- Joined
- Feb 22, 2009
- Messages
- 780 (0.13/day)
| Processor | Ryzen 7 5700X3D |
|---|---|
| Motherboard | Asrock B550 PG Velocita |
| Cooling | Thermalright Silver Arrow 130 |
| Memory | G.Skill 4000 MHz DDR4 32 GB |
| Video Card(s) | XFX Radeon RX 7800XT 16 GB |
| Storage | Plextor PX-512M9PEGN 512 GB |
| Display(s) | 1920x1200; 100 Hz |
| Case | Fractal Design North XL |
| Audio Device(s) | SSL2 |
| Software | Windows 10 Pro 22H2 |
| Benchmark Scores | i've got a shitload of them in 15 years of TPU membership |
Greetings this wonderful after Easter time! I'd like to invite every GTX970, GTX980 Ti, GTX1070 owner or any NVIDIA fan in general to join the discussion and share your experiences!
People looking for a video card to run games at 2560X1440 res. found the right place. Perhaps some folks are looking to upgrade their GeForce GTX970 to something that can handle 120 or 144 FPS at full HD? The options are pretty obvious if we are talking about new video card from first hand retailer - GTX1070 is the card to buy right now. It's positioned at the price and performance segment that is just right for 2560x1440 res. monitors. I think most people will agree with me that GTX1080 is a bit of an overkill for 2560x1440 res. monitors considering the premium price and that it is a 4k ready video card. I expect most of you agree that GTX1060 6 GB is too weak for some of the more demanding titles at 2560x1440 res. While AMD has R9 Fury X and R9 295X2 as alternatives for 2560X1440 res. gaming, as a G-SYNC monitor owner i will stick to NVIDIA this time.
The primary point of this thread is to prove that SLI GTX970 can beat GTX1070 costing much less. Everything eventually comes to the price. I was intrigued to do this test, because recently one of the local retail shops in my country started selling brand new Zotac GTX970 AMP! Omega Core video cards for 190 EU, which is 30 EU cheaper than the cheapest GTX1060 3 GB you could get. This is a great value since i have proven that GTX970 is equal to GTX1060 3 GB in this thread. Having this in mind an idea poped: what if SLI GTX970 could beat the cheapest GTX1070 you could find? The cheapest stock GTX1070 with a custom cooler costs 405 EU, but the fact of the matter is that you can buy two used custom GTX970 OC video cards for 300 EU this time around from second hand retailers; and so i bought two GTX970 video cards for 300 EU - Zotac GTX970 AMP! Omega Core and Gigabyte GTX970 G1 Gaming, as well Asus GTX1070 Dual, that costs 445 EU. I have slightly overclocked the Zotac GTX970 AMP! Omega Core to equally match the Gigabyte GTX970 G1 Gaming, so in practice for this benchmark i have two GTX970 G1 Gaming cards.
The second objective of this thread is to prove that factory overclocked GTX980 Ti can beat or at the very least tie a stock GTX1070. This itchy idea came from the original techpowerup GTX1070 review, in which a stock GTX1070 proved to be 13,5 % faster than a stock GTX980 Ti. I was, frankly speaking, displeased, and thought no way the stock GTX1070 can beat my Asus GTX980 Ti Strix OC, and that comes to play today. I could find a used custom GTX980 Ti for 300 EU these days and that means it competes directly with two GTX970 cards to dethrone the more expensive GTX1070.
For this test i have selected 25 games, but i found out that 5 of them were not supporting SLI at all: Company of Heroes 2, Call of Duty Infinite Warfare, Arma 3 Apex, Ashes of the Singularity (DX11) and Rainbow Six Siege, therefore to be FAIR, i, obviously have not included them in this benchmark and only included games, supporting at least some noticeable scaling.
BENCHMARK HARDWARE & DRIVERS
Intel Core i7 5775C 4 GHz
Gigabyte Z97X-Gaming 3
Crucial Ballistix Tactical 2X8 GB DDR3 1600 MHz C8
Samsung Evo 500 GB SSD (all games are installed in it)
Windows 7 Pro X64 (updates installed)
NVIDIA Forceware 378.66
I tested all the games with Corsair Vengeance Pro 2400 MHz C11 DDR3 and found no difference in performance, so don't talk about "low RAM speed BS please".
I had a lot of trouble finishing this benchmark at the planned time (a month ago) and i could not update the drivers to newest as something that worked with one card did not work with another later, so by the time i fixed the problem new drivers would appear and i had no time to redo all the stuff again. Please understand that.
