Upgrade to 1080? Would it be worth it?

[CAPSLOCKSTUCK]

fair enough.

 

[slozomby]

I do not fully agree with you, I had a better score than someone with an overclocked i5 6600K.

@slozomby Edit:

With i5 6600K OC'd to 4.4GHz and 1070 OC'd:

My score with i5 6500:
View attachment 80958

and here's my 1080 using the extreme hd preset of the above 2 benchmarks
stil using cpu@4.6 and gpu @1987

 

[P4-630]

and here's my 1080 using the extreme hd preset of the above 2 benchmarks
stil using cpu@4.6 and gpu @1987
View attachment 80962

What was that about:

both bad examples as CPU clock speeds will limit both cards horribly.

Including your own CPU then since your score is not extremely higher than my GTX1070....

 

[slozomby]

What was that about:


Including your own CPU then since your score is not extremely higher than my GTX1070....

-please explain how your 6500 is clocked up near 4.6

the original screenshots that I had talked about were running nehalem xeons and a 1080. theres a reason I upgraded from my Nehalem chip. because ipc growth in the last 8 years warranted it. you put forth that it was not limited by cpu. I put forth that the 1080 would perform much better than the screenshots you were showing with a high clock speed on a recent processor.

a 15% performance difference is significant. especially when running my folding profile. if you wanted to argue that the 1070 was better price/performance I would happily agree.

I know my cpu will cap out at 162 fps in valley. it does that with 0aa enabled. sli 1080's cap out at right around the same place with the 6700k at 4.7. I've show in the valley thread what effects cpu and gpu clock speed have on performance. as well as the gene


feel free to debate that in the proper thread.
https://www.techpowerup.com/forums/...getting-an-upgrade.226448/page-5#post-3543712

here's an example of going from a 6500 to a 6700k
https://www.techpowerup.com/forums/threads/unigine-valley-benchmark-is-getting-an-upgrade.226448/page-6#post-3545440

 

[headik]

nope.. it ain't happening. look at tpu's own. No way a 'slight' overclock gets it there...

... now if you are maxing out a 1070, it can reach a reference clocked 1080...don't forget, a 1080 can overclock as well and get that 20% right back.

https://www.techpowerup.com/reviews/MSI/GTX_1070_Gaming_Z/29.html

https://www.techpowerup.com/reviews/ASUS/GTX_1080_STRIX/27.html

No GTX1070 cant reach stock GTX1080 performance.Techpower up is just wrong.You should look at some youtube videos where people compare 1070 and 1080 at same clock.There is pretty huge gap between them 25-33%.There are even some videos with 1070 oc vs 1080 stock and gap is still 10-15%.

 

[EarthDog]

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.

 

[rtwjunkie]

No GTX1070 cant reach stock GTX1080 performance.Techpower up is just wrong.

W1zz is just flat out wrong, huh?

 

[headik]

He tested it in archaic old BF3 game.In recent AAA games the gap is 25-33% and you cant reach GTX1080 performace as i show you in those videos.
Here is more

 

[rtwjunkie]

He tested it in archaic old BF3 game.In recent AAA games the gap is 25-33% and you cant reach GTX1080 performace as i show you in those videos.

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

 

[EarthDog]

I agree that is an older title to check against (seriously Wiz, BF3??????), but your differences between the two cards are there around 20%. YOur videos at least the 1070 vs 1080 one, doens't show a REFERENCE 1080 versus a maxed out 1070, do they?????????????????????????????

Aren't we seeing the 1070 clock matching the 1080? That tests IPC, but not much else.......

 

[headik]

I agree that is an older title to check against (seriously Wiz, BF3??????), but your differences between the two cards are there around 20%. YOur videos at least the 1070 vs 1080 one, doens't show a REFERENCE 1080 versus a maxed out 1070, do they?????????????????????????????

Aren't we seeing the 1070 clock matching the 1080? That tests IPC, but not much else.......

my videos shows 1070 vs 1080 at same clock and gap.Gap is 25-33% in almost all games.Some videos are 1070 OC vs 1080 stock just like this:

and this

and this.....

As you can see overclocked gtx1070 cant reach stock GTX1080.

 

[EarthDog]

Not sure what to tell you there bud.. I didn't make those vids...One thing I did notice in the vid is some setting must be different between them because memory use is a lot higher on the 1070 in TW3.

