GTX 980 & 970 Benchmarks!

[Humbug]

According to that 980m has 8GB VRAM. What are you going to do with 8GB VRAM in a mobile part? Seems wasteful.

 

 

I hope you're right, this is pretty lackluster and disappointing. Literally the same price:performance. Might as well get a 780ti now.

Maxwell is not expected to be revolutionary in terms of raw performance. We have seen in the 750ti the main advantage is power consumption.

[Khvarrioiren]

-snip-

 

DirectX IS a low level API. DX9, XD10, DX11, DX12, Mantle ,OpenGL are all low-level APIs.

 

DirectX doesn't bring in itself a low level APi because it is one. Missing feature sets are a whole different thing.

[qwertywarrior]

DirectX IS a low level API. DX9, XD10, DX11, DX12, Mantle ,OpenGL are all low-levelAPIs.

 

DirectX doesn't bring in itself a low level APi because it is one. Missing feature sets are a whole different thing.

TIL DX9 is a low level api

[Khvarrioiren]

TIL DX9 is a low level api

 

I didn't get what you're trying to say.

[OlekKing]

[qwertywarrior]

I didn't get what you're trying to say.

very high draw calls are what mantle and DX12 is after and also OpenGL NG

 

since DX9 microsoft was trying to improve draw calls

DX10 did try to remedy the issue with drawcalls

Dx11 also made changes to it

but by no means were it ever a main focus  during DX11 openGL was still was around ~2X better

 

mantle was one of the first in terms of a modern Low level Api

DX9 - DX11 and openGL are high level API

while OpenGL NG DX12 and Mantle are Low level api

 

what makes then Low level is the amount of abstraction AKA reduces the CPU overhead

"A low-level API cuts the level of abstraction and in the process makes draw calls cheap"

[Khvarrioiren]

-snip-

 

http://en.wikipedia.org/wiki/List_of_3D_graphics_libraries

 

You've misunderstood quite a alot. The amount of draw calls doesn't govern wheter an API is low-level or not.

[qwertywarrior]

http://en.wikipedia.org/wiki/List_of_3D_graphics_libraries

 

You've misunderstood quite a alot. The amount of draw calls doesn't govern wheter an API is low-level or not.

when u take a step back they look like low level API

especially when you try to compare them to java 3D lol

 

 

 

dx9 cant make the same amount of draw calls compared to DX10

and dx10 cant make the same amount of draw calls as DX12

 

take a look at this

http://www.anandtech.com/show/7889/microsoft-announces-directx-12-low-level-graphics-programming-comes-to-directx

 

dx12 mantle and OpenGL NG isnt as "low" as its predecessor

[Khvarrioiren]

-snip-

 

The amount of draw calls still doesn't govern whether an API is low-level or not.

 

I'm now done trying to make you understand that.

[qwertywarrior]

The amount of draw calls still doesn't govern whether an API is low-level or not.

 

I'm now done trying to make you understand that.

dx12 mantle and OpenGL NG isnt as "low" as its predecessors

thats what im trying to explain to you

[Silent_Xaxal]

The performance gains on the mobile cards is why I'm happy I'm broke now.

[Valentyn]

According to that 980m has 8GB VRAM. What are you going to do with 8GB VRAM in a mobile part? Seems wasteful.

 

 

My old man works with Solidworks on his laptop all the time. He'd greatly appreciate it, and has been hinting to wanting na upgrade for a while now.

Might see how the Thinkpads with a quadro stack up with VRAM.

[EChondo]

Maxwell is not expected to be revolutionary in terms of raw performance. We have seen in the 750ti the main advantage is power consumption.

To be honest, at least on Nvidia's side of things, power consumption isn't exactly the main show piece...I mean, it's great that we can get the same performance with less power, but it just seems lackluster.

 

Hopefully AMD has a bit more oomf.

[Humbug]

To be honest, at least on Nvidia's side of things, power consumption isn't exactly the main show piece...I mean, it's great that we can get the same performance with less power, but it just seems lackluster.

 

Hopefully AMD has a bit more oomf.