Here are the contenders
Gigabyte GeForce GTX970 G1 Gaming
Zotac GeForce GTX970 AMP! Omega Core
Asus GeForce GTX980 Ti Strix OC
Asus GeForce GTX1070 Dual
As techpowerup proved that stock GTX1070 is somewhat faster than stock SLI GTX970 and stock GTX980 Ti, i have no intention of repeating that again despite a lot of different games in my rig. My goal is to satisfy the curiosity of many people who have custom factory overclocked GTX970 and GTX980 Ti video cards and wonder if it is worth the effort to replace them with the cheapest GTX1070 they can afford. I think i've said this clear enough. Sure, one can buy Asus GTX1070 Strix OC for 540 EU, which, i think would beat my Asus GTX980 Ti Strix OC, but at that price point it does not matter how much faster it would be...
For this test I’ve made two sets of benchmarks:
1. 2560x1440, all settings max, no anti-aliasing
2. 2560x1440, all settings max, custom anti-aliasing
The final average FPS performance summary will be based upon no anti-aliasing results. AA benchmark has been done additionally just to see how much stress does custom AA put on the video cards, but in many cases custom AA makes the game below enjoyable or simply unplayable.
What more to say, let's begin the show.
ALAN WAKE AMERICAN NIGHTMARE
There is no option to disable AA in this game. SLI scaling is 83 %, which just enough to lead the graph. Enabling 8X MSAA keeps results in similar fashion and the performance drop is not crucial.
ALIEN ISOLATION
SLI scaling of 80 % is just what SLI GTX970 need to confidently lead the graph. The strange SM T1X AA does not hurt performance and keeps results almost intact.
ASSASSIN'S CREED SYNDICATE
SLI scaling of 70 % is what makes the difference from non playable GTX970 to playable SLI GTX970, and that is also enough to beat GTX980 Ti and GTX1070; however, 4X MSAA bashes all the cards so hard, that you wished for GTX1080.
BATTLEFIELD 1
SLI scaling is 65 % and that is enough to lead the graph once more. AA options are limited and do not hurt performance very much, but what they do is bring down SLI GTX970 performance below GTX980 Ti and GTX1070.
CALL OF DUTY BLACK OPS 3
SLI scaling is 57 %, which is not enough to beat GTX980 Ti. Notice that the minimal FPS of SLI GTX970 tend to be low when compared to the other cards, but maximum FPS seem to be highest - this is a pattern that is common with all NVIDIA SLI cards, since the days when GF 9800GX2 VS GF GTX280 were benchmarked and you will see this pattern with SLI GTX970 in several more games. AA in this game was broken or "something like that" - i just could not get it stable with all the contenders and i will not bother to explain.
CRYSIS 3
SLI scaling is 68 %, which ties SLI GTX970 to GTX980 Ti. Low minimal FPS and high maximum FPS - a classic feature of NVIDIA SLI cards in demanding games like Crysis. Enabling 4X SMAA does not help SLI GTX970 any further as the minimal FPS is now seriously crippled.
DYING LIGHT
SLI scaling is 87 %, which is enough to lead the graph and the minimal FPS are stable too. Enabling unknown AA favors GTX980 Ti above the rest, but the general performance does not suffer much.
DOOM
Doom has got the worst SLI scaling of 32 %, so no wonder how come SLI GTX970 are in the bottom. SMAA 1TX does not hurt performance at all.
DRAGON AGE INQUISITION
SLI scaling is a massive 80 %, which is a confident win for SLI GTX970. Enabling 4X MSAA crushes the performance of all cards, but keeps the game playable.
FAR CRY 4
SLI scaling is an overwhelming 93 % - it is the difference between non playable and playable. In fact SLI works so well that GTX980 Ti and GTX1070 seem to be totally crushed in Far Cry 4. Enabling NVIDIA based 4X TXAA almost ties SLI GTX970 with GTX980 Ti and GTX1070.
LEFT 4 DEAD 2
In order to keep the contenders draw below 300 FPS i had to turn on 8X MSAA. SLI scaling is 87 %, which accounts for SLI GTX970 winning the graph. If you have plans for those 240 Hz monitors, SLI GTX970 might be a good choice to play L4D2. Even though this is a 2009 game, it was being updated till the the end of 2016 and remains an extremely popular multiplayer coop game to this day, therefore it is still relevant to be included in the benchmark.