I can tell you from my testing in more modern titles, that a maxed out 1070 can reach a reference 1080...

 

[headik]

Not sure what to tell you there bud.. I didn't make those vids...One thing I did notice in the vid is some setting must be different between them because memory use is a lot higher on the 1070 in TW3.

I can tell you from my testing in more modern titles, that a maxed out 1070 can reach a reference 1080...

Then show me your results.I have GTX1070 2140/9400 and it cant reach stock GTX1080 i can tell you.
I mean in new AAA games not some benchmarks like heaven or valley.Just look at dishonored 2 video.The gap is just soo huge.There is no way you can oc GTX1070 for +30%

 

[EarthDog]

I don't have a 1080 anymore... sorry.

 

[P4-630]

This is starting to give me a "Headik"....

 

[64K]

Well, if it's on Youtube then it must be true. They can't put horse shit on Youtube.

 

[headik]

Well, if it's on Youtube then it must be true. They can't put horse shit on Youtube unless it's true.

Still 100x better proof than one test in 5years old BF3.Anyway he testing alot of cards.I am sure those videos are legit.

 

[EarthDog]

Is it possible driver revisions helped the 1080 gain more ground as well?

I agree whole heartedly that BF3 shouldn't be used... maybe that is the problem here as well...

But I recall in my testing when these cards came out, that the 1070 Gaming X met the 1080 I had been using for a comparison. The difference then was around 20% between FE 1070 and FE 1080. Add an factory overclocked card and overclock it to the max, its right there... at least was at some point.

As I said, there is SOMETHING wrong with TW3 results as the memory on the 1070 was nearly 1GB higher.

EDIT: I am also wondering how in the hell BF1 uses so much vRAM at 2560xc1440. I run that res with ultra settings on a 1070 WITH 140% resolution scale, and barely break 4GB...................how are they hitting 6GB+??????????? Oh.. .I see they run DX12, which is for all intents and purposes, broken for NVIDIA last I heard............

So.... yeah, something is up there with those two vids IMO..................

 

[headik]

Is it possible driver revisions helped the 1080 gain more ground as well?

I agree whole heartedly that BF3 shouldn't be used... maybe that is the problem here as well...

But I recall in my testing when these cards came out, that the 1070 Gaming X met the 1080 I had been using for a comparison. The difference then was around 20% between FE 1070 and FE 1080. Add an factory overclocked card and overclock it to the max, its right there... at least was at some point.

As I said, there is SOMETHING wrong with TW3 results as the memory on the 1070 was nearly 1GB higher.

1GB higher?I see there 100-140mb more..Btw in what games and resolution did you test it?The gap is bigger in higher resolutions.

 

[EarthDog]

1080p/2560x1440. Yes, aware 1080p is pretty CPU bound with these cards...

As far as the vRAM, again, TW3 your last video, shows a HUGE difference in vRAM used......1.3GB worth and from the start of the video........not sure how you are missing that.

Also note I edited more before you posted..

 

[headik]

Vram usage is in those 3 videos within 200MB.Look at last line MEM first is clock and second Vram usage.

 

[EarthDog]

Oops, reading the wrong line... system ram is well over 1.3GB off in that title...

Seeing as how DX12 doesn't perform well for NVIDIA at this time, I am going to discount that test for BF1.

I see what you are saying, but, I have have faith that whatever W1z's results were, at the very minimum at that time, accurate. Perhaps that has changed with drivers, and titles. I do agree BF3 is WAY too old to be used.

 

[Grings]

some situations are not making use of every last shader unit, so a 1070 can be faster if its clocks are higher

anything making use of the cards resources to the maximum the 1080 will be faster

 

[Frag_Maniac]

I really think once the 1080 Ti comes out, the 1070 will still be popular for 1080p, while the 1080 Ti will be the new 4k card, and the 1080 will sit in no man's land and soon be forgotten. Yeah I know a 1080 can play a lot of games in 4k at decent frame rates, but many it can't, and players into 4k and high end GPUs care about that.

 

[Grings]

I really think once the 1080 Ti comes out, the 1070 will still be popular for 1080p, while the 1080 Ti will be the new 4k card, and the 1080 will sit in no man's land and soon be forgotten. Yeah I know a 1080 can play a lot of games in 4k at decent frame rates, but many it can't, and players into 4k and high end GPUs care about that.

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?