Upgrade to 1080? Would it be worth it?

[Recon-UK]

I trust W1z and his GPU reviews over 99% of the muppets out on YT bud... as we said and results at other places show, a 1070 overclocked to the max can reach a reference clocked 1080.

Now, we are aware you can overclock the 1080 too (I think I said that), but that also isn't the point.

W1zz is just flat out wrong, huh?

Well, you can become a respected reviewer and then start your own website and use what methods you prefer.

Ignore the troll.

And as a side note, every gen of GPU has been able to have the second from top card able to OC to it's performance, even beating it.

GTX 280 - 285
GTX 470 - 480
GTX 480 - 580
GTX 570 - 580
GTX 670 more or less equal to 680 anyway -680
680 - 770

The list goes on.

 

[Frag_Maniac]

Is this because once the 980 Ti came out, the 970 was still popular for 1080p, While the 980 Ti was the new 2k card, and the 980 sat in no mans land and was soon forgotten?

It has all got a bit samey hasnt it?

Yeah it has, but I do think 4k has changed it some. People have been waiting a long time for a legit single card solution for it, and if the 1080 Ti is it, the dust should settle some, and we can all go back to more traditionally slotted target niches for GPUs.

 

[Grings]

4k is just where we should be if you look at how res has gone up over the years, its just not quite mainstream yet because panels and gpu's cost too much

This also supports the theory that the 1080 (and 980 before it) are midrange cards, 256bit memory bus and all, being sold as high end and not quite up to the task of running max at (what should be) the standard optimal gaming res

 

[slozomby]

not quite up to the task of running max at (what should be) the standard optimal gaming res

they run ultra wide 1440p just fine.

 

[Vayra86]

4k is just where we should be if you look at how res has gone up over the years, its just not quite mainstream yet because panels and gpu's cost too much

This also supports the theory that the 1080 (and 980 before it) are midrange cards, 256bit memory bus and all, being sold as high end and not quite up to the task of running max at (what should be) the standard optimal gaming res

I beg to differ. Resolution going up isn't really what's been going on the past decade or so. CRT already had 1600x1200. What goes up is graphical quality and detail. Resolution is on the sideline, and it is generally the wisest decision to have mainstream resolutions. Doing otherwise was always problematic and/or needed special care. If any res is going mainstream for PC I would say right now it's 1440p. Comfortable to push, within a reasonable PPI for the view distance on PC monitors, and a noticeable step up the ladder, and also within current spec of the majority of display interfaces. 4K is definitely coming, but I'd give it 2-3 years more to say mainstream.

 

[EarthDog]

I assume you mean 2560 x 1440? 1440p can mean a couple of resolutions.

 

[Frag_Maniac]

4k is just where we should be if you look at how res has gone up over the years, its just not quite mainstream yet because panels and gpu's cost too much

This also supports the theory that the 1080 (and 980 before it) are midrange cards, 256bit memory bus and all, being sold as high end and not quite up to the task of running max at (what should be) the standard optimal gaming res

I'm just saying you don't see resolution standards quadruple in pixels very often, and the 1080 is on the cusp of being in no man's land between 4k and 1080p, whereas normally it wouldn't suffer any consequences. Once 4k settles in, you won't see the same scenario for some time.

 

[slozomby]

I'm just saying you don't see resolution standards quadruple in pixels very often, and the 1080 is on the cusp of being in no man's land between 4k and 1080p, whereas normally it wouldn't suffer any consequences. Once 4k settles in, you won't see the same scenario for some time.

except they've been talking about 8k TV's for a couple of years now.

 

[headik]

Ignore the troll.

And as a side note, every gen of GPU has been able to have the second from top card able to OC to it's performance, even beating it.

GTX 280 - 285
GTX 470 - 480
GTX 480 - 580
GTX 570 - 580
GTX 670 more or less equal to 680 anyway -680
680 - 770

The list goes on.

Yeah call me troll and dont provide any proof at all.Just great.
Did you look at Specs?
GTX1070 is more cutdown than any previous card in past.
GTX1080 have 33% more SP and 25% memory bandwidth.
GTX1070 also have disabled ONE GPC cluster.So all things in that GPC are disabled like rasterizer-you know that thing that have huge impact on pixel fillrate/triangle rate.GTX1070 have rating only 3triangles/s with 3x GPC.GTX1080 have rating 4x triangles/s with 4xGPC.GPC cluster have HUGE impact on performance.

GTX1080 vs 1070
33%more SP
25%more memory bandwidth
+1 GPC

GTX980 vs 970
23% more SP
14% more memory bandwidth(gtx970 is 224bit card)
Both have 4x GPC

GTX680 vs 670
14% more SP
same memory bandwidth
Same number of GPC

GTX580 vs 570
6.6% more Sp
26% more memory bandwidth
Same number of GPC

GTX480 vs 470
7% more SP
32% more memory bandwidth
same number of GPU

As you can see GTX1070 is more cutdown than any of these cards.That disabled GPC cluster make it even far worse than "normal" cutdown.

