Titan X Benchmarks.

[Kaapstad]

If any one gets a Titan X you can post your benchmarks in this thread.

Here is one I did earlier.

4 x Titan X @1392/1952
5960X @4.5
Watercooled.

http://www.3dmark.com/3dmv/5249873

I don't think I will stay at the top of that table for long.

 

[Dawelio]

I don't think I will stay at the top of that table for long.

How so?...

 

[JR23]

How so?...

Because someone will whip out their LN2 pots and smash it

I don't think i'll be joining the Titan X club. How do they do in 1080p Heaven, all the frames?

JR

 

[Dawelio]

Because someone will whip out their LN2 pots and smash it

JR

And this is why I'm a newbie ... No idea what it is or means

 

[Kaapstad]

Because someone will whip out their LN2 pots and smash it

I don't think i'll be joining the Titan X club. How do they do in 1080p Heaven, all the frames?

JR

Here is a link to the Heaven 4 1080p Single GPU scoreboard on a benching thread I run.

http://forums.overclockers.co.uk/showpost.php?p=25214819&postcount=1741

There are some Titan Xs at the top.

Below is my Single card 1080p score

Titan X @1446/2002
5960X @4.5

 

[Warchild]

Hi Kaapstad, perhaps you can do the same benches with 1, 2 and 3 card configs to have a good comparison for others with these setups. Not that I will ever see that many with more than 2 cards

 

[Kaapstad]

Hi Kaapstad, perhaps you can do the same benches with 1, 2 and 3 card configs to have a good comparison for others with these setups. Not that I will ever see that many with more than 2 cards

Here is Tomb Raider on 1 and 4 cards, you can work out 2 and 3 cards as the game scales really well with SLI. I set the same clocks for both 1 and 4 cards but the single card boosts slightly higher.

Tomb Raider Bench
Maximum settings
2160p

1 x Titan X @1435/2002
5960X @4.0

4 x Titan X @1405/2002
5960X @4.0

 

[SeanAngelo]

Me:

 

[Kaapstad]

Same settings and clockspeeds as above but with the resolution @1080p maxed

Single Titan X

4 X Titan X

 

[MrKambo]

I've always wondered, do higher clock speeds on each card in SLI matter? Or do they all need to be the same?

 

[Warchild]

I've always wondered, do higher clock speeds on each card in SLI matter? Or do they all need to be the same?

I think its irrelevant. From what I gather, when in SLI, the cards will run at the speed of the slowest card.

 

[JR23]

I think its irrelevant. From what I gather, when in SLI, the cards will run at the speed of the slowest card.

Not any more they all boost individually, you can overclock them individually or bind them together.

JR

 

[Kaapstad]

With 4 GPUs even if you set them to run the same when they boost they often end up running slightly different clockspeeds but normally within 13mhz of each other.

On a stock GTX 690 you also see this with the two cores boosting slightly differently.

Having said that my 4 TXs all seem to boost the same when working together, this is the first time I have seen this with 4 way SLI.

 

[Warchild]

Not any more they all boost individually, you can overclock them individually or bind them together.

JR

Interesting. So would it be better then to OC the cards together until you hit your maximum , then try to tweak each one individually to see which one of the two could go higher?

 

[JR23]

Interesting. So would it be better then to OC the cards together until you hit your maximum , then try to tweak each one individually to see which one of the two could go higher?

Like Kaap said they all kind of reserve the right to do their own thing even when you do tie them together. On air my 780's always wanted to do something different to each other. But under water they seem to behave exactly the same, GPU boost 2.0 keeps its eye on a bunch of parameters and it won't hold back the array because one card is outside of the targets. I think when they a properly cooled, properly powered and have excessive power limits set they should theoretically do the same clock.

I've never tried to overclock mine individually because I run out of power before I can max them out. There isn't much difference between them though, one can do a single bin higher on the core and the other will go an extra ~100MHz on the memory.

JR

 

[Kaapstad]

Interesting. So would it be better then to OC the cards together until you hit your maximum , then try to tweak each one individually to see which one of the two could go higher?

Some people do the above.

It is also a good idea to put the fastest card in slot one

The next fastest in slot three (in case you want to run 2 way SLI).

The third fastest in slot two

The slowest in slot four.

If you have the fastest card in slot one it can also pull a tiny bit more performance out of the slower cards, more so in a two way system.



Sleeping Dogs @2160p
5960X @4.5
Titan X @1405/2002

Settings Maxed

Single card

4 cards

Hitman benchmark max settings.

Single TitanX @1526/2002
5960X @2002
350.05 Drivers

1080p

2160p

 

[AlienALX]

I've always wondered, do higher clock speeds on each card in SLI matter? Or do they all need to be the same?

Just use Precision X or Afterburner to sync the clocks and lock them to the same frequency.

 

[Kaapstad]

5960X @4.5
4 x TitanX @1455/2002
2160p
350.12 Drivers

Just use Precision X or Afterburner to sync the clocks and lock them to the same frequency.

This is what I normally do but even then they will still boost to different levels.

Here is a Heaven 4 bench maxed @1440p, it is a higher score than the original Titans could do @1080p

4 X TitanX @1480/2002
5960X @4.5
1440p
350.05 Drivers

4 X TitanX @1480/2002
5960X @4.0
2160p
350.05 Drivers

 

[Dawelio]

For people who have Titan X's, how will one perform in Battlefield 4 on Ultra at 1440p?...

And also, how are the noise levels etc of the cooler under heavy load?... is it loud?...

 

[Kaapstad]

For people who have Titan X's, how will one perform in Battlefield 4 on Ultra at 1440p?...

And also, how are the noise levels etc of the cooler under heavy load?... is it loud?...

The cooler can be a little noisy much the same as the original Titans was.

You can set a custom fan profile which helps.

Most people don't have a problem with the cooler but a few do not like it.

As to BF4 I will give it a run next time I am off work but IIRC a single Titan X should max it @1440p.