CPU Reviews 720p/1080p/4K or 1080p/1440p/4K?

[Caring1]

TPU members have spoken and voted! It seems overly clear that 720p should be dropped if you really want/need to drop something...

Or it clearly shows a majority don't know what they are talking about.

 

[bug]

TPU members have spoken and voted! It seems overly clear that 720p should be dropped if you really want/need to drop something...

View attachment 87717

Good thing we didn't have on 8k option in there. That would have got even more votes

 

[EarthDog]

Final take...

WHO CARES what a CPU does at 720p. It is NOT indicative of 1080p or higher! The ONLY information which comes out it that testing is how these cpus perform with a 1080 at 720p. Are we catering to the people that buy such a lopsided system, or are we looking to see how cpus perform at resolutions we play at with gpus appropriate for the resolution?? I can pull a resolution out of my arse and strap a 1080ti on it and get a dataset but.....

...what would it mean, really???? Nothing as the dataset doesnt scale up..useless except for the few knuckleheads that have a gtx 1080 and play at 720p. Contrary to what i saw posted earlier, it doesnt show 'what a cpu would do' as that same cpu 'wouldnt do' the same at higher res. Useless dataset except for owners of a lopsided rig.

Can someone explain what use that dataset gives others?

 

[bug]

Final take...

WHO CARES what a CPU does at 720p. It is NOT indicative of 1080p or higher! The ONLY information which comes out it that testing is how these cpus perform with a 1080 at 720p.

Omg. Testing at 720 will show you which CPU has more HP to push a video card. I thought that was beyond obvious.
We're still talking about CPU benchmarks, aren't we?

 

[EarthDog]

No, it wont... that is exactly my point. It doesnt tell you squat, except for at 720p as results do not scale and will be different at the higjer resolutions. You are in essence using irrelevant data unless you are one of the knuckleheads that have a 1080 and play at 720p.

So, its ok to use a completely unrealistic testing environment (720p and 1080) to get a dataset that cannot extrapolate its results to any other res because its such a lopsided and unrealistic system? Testing there is fabricating testing for a specific result, which it doesnt even show anything outside of that environment.

Gotcha. Dont agree..at all...but understand.

Answer this. What good is that dataset for those on 1080p or higher? Why cater to such an extreme minority (720p and 1080+), particularly when all that testing does is magnify results...?!?! In other words, what will the 19 people wanting this testing going to do with that singular result??? What do 720p results tell us when we play at 1080? 1440p? 4k uhd? Why include it when it literally tells us nothing useful. Its a random fact which a minority of people believe it tells more of a story than it does.


Edit: what would be great to see are cpus all tested at the same clockspeed as well as stock. This would truly show the differences between each cpu showing what IPC, arch, and cores can do to a specific title/res.

 

[bug]

What good is measuring the HP of a car engine? It's done under completely unrealistic conditions that you'll never see in traffic.

 

[EarthDog]

What good is measuring the HP of a car engine? It's done under completely unrealistic conditions that you'll never see in traffic.

Nice try...

But the car will make the same amount of horsepower on the road as it does with a dyno....same engine. I can mash the gas on the road and get all the ponies going.

 

[64K]

Well, I will throw my two cents in one more time since us 720p backers are taking a beating. What 720p results tell us that are gaming at 1080p and above is just how many FPS we can achieve with the reviewed CPU providing we also buy enough GPU to reach that goal. It's mostly for people that want very high FPS but it's also nice for determining what the reviewed CPU is actually capable of achieving in a game. We have many, many reviews of GPUs that tell us what we can expect from a GPU at a given resolution.

Edit: One example might be that a CPU with a GTX 1080 benched at 85 FPS on 1080p. Would that CPU get them to 120 FPS if they paid a lot more for a 1080 Ti or is the CPU already at it's limit of performance and more GPU would just be a waste.

 

[Outback Bronze]

Voted 1080p/1440p/4k

Don't care what 720p does even if it does show more cpu utilization with certain cpu's.

 

[EarthDog]

Well, I will throw my two cents in one more time since us 720p backers are taking a beating. What 720p results tell us that are gaming at 1080p and above is just how many FPS we can achieve with the reviewed CPU providing we also buy enough GPU to reach that goal. It's mostly for people that want very high FPS but it's also nice for determining what the reviewed CPU is actually capable of achieving in a game. We have many, many reviews of GPUs that tell us what we can expect from a GPU at a given resolution.

but, it doesnt tell you that...its what im trying to explain..

 

[bug]

Well, I will throw my two cents in one more time since us 720p backers are taking a beating. What 720p results tell us that are gaming at 1080p and above is just how many FPS we can achieve with the reviewed CPU providing we also buy enough GPU to reach that goal. It's mostly for people that want very high FPS but it's also nice for determining what the reviewed CPU is actually capable of achieving in a game. We have many, many reviews of GPUs that tell us what we can expect from a GPU at a given resolution.

I'd argue that 720p results are not about FPS at all. They're about how fast the CPU can crunch numbers with a workload other than media encoding, 3D rendering or superpi computations.

 

[Deleted member 67555]

I have a 3 year old APU that cost $100... (motherboard included)
It was purchased 3 years ago in 2014 and it does 720p gaming like a champ....no gfx needed
Why 3 years later should any fucks be given about 720p on any modern enthusiast platform or even mainstream platform?
It's completely irrelevant... Time to move forward... Let's leave aspirations of 2007 behind.

 

[EarthDog]

I'd argue that 720p results are not about FPS at all. They're about how fast the CPU can crunch numbers with a workload other than media encoding, 3D rendering or superpi computations.

so...what does that tell us in gaming which is what this testing for? How is that actually useful at a higher res where the results arent the same?

Helo me understand the point..

