NVIDIA GeForce GTX 960 Launch Date Revealed

Ian Evey · 1 Jan 2015 at 15:11

Orangey said:
770 = 680
280X = 7970GHz, which pwned the 680

Makes sense to me.

Yeah 1 year after release it did, If only amd had released it In The first place like it should have been, They would have severely damaged the sales of the then faster 680.

humbug · 1 Jan 2015 at 15:16

Ian Evey said:
http://www.guru3d.com/articles-pages/gigabyte-geforce-gtx-970-g1-gaming-review,7.html

Depends if you regard upto 100 watts less not worth it, many sites have stated around £30 per year saving. All depending on what exactly you do with the the card, like overclocking.

It depends on what game you are playing, or even what part of the game. It varies a lot because some games stress the card more than others, this doesn't matter with Hawaii cards as they always run at full power.

Maxwell doesn't, unless the game demands it, in which case it pull just as much as the 290.

This is what 90% of those who buy them for efficiency don't realise

Tom's did some very extensive testing to figure out what's going on with them, its an old link but i can dig it up if you like?

Ian Evey · 1 Jan 2015 at 15:29

What is it at 1080p though? I'll be sure to point anyone thinking of buying a 970 or two and is playing at 1600 res and is thinking about power consumption to you post.

Darujhistan · 1 Jan 2015 at 15:34

Ian Evey said:
http://www.guru3d.com/articles-pages/gigabyte-geforce-gtx-970-g1-gaming-review,7.html

Depends if you regard up to 100 watts less not worth it, many sites have stated around £30 per year saving. All depending on what exactly you do with the the card, like overclocking.

Is that all? I'm not going to worry about that.

I'm not an overclocker/bencher I'll have a play around for a couple of days and see what a new card can do compared to the old card, then I get bored.

Tonester0011 · 1 Jan 2015 at 15:35

My 970 was chugging a good 70+watts less than my 780ti during all tests i done. That's at 1440.

Kaapstad · 1 Jan 2015 at 15:50

I would not take Tom's review too seriously as I think they are being a bit economical with the truth. I have tested out the power consumption of the 980s and 290Xs and it is quite a bit different to what Tom's got.

1. I agree that the 980s do use more than their 178 TDP and it is quite easy to do this.

2. The 290Xs use a lot more than the 980s.

What I did was check the power usage 4 up with the 980s and 290Xs using Firestrike Extreme which will push power usage higher than games do.

One important difference between the two systems I used was the 980s were paired up with a 5960X and the 290Xs were paired up with a 4930k. What is important to note here is a 5960X will use more watts overclocked than a 4930k will.

This is the max watts used

Firestrike Extreme

All GPUs and CPUs were overclocked to near their max limits.

4 x 290Xs and 4930k = 1778 Watts

4 x 980s and 5960X = 1200 Watts

As you can see from the above the way I tested was a bit crude but it is a non contest, despite being paired with the more power hungry 5960X the 980s used a lot less watts. I think Tom's are being a bit naughty and can go and take a jump on this one.

As a rough guide the 980s use about 200 Watts each when overclocked.

Geeman1979 · 1 Jan 2015 at 15:53

Good info ^^^

But no real surprise...

Tonester0011 · 1 Jan 2015 at 16:00

It's also worth noting that those tests are done on a OC card with a decent cooler. The reference 980 with the normal 178tdp cooler used, 180w iirc?

humbug · 1 Jan 2015 at 16:21

Kaapstad said:
I would not take Tom's review too seriously as I think they are being a bit economical with the truth. I have tested out the power consumption of the 980s and 290Xs and it is quite a bit different to what Tom's got.

1. I agree that the 980s do use more than their 178 TDP and it is quite easy to do this.

2. The 290Xs use a lot more than the 980s.

What I did was check the power usage 4 up with the 980s and 290Xs using Firestrike Extreme which will push power usage higher than games do.

One important difference between the two systems I used was the 980s were paired up with a 5960X and the 290Xs were paired up with a 4930k. What is important to note here is a 5960X will use more watts overclocked than a 4930k will.