 

[Vayra86]

Yeah call me troll and dont provide any proof at all.Just great.
Did you look at Specs?
GTX1070 is more cutdown than any previous card in past.
GTX1080 have 33% more SP and 25% memory bandwidth.
GTX1070 also have disabled ONE GPC cluster.So all things in that GPC are disabled like rasterizer-you know that thing that have huge impact on pixel fillrate/triangle rate.GTX1070 have rating only 3triangles/s with 3x GPC.GTX1080 have rating 4x triangles/s with 4xGPC.GPC cluster have HUGE impact on performance.

GTX1080 vs 1070
33%more SP
25%more memory bandwidth
+1 GPC

GTX980 vs 970
23% more SP
14% more memory bandwidth(gtx970 is 224bit card)
Both have 4x GPC

GTX680 vs 670
14% more SP
same memory bandwidth
Same number of GPC

GTX580 vs 570
6.6% more Sp
26% more memory bandwidth
Same number of GPC

GTX480 vs 470
7% more SP
32% more memory bandwidth
same number of GPU

As you can see GTX1070 is more cutdown than any of these cards.That disabled GPC cluster make it even far worse than "normal" cutdown.

Important omission

64 ROPs versus 64 ROPs between 1070 and 1080.

Combined with a slight VRAM bottleneck on the 1070 (memory OC is more useful than core OC and the core auto-OC's to within 3-10% of max clocks), it is quite easy to see why in certain games and especially at lower resolutions (!) the 1070 can equal a 1080. This effect is negated once you move to 1440p > 4K where the 30% advantage of the 1080 starts to show. At the same time, the 1070 and the 1080 suffer from a CPU bottleneck in many cases at 1080p resolution because they crank out over 100 fps. This will make them end up closer to each other still.

You can drop it now, trust us.

 

[Recon-UK]

Yeah call me troll and dont provide any proof at all.Just great.
Did you look at Specs?
GTX1070 is more cutdown than any previous card in past.
GTX1080 have 33% more SP and 25% memory bandwidth.
GTX1070 also have disabled ONE GPC cluster.So all things in that GPC are disabled like rasterizer-you know that thing that have huge impact on pixel fillrate/triangle rate.GTX1070 have rating only 3triangles/s with 3x GPC.GTX1080 have rating 4x triangles/s with 4xGPC.GPC cluster have HUGE impact on performance.

GTX1080 vs 1070
33%more SP
25%more memory bandwidth
+1 GPC

GTX980 vs 970
23% more SP
14% more memory bandwidth(gtx970 is 224bit card)
Both have 4x GPC

GTX680 vs 670
14% more SP
same memory bandwidth
Same number of GPC

GTX580 vs 570
6.6% more Sp
26% more memory bandwidth
Same number of GPC

GTX480 vs 470
7% more SP
32% more memory bandwidth
same number of GPU

As you can see GTX1070 is more cutdown than any of these cards.That disabled GPC cluster make it even far worse than "normal" cutdown.

GPC / GPU???

GTX 470 for instance was a big cut down from the 480 including memory bus and framebuffer size. It took a very big OC to match the 480, but it could be done.

 

[headik]

Important omission

64 ROPs versus 64 ROPs between 1070 and 1080.

Combined with a slight VRAM bottleneck on the 1070 (memory OC is more useful than core OC and the core auto-OC's to within 3-10% of max clocks), it is quite easy to see why in certain games and especially at lower resolutions (!) the 1070 can equal a 1080. This effect is negated once you move to 1440p > 4K where the 30% advantage of the 1080 starts to show. At the same time, the 1070 and the 1080 suffer from a CPU bottleneck in many cases at 1080p resolution because they crank out over 100 fps. This will make them end up closer to each other still.

You can drop it now, trust us.

Do you have both cards?I have 1070 2140/9400 and it cant match stock GTX1080 in 1440P.I also linked alot of videos in 1440P whe you can see that 1070 at 2050-2080/9400 cant reach stock GTX1080 performance.(still around 10% slower.In BF1 the gap is even bigger 10-15%)Why the hell you guys still saying GTX1070 can match stock GTX1080?

 

[headik]

GPC / GPU???

GTX 470 for instance was a big cut down from the 480 including memory bus and framebuffer size. It took a very big OC to match the 480, but it could be done.

Yeah GPC.There are 4x GPC in full GP104.GTX1080 have all 4 enabled.GTX1070 have one disabled.
This is GTX1070 diagram:

 

[Recon-UK]

Yeah GPC.There are 4x GPC in full GP104.GTX1080 have all 4 enabled.GTX1070 have one disabled.
This is GTX1070 diagram:

The tests conducted by the very best benchmarker out there from this site are legit, and i take them with more evidence than some random YT video's.

