even advertised quantum break and halo on their own website like they had a partnership.

>>56954733
LOL

>>56956430
Quantum Break isn't even in AMD's list of featured games, dummy. Also both companies advertise whatever games are popular on their websites
http://www.amd.com/en-us/markets/game/featured

>>56953643
>under 9fps average because of my CPU
fuck yeah, time to get a 4k monitor for that glorious ~9fps

>>56956576
they literally produced a full slide of dx12 microsoft exclusive games labeled under radeon technologies.

>>56956178
Why is the 280x above a 380x if it performs slightly worse?

>>56956748
And? Just saying "This game can use DX12" doesn't mean that they were partnered during development. It isn't a Gaming Evolved title

>>56956785
They are sorted after minimum framerates.

>>56956785
Wider bandwidth, but your statement contradicted itself. The 280x beats the 380x usually. In some new games the 380x is outperforming the 280x (7970) because of the newer arch

>>56956791
>Just saying "This game can use DX12" doesn't mean that they were partnered during development.

actually it does show signs of a partnership. with microsoft. they can't produce a slide with copyrighted content like that and also advertise their own technologies alongside it.

>>56956851
i know you are. you should get some sleep.

File: 1355472121248.jpg (283KB, 484x484px) Image search: [Google]

283KB, 484x484px

>>56956832
Full retard

Forgot image

>>56956802
But they have the same minimum too.

>>56956813
But they get the same minimum and the 280x gets 1 fps lower average but is still on top of the 380x in this specific one benchmark that I was replying to.
This one in case you missed it >>56956178

>>56956944
Right, didn't notice that particular anomaly, but it's common in gameGPU charts. Could be a flaw of the software they're using to chart or the underlying data has a precision level we can't see, like 29.1 vs 29.3.

>>56954461
>Linux port for one of MS two or three relevant exclusives
heh

>>56956269
gamers nexus updated tests say that there wasnt any improvement using the newer drivers from amd

>>56954300
in case you havent noticed gameGPU is extremely biased towards nvidia

File: 1472013187349.gif (2MB, 500x281px) Image search: [Google]

2MB, 500x281px

>>56958272
>biased towards nvidia

>Any benchmarks that show nvidia at #1 is biased.

Why on the benchmark am I getting average of 122 fps with a 480 nitro+4790k @1080 yet this chart shows and struggling to get 60+?

>>56955789
That's Gears of War Ultimate Edition, the remake of GOW1, not GOW4.

That one was also a shit port.