 

[Deleted member 67555]

Every single modern CPU is fast enough to flick over everything needed for 720p... Making testing that pointless...
Gauging it's abilities where some are limited makes sense tho..right?

 

[bug]

so...what does that tell us in gaming which is what this testing for? How is that actually useful at a higher res where the results arent the same?

Helo me understand the point..

Why are you clinging to gaming? It's a CPU review. It's supposed to stress the CPU with as many workloads as possible. I don't know how to make it any more clear.

 

[EarthDog]

Why are you clinging to gaming? It's a CPU review. It's supposed to stress the CPU with as many workloads as possible. I don't know how to make it any more clear.

Is it? The review(s) is(are) covering games (among other things?) and the performance of said cpus with a bunch of games..its pretty clear to me, the purpose of this thread and why everyone is clinging to the scope of gaming...because that is what was asked of us...

From the first post...

This is not for GPU reviews, but for CPU Reviews w/ GTX 1080, for game testing.

Did i misunderstand? Seems to me its a review of CPUs with a GTX 1080 and games. I dont know if its a round up of results with this config or just setting framework for game testing for new CPUs though...

 

[W1zzard]

just setting framework for game testing for new CPUs though...

that

 

[Vayra86]

Is it? The review(s) is(are) covering games (among other things?) and the performance of said cpus with a bunch of games..its pretty clear to me, the purpose of this thread and why everyone is clinging to the scope of gaming...because that is what was asked of us...

From the first post...

Did i misunderstand? Seems to me its a review of CPUs with a GTX 1080 and games. I dont know if its a round up of results with this config or just setting framework for game testing for new CPUs though...

I give up. You're right entirely about 'typical gaming scenarios' and the fact that a 1080p result is not a 720p result. But that only gives me new things to wonder...

When it comes to gaming, what matters FAR more than whatever res you test on, is the benchmarks used. I would love to see CPU heavy games, and not 20 different kinds of third and first person semi-RPG shooter junk that is GPU limited by design. Because in the end, its that shooter junk that you'd *situationally* play on low res/lowest detail levels - once again underlining the importance of the lower resolution. You say knucklehead, I say competitive gaming / high refresh gaming. But apparently we don't want to hit that market, and we do want to hit every resolution above that. Mkay

In addition: what the opposition to 720p is actually saying, is self-contradictory. Because how is a 1080p - low detail settings indicative of real world performance, because apparently CPU load changes drastically if we increase detail levels... It does with resolution, right... (even though in the vast majority of engines the CPU load is very, very similar between all resolutions and scales with FPS, but apparently this is wrong, even though its precisely what we see confirmed in literally every GPU review - look at how the 1080ti caps out (hits a CPU wall) on 1080p and its nuff said)

Why is 1080p LOW perceived to be some sort of relevant benchmark, when 720p LOW isn't. I would daresay there are more people on shitty laptops with 1366x768 res than there are full blown rigs with a dedicated GPU that game at 1080p/low. Same argument goes for 1440p/low. 4K @ low I can fully understand, this is an interesting bench to do, because its something people would realistically touch on.

To underline the laptop POV: http://www.laptopmag.com/articles/laptop-screen-resolution-ripoff
But any realistic laptop buyer knows this.

All in all, there is so much contradiction here...

so...what does that tell us in gaming which is what this testing for? How is that actually useful at a higher res where the results arent the same?

Helo me understand the point..

Relative maximum performance, maximum achievable FPS.

It answers questions such as:
- how many FPS could I win by upgrading compared to my last CPU, regardless of GPU (could be integrated for all we know)
- it allows you to mirror with results from way way back
- is it useful to even buy a 240hz panel
- can I hit 120 fps reliably with CPU XYZ
- What is the difference of gaming @ 720p old system versus 1080p new system (not everybody, including TPU visitors, is rich)

And most importantly: Can it run Crysis? (I kid you not, Crysis 3 might prove surprising @ 720p)

Suffice to say there are many arguments to make for 720p, it would do people well to think about this in the scope of a wide range of readers, not just your own (enthusiast) situation.

 

[P4-630]

Oh the good old days on 768x1024 and 1024x1280....

My 1440p is coming!!

 

[Duality92]

Since you're using a GTX 1080 for your testing, I think we can safely assume that it will be more CPU limited in 1080p trying to pump out all those frames. I think the results will just be amplified in 720p, thus we're able to somewhat extrapolate for it. With this in mind, I think we can stick to 1080p, 1440p and 2160p.

edit : using a different GPU (lower end 1060/480) might want to get 720p in the mix.

 

[newtekie1]

Omg. Testing at 720 will show you which CPU has more HP to push a video card. I thought that was beyond obvious.
We're still talking about CPU benchmarks, aren't we?

That is only going to be the case if there is a GPU bottleneck at 1080p. With a GTX1080, there really isn't a bottleneck at 1080p, so going lower won't give us any more information.

 

[bug]

That is only going to be the case if there is a GPU bottleneck at 1080p. With a GTX1080, there really isn't a bottleneck at 1080p, so going lower won't give us any more information.

Now that is a much better argument. Yes, if games are chosen not to be GPU limited at FHD, that could replace 720p testing.

 

[Vayra86]

Now that is a much better argument. Yes, if games are chosen not to be GPU limited at FHD, that could replace 720p testing.

Indeed, like I've pointed out, the games used are critical, much more so than res.

 

[alucasa]

Look at the fight W1zzard has started. Ban the guy.

 

[EarthDog]

Had I known less people understood a 1080 or greater isnt holding anything back at 1080p, i would have said that a lot earlier...

Games used are critical, however, i would only say having one or two of each genre is most important. Mmos and rpgs typically need more cpu than fps...so it eill behave differently for each. You cant extrapolate performance outside of each game's results anyway, but gives a bigger picture to form a conclusion from.