This is the max watts used

Firestrike Extreme

All GPUs and CPUs were overclocked to near their max limits.

4 x 290Xs and 4930k = 1778 Watts

4 x 980s and 5960X = 1200 Watts

As you can see from the above the way I tested was a bit crude but it is a non contest, despite being paired with the more power hungry 5960X the 980s used a lot less watts. I think Tom's are being a bit naughty and can go and take a jump on this one.

As a rough guide the 980s use about 200 Watts each when overclocked.

3DMark isn't at all stressful, nor would it be as there is no Direct Compute Lighting, Shading ecte... there isn't actually a lot going on for the GPU to do, its not a game. its an arbitrary benchmark tool.

Something Like Tomb Raider has a lot more going on, and pushes the GPU much harder.

Kaapstad · 1 Jan 2015 at 19:38

humbug said:
3DMark isn't at all stressful, nor would it be as there is no Direct Compute Lighting, Shading ecte... there isn't actually a lot going on for the GPU to do, its not a game. its an arbitrary benchmark tool.

Something Like Tomb Raider has a lot more going on, and pushes the GPU much harder.

Firestrike Extreme is very stressful on the GPU tests.

It is also harder to do big overclocks for the same reason.

Kaapstad · 1 Jan 2015 at 20:10

humbug said:
3DMark isn't at all stressful, nor would it be as there is no Direct Compute Lighting, Shading ecte... there isn't actually a lot going on for the GPU to do, its not a game. its an arbitrary benchmark tool.

Something Like Tomb Raider has a lot more going on, and pushes the GPU much harder.

Just ran this for the Tomb Raider bench

4 x 980 @1482/2002
5960X @4.5
1080p
347.09 drivers

This was run using the max volts available to the cards and also way over the top on vcore for the CPU (1.46v which is a lot for a 5960X).

It is the highest score on that bench.

How many watts max did I measure for this at the wall ?

@humbug or anyone else if they want to guess.

Tonester0011 · 1 Jan 2015 at 20:12

1005w

Geeman1979 · 1 Jan 2015 at 20:14

1100w

Kaapstad · 1 Jan 2015 at 20:17

The cards are the EVGA 980 SC version so they are already running a big factory oc @stock.

Below is the additional volts and clocks I added.

Davedree · 1 Jan 2015 at 20:26

is that 1.21v + .087
or 1.24v + .087 ?

Ian Evey · 1 Jan 2015 at 20:32

Kaapstad said:
Just ran this for the Tomb Raider bench

4 x 980 @1482/2002
5960X @4.5
1080p
347.09 drivers

This was run using the max volts available to the cards and also way over the top on vcore for the CPU (1.46v which is a lot for a 5960X).

It is the highest score on that bench.

How many watts max did I measure for this at the wall ?

@Humbug or anyone else if they want to guess.

What would be interesting is what the 980's and 290x's get in tombraider as humbug "knows" they are very similar in terms of wattage

My guess is 1067 watts.

Kaapstad · 1 Jan 2015 at 21:34

Ian Evey said:
What would be interesting is what the 980's and 290x's get in tombraider as humbug "knows" they are very similar in terms of wattage

My guess is 1067 watts.

For anyone who wants to know what the 290Xs use when overclocked I just ran this.

4 x 290X @1230/1500
4930k @4.7
1080p
14.12 drivers

Watts used max was 1689 watts.

Running the 290Xs @stock and the CPU at the clock above max watts is 500 less.

So anymore guesses on what 4 heavily oced 980s and a 5960X use ?

Davedree · 1 Jan 2015 at 21:43

1110 watts

triss · 1 Jan 2015 at 21:46

999 watts

Kaapstad · 1 Jan 2015 at 21:53

Can not wait for humbug so here is the figures.

Kaapstad said:
4 x 980 @1482/2002
5960X @4.5
1080p
347.09 drivers

Watts used max was 1068 watts.

Kaapstad said:
4 x 290X @1230/1500
4930k @4.7
1080p
14.12 drivers

Watts used max was 1689 watts.