Power consumption is very important in the mobile world. In laptops how far you can go on battery power is key.

 

But ya on desktop high end graphics the buying decision between Nvidia and AMD never really comes to the down to power consumption.

 

I think in terms of performance the next significant jump from both AMD and Nvidia will be when they transition past 28nm.

[SSOB]

http://nl.hardware.info/nieuws/41303/msi-geforce-gtx-970-en-gtx-980-4-gb-in-prijsvergelijker

 

 

{not official but trustworthy] Clockspeeds : 

GTX980: GPU : 1127 MHz

Memory             : 7012 MHz

Read article for more info (translate)

 

Where to buy: ( Netherlands )

https://maxict.nl/product/4340015/msi-v801-1275r-vga-gefpunt-gtx-980-4gd5

 

No picture yet but the source is trusted.

 

Price :

 

€530,- ex. VAT

€640,- Incl. VAT

 

So the R.R.P. will probably be around $650, but if they use the stupid direct conversion thingy it may be $600

 

For the GTX 970:

 

https://maxict.nl/product/4340014/msi-v316-001r-vga-gefpunt-gtx-970-gaming-4g

 

€350.- ex VAT

€430,- Incl. VAT

 

 Again probably around $400-$450 dollars.

 

Manafacturer Serial numbers:

 

GTX970:

V316-001R

EAN: 4719072365752

 

GTX980:

V801-1275R

EAN: 4719072366292

 

 

Both cards will have 4GB of GDDR5 vRAM.

 

Topic will be updated soon...

 

Google translate of the article:

 

In recent days there have been several video cards surfaced based on GM204-400 video chip including a Zotac GeForce GTX 970 and GeForce GTX 980 from Inno3D. The GTX 900 series is expected to be introduced, in which the first GTX 980 and GTX 970 video cards will appear. September 19 The price Hardware.info the first cards are now also surfaced both derived from MSI. 

 

The MSI GeForce GTX 970 4GB Gaming graphics is currently be ordered from a single provider for € 427.11. The MSI GeForce GTX 980 4GB Gaming can be ordered for € 640.47. Well at Ebuyer Further specifications are not yet revealed, but presumably gets the GTX 980 a clock frequency of 1127 MHz and a memory clock frequency of 1753 MHz, which is an effective clock frequency of 7012 MHz.

 

 

Twin Frozr V Cooler. Will most likely will be used with GTX 980.

[Bloodyvalley]

FINALLY 4GB now please increase your bus size, nVidia!!

[AlwaysFSX]

Yeah that graph has been shown over and over again...

[Sidiox]

I just hope that that pricing is true, it makes the 970 very appealing

[SSOB]

Yeah that graph has been shown over and over again...

Yeah I know , i just put it in there for dramatic effect, the most important things are this:

 

 

Manafacturer Serial numbers:

 

GTX970:

V316-001R

EAN: 4719072365752

 

GTX980:

V801-1275R

EAN: 4719072366292

With this it will probably be possible to figure the exact specs out and even pictures etc.

[AlwaysFSX]

Yeah I know , i just put it in there for dramatic effect, the most important things are this:

 

With this it will probably be possible to figure the exact specs out and even pictures etc.

If it's true, it shows how bad prerelease drivers are. Nvidia usually gets 95+5 scaling on SLI, not 66%. So we're still looking at GREAT performance.

[Swndlr]

i wanna see the real deal evga and asus models #zotacnogood

[themaniac]

just wish it would show the performance of the 960

[AlwaysFSX]

Are you serious..? GTX 770 is ~€300 right now, it's a HUGE increase.

770 and 680 release prices though.. it isn't too bad for the performance.

[Sidiox]

Are you serious..? GTX 770 is ~€300 right now, it's a HUGE increase.

 

 

770 and 680 release prices though.. it isn't too bad for the performance.

Exactly what this ^, it is a release price and it will drop quite fast (or so I hope) 

[AlwaysFSX]

Exactly what this ^, it is a release price and it will drop quite fast (or so I hope) 

Given what AMD may or may not be offering for their 3xx series. It isn't too bad for a top tier card no longer being $650/$700+.