MAD MAX FURY ROAD
SLI scaling is 89 %, therefore SLI GTX970 lead the graph. Unknown AA does not hurt performance that much.
METRO LAST LIGHT REDUX
SLI scaling of 54 % i consider good, but not enough to beat GTX1070 or GTX980 Ti. This 2014 updated game for some reason has got only one type of AA - super sampling, which is the most punishing. That being said, enabling just 2X SSAA absolutely crushes all the cards to half of their no AA performance! SLI GTX970 start showing their typical low minimal FPS and high maximum FPS pattern. But the most disappointing is GTX1070!
MIDDLE EARTH SHADOW OF MORDOR
SLI scaling is 82 % - great! Another typical picture - SLI GTX970 beat GTX980 Ti, GTX980 Ti beats GTX1070. And of course SLI GTX970 low minimal FPS and high maximum FPS.
MIRROR'S EDGE CATALYST
SLI scaling is 90 % - excellent! Another crushing victory for SLI GTX970. Enabling Post 3 AA in the config file does not hurt performance too much. There are no AA options in the game menu.
PROJECT CARS
SLI scaling is 62 %, which is just enough for SLI GTX970 to lead the graph. Enabling SMAA to high preset moderately drops FPS as well as drops SLI GTX970 performance below GTX980 Ti and GTX1070.
RISE OF TOMB RAIDER
SLI scaling is 64 %. This is a mighty victory for GTX980 Ti and the minimal FPS of SLI GTX970 look weak when compared to GTX980 Ti and GTX1070; however, the minimal FPS of SLI GTX970 are still much better than those of single GTX970. Enabling 2X SSAA ties the minimal FPS of all cards to the same level and surprisingly enough, makes SLI GTX970 the victor. Unlike in Metro Redux, 2X SSAA is not that punishing.
THIEF
This game is just not optimized properly or CPU bound - whatever! I will not use this game in my benchmarks again! SLI scaling is an underwhelming 30 %, but much worse is the fact that single GTX970 is almost as good as GTX1070... Enabling SSAA to high preset drops the performance very lightly when compared even with the Rise of Tomb Raider.
WATCH DOGS 2
SLI scaling is 52 % - not quite enough to beat GTX1070 and GTX980 Ti. Funny thing - all settings are set to the highest possible and the game demands 5500 MB VRAM, yet SLI GTX970, which have only 3500 MB GDDR5 running on 224 bit bus, and 500 MB GDDR5 running on 32 bit bus do not seem to be far away from GTX1070 8000 MB GDDR5. If SLI scaling had been anywhere as good as in Far Cry 4, i believe SLI GTX970 would crush GTX1070 even with their VRAM shortage. Enabling SMAA and 2X MSAA absolutely annihilates all the cards! So this is indeed the most lagging game i have ever tested! Only after i made the benchmarks i found out that temporal filtering greatly increases performance. It was turned off during this test.
WITCHER 3 WILD HUNT
SLI GTX970 scaling is an awesome 84 % and so SLI GTX970 outperform GTX980 Ti and GTX1070. Enabling AA, and the very demanding NVIDIA Hairworks feature with it's own AA changes the picture - now GTX980 Ti confidently leads the graph above 60 FPS with all settings maximum and Hairworks ON, high preset with 4X AA. What a beast GTX980 Ti is for this game! GTX1070 is as disappointing as ever.
----------------------------------------------------------------------------------------------------
AVERAGE FPS PERFORMANCE SUMMARY
------------------------------------------------------------------------------------------------------
So, SLI GTX970 have been honorably redeemed! SLI scaling of 70 % in 20 games seems like a good deal for 300 EU. SLI GTX970 were 9 % faster than GTX1070! Having in mind the cheapest GTX1070 is 400 EU, it's a no brainier to get SLI GTX970 instead.
When it comes to GTX980 Ti, it's obvious that even a factory overclocked video card can easily defeat a stock GTX1070 as Asus GTX980 Ti Strix OC was 6 % faster than Asus GTX1070 Dual and outperformed it in every single game. If you still have doubt in the huge potential of GTX980 Ti then watch this video, in which a custom overclocked GTX980 Ti beats a factory overclocked Gigabyte GTX1070 G1 Gaming! This makes for a conclusion that a slightly overclocked GTX980 Ti can match the stock GTX1070 and a good manual overclock on GTX980 Ti can beat the best factory overclocked GTX1070.