If you were trying to show numbers based on YOUR platform, that's fine, but trying to mock up results with different systems? come back when your head is out of the gutter mate, you need to be far better set up than you are right now, hard data, hard graphs, use a method that cannot be twisted to your liking.

 

[headik]

The tests conducted by the very best benchmarker out there from this site are legit, and i take them with more evidence than some random YT video's.

If you were trying to show numbers based on YOUR platform, that's fine, but trying to mock up results with different systems? come back when your head is out of the gutter mate, you need to be far better set up than you are right now, hard data, hard graphs, use a method that cannot be twisted to your liking.

What test??That test in 5 or 6 years old BF3 game?Really?
Btw GTX1070 max OC is even slower than max OC 980TI in most games.
That mean 980TI OC is even faster than stock GTX1080?(by your logic)

 

[Recon-UK]

What test??That test in 5 or 6 years old BF3 game?Really?

BF3 is still GPU heavy, it's a good game to base it off as it shows GPU overclocks nicely, it did back when i was on a GTX 480, a clock bump from stock to 920mhz netted me a massive 19FPS increase.

 

[headik]

BF3 is still GPU heavy, it's a good game to base it off as it shows GPU overclocks nicely, it did back when i was on a GTX 480, a clock bump from stock to 920mhz netted me a massive 19FPS increase.

well..Nope
BF3 is outdated crap.Look at Bf1 results.
1070 2080/9400 is 10-15% slower than stock GTX1080.

 

[EarthDog]

BF3 is NOT really the best test for modern GPUs... that needs to go...

 

[Recon-UK]

Looking around a bit more and i can see even in GTA V a OC'd 2090mhz 1070 only matches the 1080 FE edition, and that's at 4k.

Don't know why Nvidia cut it down so much but the performance is still great, 5-6FPS difference is not that huge.

 

[EarthDog]

Looking around a bit more and i can see even in GTA V a OC'd 2090mhz 1070 only matches the 1080 FE edition, and that's at 4k.

That is what we(I) are saying, in some titles, even modern titles, it can make up that 20% difference.

Though you may want to provide a link for Uncle Headik.

 

[headik]

You guys only talks, but no one post any proof at all.Show me some results with GTX1070 max OC vs stock GTX1080 in 1440P in lastest games where GTX1070 can match GTX1080 performance..I posted already many proofs that it cant do that.

Here is another proof in latest COD

Look at that video at GTX1080 stock performance and gap vs stock 1070.I have this game and i tested it in exactly same settings with my GTX1070 2114/9400 and in same test scenes.As you can see its still slower than stock GTX1080.Its way faster than his stock GTX1070, but its slower than stock GTX1080.The gap is just too big.You cant OC GTX1070 to GTX1080 level.

https://abload.de/img/iw7_ship_2016_11_06_1eklod.png
https://abload.de/img/iw7_ship_2016_11_06_1eklod.png
https://abload.de/img/iw7_ship_2016_11_06_1o5xk7.png
https://abload.de/img/iw7_ship_2016_11_06_1nslfi.png
https://abload.de/img/iw7_ship_2016_11_06_10mudv.png
https://abload.de/img/iw7_ship_2016_11_06_12bu9i.png
https://abload.de/img/iw7_ship_2016_11_06_1eklod.png
https://abload.de/img/iw7_ship_2016_11_06_1o5xk7.png
https://abload.de/img/iw7_ship_2016_11_06_1nslfi.png
https://abload.de/img/iw7_ship_2016_11_06_10mudv.png
https://abload.de/img/iw7_ship_2016_11_06_12bu9i.png

 

[Recon-UK]

This guy still going? lol

 

[headik]

This guy still going? lol

We both know i have truth.Show some proof or dont post/troll here.

 

[Recon-UK]

We both know i have truth.Show some proof or dont post/troll here.

Do you read anything? LOL

 

[Vayra86]

We both know i have truth.Show some proof or dont post/troll here.

64 ROPs versus 64 ROPs between 1070 and 1080.

Combined with a slight VRAM bottleneck on the 1070 (memory OC is more useful than core OC and the core auto-OC's to within 3-10% of max clocks), it is quite easy to see why in certain games and especially at lower resolutions (!) the 1070 can equal a 1080. This effect is negated once you move to 1440p > 4K where the 30% advantage of the 1080 starts to show. At the same time, the 1070 and the 1080 suffer from a CPU bottleneck in many cases at 1080p resolution because they crank out over 100 fps. This will make them end up closer to each other still.

You can drop it now, trust us.

@headik if you consider responding to this post. Please don't, just repeat that last line of my post in your head and re-read the bold and underlined parts another two dozen times.

We've already confirmed your statements and responded to them. The bottom line is: Youtube is not, will not, and has never been a trustworthy source. If my explanation above isn't enough for you, the only possible conclusion here is you have a *very* thick skull and serious problems with reading comprehension.

Enjoy your day.