Finnaly, if you are looking for a summary chart to see which games support SLI and which do not, take a look at this great article that i have found:
http://www.babeltechreviews.com/sad-state-crossfire-sli-today/3/
If you are plaining a new setup, go for GTX1070 or a used GTX980 Ti. But if you already have a GTX970 and a SLI compatible motherboard, then definitely get a second GTX970!
People looking for a video card to run games at 2560X1440 res. found the right place. Perhaps some folks are looking to upgrade their GeForce GTX970 to something that can handle 120 or 144 FPS at full HD? The options are pretty obvious if we are talking about new video card from first hand retailer - GTX1070 is the card to buy right now. It's positioned at the price and performance segment that is just right for 2560x1440 res. monitors. I think most people will agree with me that GTX1080 is a bit of an overkill for 2560x1440 res. monitors considering the premium price and that it is a 4k ready video card. I expect most of you agree that GTX1060 6 GB is too weak for some of the more demanding titles at 2560x1440 res. While AMD has R9 Fury X and R9 295X2 as alternatives for 2560X1440 res. gaming, as a G-SYNC monitor owner i will stick to NVIDIA this time.
The primary point of this thread is to prove that SLI GTX970 can beat GTX1070 costing much less. Everything eventually comes to the price. I was intrigued to do this test, because recently one of the local retail shops in my country started selling brand new Zotac GTX970 AMP! Omega Core video cards for 190 EU, which is 30 EU cheaper than the cheapest GTX1060 3 GB you could get. This is a great value since i have proven that GTX970 is equal to GTX1060 3 GB in this thread. Having this in mind an idea poped: what if SLI GTX970 could beat the cheapest GTX1070 you could find? The cheapest stock GTX1070 with a custom cooler costs 405 EU, but the fact of the matter is that you can buy two used custom GTX970 OC video cards for 300 EU this time around from second hand retailers; and so i bought two GTX970 video cards for 300 EU - Zotac GTX970 AMP! Omega Core and Gigabyte GTX970 G1 Gaming, as well Asus GTX1070 Dual, that costs 445 EU. I have slightly overclocked the Zotac GTX970 AMP! Omega Core to equally match the Gigabyte GTX970 G1 Gaming, so in practice for this benchmark i have two GTX970 G1 Gaming cards.
The second objective of this thread is to prove that factory overclocked GTX980 Ti can beat or at the very least tie a stock GTX1070. This itchy idea came from the original techpowerup GTX1070 review, in which a stock GTX1070 proved to be 13,5 % faster than a stock GTX980 Ti. I was, frankly speaking, displeased, and thought no way the stock GTX1070 can beat my Asus GTX980 Ti Strix OC, and that comes to play today. I could find a used custom GTX980 Ti for 300 EU these days and that means it competes directly with two GTX970 cards to dethrone the more expensive GTX1070.
For this test i have selected 25 games, but i found out that 5 of them were not supporting SLI at all: Company of Heroes 2, Call of Duty Infinite Warfare, Arma 3 Apex, Ashes of the Singularity (DX11) and Rainbow Six Siege, therefore to be FAIR, i, obviously have not included them in this benchmark and only included games, supporting at least some noticeable scaling.
BENCHMARK HARDWARE & DRIVERS
Intel Core i7 5775C 4 GHz
Gigabyte Z97X-Gaming 3
Crucial Ballistix Tactical 2X8 GB DDR3 1600 MHz C8
Samsung Evo 500 GB SSD (all games are installed in it)
Windows 7 Pro X64 (updates installed)
NVIDIA Forceware 378.66
I tested all the games with Corsair Vengeance Pro 2400 MHz C11 DDR3 and found no difference in performance, so don't talk about "low RAM speed BS please".
I had a lot of trouble finishing this benchmark at the planned time (a month ago) and i could not update the drivers to newest as something that worked with one card did not work with another later, so by the time i fixed the problem new drivers would appear and i had no time to redo all the stuff again. Please understand that.
Here are the contenders
Gigabyte GeForce GTX970 G1 Gaming
Zotac GeForce GTX970 AMP! Omega Core
Asus GeForce GTX980 Ti Strix OC
Asus GeForce GTX1070 Dual
As techpowerup proved that stock GTX1070 is somewhat faster than stock SLI GTX970 and stock GTX980 Ti, i have no intention of repeating that again despite a lot of different games in my rig. My goal is to satisfy the curiosity of many people who have custom factory overclocked GTX970 and GTX980 Ti video cards and wonder if it is worth the effort to replace them with the cheapest GTX1070 they can afford. I think i've said this clear enough. Sure, one can buy Asus GTX1070 Strix OC for 540 EU, which, i think would beat my Asus GTX980 Ti Strix OC, but at that price point it does not matter how much faster it would be...
For this test I’ve made two sets of benchmarks:
1. 2560x1440, all settings max, no anti-aliasing
2. 2560x1440, all settings max, custom anti-aliasing
The final average FPS performance summary will be based upon no anti-aliasing results. AA benchmark has been done additionally just to see how much stress does custom AA put on the video cards, but in many cases custom AA makes the game below enjoyable or simply unplayable.
What more to say, let's begin the show.
ALAN WAKE AMERICAN NIGHTMARE
There is no option to disable AA in this game. SLI scaling is 83 %, which just enough to lead the graph. Enabling 8X MSAA keeps results in similar fashion and the performance drop is not crucial.
ALIEN ISOLATION
SLI scaling of 80 % is just what SLI GTX970 need to confidently lead the graph. The strange SM T1X AA does not hurt performance and keeps results almost intact.
ASSASSIN'S CREED SYNDICATE
SLI scaling of 70 % is what makes the difference from non playable GTX970 to playable SLI GTX970, and that is also enough to beat GTX980 Ti and GTX1070; however, 4X MSAA bashes all the cards so hard, that you wished for GTX1080.
BATTLEFIELD 1
SLI scaling is 65 % and that is enough to lead the graph once more. AA options are limited and do not hurt performance very much, but what they do is bring down SLI GTX970 performance below GTX980 Ti and GTX1070.
CALL OF DUTY BLACK OPS 3
SLI scaling is 57 %, which is not enough to beat GTX980 Ti. Notice that the minimal FPS of SLI GTX970 tend to be low when compared to the other cards, but maximum FPS seem to be highest - this is a pattern that is common with all NVIDIA SLI cards, since the days when GF 9800GX2 VS GF GTX280 were benchmarked and you will see this pattern with SLI GTX970 in several more games. AA in this game was broken or "something like that" - i just could not get it stable with all the contenders and i will not bother to explain.
CRYSIS 3
SLI scaling is 68 %, which ties SLI GTX970 to GTX980 Ti. Low minimal FPS and high maximum FPS - a classic feature of NVIDIA SLI cards in demanding games like Crysis. Enabling 4X SMAA does not help SLI GTX970 any further as the minimal FPS is now seriously crippled.
DYING LIGHT
SLI scaling is 87 %, which is enough to lead the graph and the minimal FPS are stable too. Enabling unknown AA favors GTX980 Ti above the rest, but the general performance does not suffer much.
DOOM
Doom has got the worst SLI scaling of 32 %, so no wonder how come SLI GTX970 are in the bottom. SMAA 1TX does not hurt performance at all.
DRAGON AGE INQUISITION
SLI scaling is a massive 80 %, which is a confident win for SLI GTX970. Enabling 4X MSAA crushes the performance of all cards, but keeps the game playable.
FAR CRY 4
SLI scaling is an overwhelming 93 % - it is the difference between non playable and playable. In fact SLI works so well that GTX980 Ti and GTX1070 seem to be totally crushed in Far Cry 4. Enabling NVIDIA based 4X TXAA almost ties SLI GTX970 with GTX980 Ti and GTX1070.
LEFT 4 DEAD 2
In order to keep the contenders draw below 300 FPS i had to turn on 8X MSAA. SLI scaling is 87 %, which accounts for SLI GTX970 winning the graph. If you have plans for those 240 Hz monitors, SLI GTX970 might be a good choice to play L4D2. Even though this is a 2009 game, it was being updated till the the end of 2016 and remains an extremely popular multiplayer coop game to this day, therefore it is still relevant to be included in the benchmark.
MAD MAX FURY ROAD
SLI scaling is 89 %, therefore SLI GTX970 lead the graph. Unknown AA does not hurt performance that much.
METRO LAST LIGHT REDUX
SLI scaling of 54 % i consider good, but not enough to beat GTX1070 or GTX980 Ti. This 2014 updated game for some reason has got only one type of AA - super sampling, which is the most punishing. That being said, enabling just 2X SSAA absolutely crushes all the cards to half of their no AA performance! SLI GTX970 start showing their typical low minimal FPS and high maximum FPS pattern. But the most disappointing is GTX1070!
MIDDLE EARTH SHADOW OF MORDOR
SLI scaling is 82 % - great! Another typical picture - SLI GTX970 beat GTX980 Ti, GTX980 Ti beats GTX1070. And of course SLI GTX970 low minimal FPS and high maximum FPS.
MIRROR'S EDGE CATALYST
SLI scaling is 90 % - excellent! Another crushing victory for SLI GTX970. Enabling Post 3 AA in the config file does not hurt performance too much. There are no AA options in the game menu.
PROJECT CARS
SLI scaling is 62 %, which is just enough for SLI GTX970 to lead the graph. Enabling SMAA to high preset moderately drops FPS as well as drops SLI GTX970 performance below GTX980 Ti and GTX1070.
RISE OF TOMB RAIDER
SLI scaling is 64 %. This is a mighty victory for GTX980 Ti and the minimal FPS of SLI GTX970 look weak when compared to GTX980 Ti and GTX1070; however, the minimal FPS of SLI GTX970 are still much better than those of single GTX970. Enabling 2X SSAA ties the minimal FPS of all cards to the same level and surprisingly enough, makes SLI GTX970 the victor. Unlike in Metro Redux, 2X SSAA is not that punishing.
THIEF
This game is just not optimized properly or CPU bound - whatever! I will not use this game in my benchmarks again! SLI scaling is an underwhelming 30 %, but much worse is the fact that single GTX970 is almost as good as GTX1070... Enabling SSAA to high preset drops the performance very lightly when compared even with the Rise of Tomb Raider.
WATCH DOGS 2
SLI scaling is 52 % - not quite enough to beat GTX1070 and GTX980 Ti. Funny thing - all settings are set to the highest possible and the game demands 5500 MB VRAM, yet SLI GTX970, which have only 3500 MB GDDR5 running on 224 bit bus, and 500 MB GDDR5 running on 32 bit bus do not seem to be far away from GTX1070 8000 MB GDDR5. If SLI scaling had been anywhere as good as in Far Cry 4, i believe SLI GTX970 would crush GTX1070 even with their VRAM shortage. Enabling SMAA and 2X MSAA absolutely annihilates all the cards! So this is indeed the most lagging game i have ever tested! Only after i made the benchmarks i found out that temporal filtering greatly increases performance. It was turned off during this test.
WITCHER 3 WILD HUNT
SLI GTX970 scaling is an awesome 84 % and so SLI GTX970 outperform GTX980 Ti and GTX1070. Enabling AA, and the very demanding NVIDIA Hairworks feature with it's own AA changes the picture - now GTX980 Ti confidently leads the graph above 60 FPS with all settings maximum and Hairworks ON, high preset with 4X AA. What a beast GTX980 Ti is for this game! GTX1070 is as disappointing as ever.
----------------------------------------------------------------------------------------------------
AVERAGE FPS PERFORMANCE SUMMARY
------------------------------------------------------------------------------------------------------
So, SLI GTX970 have been honorably redeemed! SLI scaling of 70 % in 20 games seems like a good deal for 300 EU. SLI GTX970 were 9 % faster than GTX1070! Having in mind the cheapest GTX1070 is 400 EU, it's a no brainier to get SLI GTX970 instead.
When it comes to GTX980 Ti, it's obvious that even a factory overclocked video card can easily defeat a stock GTX1070 as Asus GTX980 Ti Strix OC was 6 % faster than Asus GTX1070 Dual and outperformed it in every single game. If you still have doubt in the huge potential of GTX980 Ti then watch this video, in which a custom overclocked GTX980 Ti beats a factory overclocked Gigabyte GTX1070 G1 Gaming! This makes for a conclusion that a slightly overclocked GTX980 Ti can match the stock GTX1070 and a good manual overclock on GTX980 Ti can beat the best factory overclocked GTX1070.
Finnaly, if you are looking for a summary chart to see which games support SLI and which do not, take a look at this great article that i have found:
http://www.babeltechreviews.com/sad-state-crossfire-sli-today/3/
If you are plaining a new setup, go for GTX1070 or a used GTX980 Ti. But if you already have a GTX970 and a SLI compatible motherboard, then definitely get a second GTX970!
Last edited:
