Raja Koduri goal for 2017 is to have 1000$ 4K ready PC

[PlayStation 2]

9 hours ago, deXxterlab97 said:

More resolution = more gpu stress and less cpu. An i5 and 1080 could do

But... that's... not... how... it... works...

[Doobeedoo]

9 minutes ago, Sauron said:

You'd be surprised, some settings barely make any difference in visual quality whereas a higher resolution is a night and day difference.

True for some settings yes. Though just for reference 1080p or like 1440p maxed out high settings will look better than 4K on low/med settings. It also depends on game too though, for like highly vegetation like open world games with many tiny details like such 4K will be better so you don't see pixels on high ppi tho.

[Misanthrope]

2 hours ago, Technicolors said:

oh shit is this nines rodriguez from masquerade bloodlines? liked playing that game last year 

Indeed! I played it back when it first came out (and got into the PnP game in the late 90s) such a great game.

[Memories4K]

4 minutes ago, Misanthrope said:

Indeed! I played it back when it first came out (and got into the PnP game in the late 90s) such a great game.

Great game, was hoping the remake would go through last year but that got shut down real quick.

I'll probably play+mod it again sometime soon.

[LAwLz]

6 hours ago, Prysin said:

CPU % utilization is fine if it WAS 200FPS, i do remember it being more like 55-70 FPS in the cases i am talking of. Still above or around 60FPS, but not "enough" to make it a "non issue".

I'd say 55-70 FPS is enough to make it a "non issue". You can hardly say "only" having 55-70 FPS with an i5 makes it unsuitable for gaming (assuming that's even true).

If you do then I think your definition of "issue" differs from 99% of everyone else.

 

 

6 hours ago, Prysin said:

4 cores/4 threads isnt enough for modern AAA titles. It just isn't. You may claim it is, but you'll be hurting the PC Gaming scene by doing so. Sure an i5 is still enough, just barely, it's not a major bottleneck today, but it is "tomorrow". It isn't enough for high end gaming a couple of years from now.

It isn't enough for modern AAA titles according to whom? At what settings is it not enough? Where is the line between "not enough" and "enough" drawn?

Future proofing is a terrible idea. Like you said, the i5 is enough today, but in the future it might be. So how about buying the i5 today, saving a bit of money, and then putting that money towards a better CPU in the future when games actually demand the extra performance (assuming they do)?

You wouldn't say "hey, don't but that RX480! Buy the GTX 1070 instead! In a couple of years the RX480 will not be enough for high end gaming!" right? You should buy what you need, when you need it.

 

 

6 hours ago, Prysin said:

I could go on, but i am sure your personal opinion is too far entrenched into one camp or the other regarding this subject

What personal opinion do I have? The only things I have said are that:

1) Looking at Windows' CPU utilization to determine where your bottleneck is, is a terrible idea.

2) Different people have different opinions regarding what is and isn't acceptable FPS dips.

3) Generalizations are often bad.

4) You can make misleading conclusions if you don't present all the facts.

 

All of those seem reasonable to me.

 

By the way, I have been shouting that Intel should give us a 6-core CPU for the mainstream socket for ages, because I think it's time we start moving away from quad cores.

I am all for more CPU performance. If it weren't for Zen coming out soon, I would be sitting on a 6800K. I want to see what the 6 and 8 core Zen chips will perform like. I am by no means going "I only have an i5 so that should be enough for everyone!".

[Okjoek]

Now they're going for 4K for the masses? That's cool. I mean I actually prefer PCs under 600 dollars, but it's getting there.

[RGProductions]

Wow. If this is any hint at zen+vega pricing, we should be excited.

[SteveGrabowski0]

So AMD thinks we should pirate Windows?

[SteveGrabowski0]

1 hour ago, Sauron said:

You'd be surprised, some settings barely make any difference in visual quality whereas a higher resolution is a night and day difference.

Most settings at max barely make a difference. I can't tell the difference between high and ultra in the vast majority of games I have played. Right off hand I can think of one ultra setting that's easy for me to pick out vs the next step down, that being ultra grass vs very high grass on GTA V.

[SteveGrabowski0]

A $1000 4k gaming PC sounds pretty tough. Vega 10 would have to be offering around GTX 1080 level performance for say $550.

 

CPU: Ryzen quadcore chip ($200)

GPU: Vega 10 ($550)

MB: Decent AM4 board ($80)

RAM: 8GB DDR4-2133 ($50)

STORAGE: 1TB HDD ($50)

CASE: Some cheap crap ($30)

PSU: A lower end EVGA or Corsair unit ($40)

 

And then I doubt AMD has gotten their efficiency down enough to where you'd want to use a $40 psu. Polaris isn't all that efficient. Hopefully Vega is but it wouldn't surprise me if they're 250-300W cards.

[NumLock21]

OP needs to change title, to $1000 4K Ready Gaming PC. There is a difference between 4K and 4K Gaming PC. A $300 system with onboard graphics already supports 4K.

[Prysin]

8 minutes ago, LAwLz said:

I'd say 55-70 FPS is enough to make it a "non issue". You can hardly say "only" having 55-70 FPS with an i5 makes it unsuitable for gaming (assuming that's even true).

If you do then I think your definition of "issue" differs from 99% of everyone else.

 

 

It isn't enough for modern AAA titles according to whom? At what settings is it not enough? Where is the line between "not enough" and "enough" drawn?

Future proofing is a terrible idea. Like you said, the i5 is enough today, but in the future it might be. So how about buying the i5 today, saving a bit of money, and then putting that money towards a better CPU in the future when games actually demand the extra performance (assuming they do)?

You wouldn't say "hey, don't but that RX480! Buy the GTX 1070 instead! In a couple of years the RX480 will not be enough for high end gaming!" right? You should buy what you need, when you need it.

 

 

What personal opinion do I have? The only things I have said are that:

1) Looking at Windows' CPU utilization to determine where your bottleneck is, is a terrible idea.

2) Different people have different opinions regarding what is and isn't acceptable FPS dips.

3) Generalizations are often bad.

4) You can make misleading conclusions if you don't present all the facts.

 

All of those seem reasonable to me.

 

By the way, I have been shouting that Intel should give us a 6-core CPU for the mainstream socket for ages, because I think it's time we start moving away from quad cores.

I am all for more CPU performance. If it weren't for Zen coming out soon, I would be sitting on a 6800K. I want to see what the 6 and 8 core Zen chips will perform like. I am by no means going "I only have an i5 so that should be enough for everyone!".

at which point did i state that i was looking at Windows CPU utilization? Now, i am not sure what metrics FCAT and FRAPS draw their measurements from, could be the windows API, could be the on-die monitors of most modern CPUs.
 

Spoiler

 

http://www.eurogamer.net/articles/digitalfoundry-2015-intel-skylake-core-i7-6700k-review

 

 

 

I need say no more really.

 

[Energycore]

8 hours ago, VerticalDiscussions said:

Where i live, thats nearly 1900€....

That build was not optimized for where you live, that said there's nothing I can do if 1080s start at 800 euro like they do

[VerticalDiscussions]

13 minutes ago, Energycore said:

-snip-

Yeah its understandable, np. But the import taxes, man, imagine in India where they must be over 1000€ .-. and Titan XP's nearly 2000€.

 

If you live in Brazil, it must cost a whole human organ o-o

[Energycore]

4 minutes ago, VerticalDiscussions said:

Yeah its understandable, np. But the import taxes, man, imagine in India where they must be over 1000€ .-. and Titan XP's nearly 2000€.

 

If you live in Brazil, it must cost a whole human organ o-o

Trust me, I know

[LAwLz]

1 hour ago, Prysin said:

at which point did i state that i was looking at Windows CPU utilization? Now, i am not sure what metrics FCAT and FRAPS draw their measurements from, could be the windows API, could be the on-die monitors of most modern CPUs.

And even if the % is correct, it is still a terrible way of looking for bottlenecks. Because it is not an accurate way of measuring work being done.

 

Quote

<videos>

 

You do realize that DigitalFoundry were deliberately trying to make the CPU the bottleneck in the first video right? Here is a quote straight from the video:

Quote

We should remember that our tests here are designed to prepare our CPU limitations to the forefront, and contention is that in most titles where GPU is the bottleneck the difference will be very hard to detect.

[...]

But generally speaking, our gut feeling is that the i5 still offers the best sweet-spot in terms of price versus performance.

 

This is exactly the type of "generalizations based on edge cases" I was talking about. You can't just look at a video where they are deliberately trying to bottleneck their CPU, and then go "see? It bottlenecks in this video so therefore the i5 is not suitable for high end gaming!", despite it getting really high FPS numbers.

 

Also, the games all run at ~60 FPS even on the stock i5. Are you seriously going to say that the i5 is not suitable for high end gaming because it is "only" getting 60 FPS when the i7 can get 80? Oh right, but the i5 dips down to 50 FPS sometimes. Totally not suitable for gaming then, right? Oh it's only getting 120 FPS in Battlefield 4. Totally unplayable...

 

I think it's pretty funny that even in the videos you linked, the i5 keeps up with the i7 most of the time. There are some instances where the i7 is more consistent, while the i5 dips a bit, but those are the exceptions rather than the norm. For something claimed to "handily get 30% higher performance", the gap sure is very small in most cases.

 

 

Not sure what the second video was meant to show. They didn't even show an i7 in that video. Is that the video where you draw the conclusion that "100% CPU utilization = CPU bottleneck"? Because again, that is not necessarily true. 100% GPU utilization does not mean you have a GPU bottleneck either. It goes the opposite way too. You can be CPU bottlenecked even if you are only sitting at 70% CPU usage. That's because CPUs are not as simple as that.

 

Think of it like this.

I am transcoding a video. My GPU is doing the decoding, and my CPU is doing the encoding.

Since the "GPU % usage" doesn't factor in PureVideo workload my GPU will say it is sitting at close to 0 load. In reality, the SIP core is at 100% load. So now we're already seeing an issue. The system is reporting near 0 load on the GPU, but the part of the GPU that is actually important is at 100% load.

 

Now over to the CPU... Do CPU utilization take into account the different pipeline stages? Because video encoding is usually very integer heavy. So the integer units might be fully loaded but the rest of the CPU is mostly idling. If the CPU utilization counter takes all of them into consideration, then we end up in a scenario where we might be at 60% CPU utilization, but the CPU still can't do anything more for this particular type of workload. It could do other tasks, but that does not help for our program.

So in this case we might be CPU bottlenecked even though we are at 60% load, but we might also be bottlenecked by the PureVideo engine in the GPU.

If the % counter doesn't take the different execution units into consideration then you end up in a situation where we have 100% CPU load, but the CPU can still do more things if we hit it with a different type of load in addition to our encoding. And in this case it is also unreliable because 100% doesn't mean 100% on all different types of loads it could do at the same time.

 

See how % utilization is a meaningless number? I was just using simple examples as well. The reality is far more complicated.

[Prysin]

10 minutes ago, LAwLz said:

And even if the % is correct, it is still a terrible way of looking for bottlenecks. Because it is not an accurate way of measuring work being done.

 

 

You do realize that DigitalFoundry were deliberately trying to make the CPU the bottleneck in the first video right? Here is a quote straight from the video:

 

This is exactly the type of "generalizations based on edge cases" I was talking about. You can't just look at a video where they are deliberately trying to bottleneck their CPU, and then go "see? It bottlenecks in this video so therefore the i5 is not suitable for high end gaming!", despite it getting really high FPS numbers.

 

Also, the games all run at ~60 FPS even on the stock i5. Are you seriously going to say that the i5 is not suitable for high end gaming because it is "only" getting 60 FPS when the i7 can get 80? Oh right, but the i5 dips down to 50 FPS sometimes. Totally not suitable for gaming then, right? Oh it's only getting 120 FPS in Battlefield 4. Totally unplayable...

 

I think it's pretty funny that even in the videos you linked, the i5 keeps up with the i7 most of the time. There are some instances where the i7 is more consistent, while the i5 dips a bit, but those are the exceptions rather than the norm. For something claimed to "handily get 30% higher performance", the gap sure is very small in most cases.

 

 

Not sure what the second video was meant to show. They didn't even show an i7 in that video. Is that the video where you draw the conclusion that "100% CPU utilization = CPU bottleneck"? Because again, that is not necessarily true. 100% GPU utilization does not mean you have a GPU bottleneck either. It goes the opposite way too. You can be CPU bottlenecked even if you are only sitting at 70% CPU usage. That's because CPUs are not as simple as that.

 

Think of it like this.

I am transcoding a video. My GPU is doing the decoding, and my CPU is doing the encoding.

Since the "GPU % usage" doesn't factor in PureVideo workload my GPU will say it is sitting at close to 0 load. In reality, the SIP core is at 100% load. So now we're already seeing an issue.

 

Now over to the CPU... Do CPU utilization take into account the different pipeline stages? Because video encoding is usually very integer heavy. So the integer units might be fully loaded but the rest of the CPU is mostly idling. If the CPU utilization counter takes all of them into consideration, then we end up in a scenario where we might be at 60% CPU utilization, but the CPU still can't do anything more for this particular type of workload. It could do other tasks, but that does not help for our program.

So in this case we might be CPU bottlenecked even though we are at 60% load, but we might also be bottlenecked by the PureVideo engine in the GPU.

If the % counter doesn't take the different execution units into consideration then you end up in a situation where we have 100% CPU load, but the CPU can still do more things if we hit it with a different type of load in addition to our encoding. And in this case it is also unreliable because 100% doesn't mean 100% on all different types of loads it could do at the same time.

 

See how % utilization is a meaningless number? I was just using simple examples as well. The reality is far more complicated.

you really enjoy twisting words until it fits your agenda, do you not? Not to mention you are still not paying attention to what i've been saying all along. But i shall let you go over and re-read my posts as many times as you need, because i feel you are too spoiled to have the answer served to you on a silver platter this time.

 

Although, regarding what i highlighted in the quote, that is true, however i must stress that by common sense, the higher the CPU load, the more likely you are to experience a CPU bottleneck. You are highly unlikely to be having a CPU bottleneck at sub 50% load, If you are experiencing a CPU bottleneck at that load, it is probably not even the CPU, it's probably somewhere else, like IO or memory, but the fact remains that the higher % load on the component, the more likely you are to see a bottleneck and the less likely it is that you are not.

 

As for the second part i highlighted:
Strawman. We are talking about games, encoding has jack shit to do with the workloads you typically see in games. Games use a healthy variety of compare, call, IO, integer and float. You need all these types of workloads for them to work. in addition to this is the OS, and i have no idea what kinda workloads background OS tasks use, so i will refrain from commenting on that effect other then "it will add work".

 

 

[LAwLz]

6 minutes ago, Prysin said:

you really enjoy twisting words until it fits your agenda, do you not? Not to mention you are still not paying attention to what i've been saying all along. But i shall let you go over and re-read my posts as many times as you need, because i feel you are too spoiled to have the answer served to you on a silver platter this time.

No, I actually think it's you who are twisting words around (from the Digital Foundry video).

Or do you mean I am twisting your words around? Because I think everyone here will have interpreted your posts the way I did. I mean, I don't really have to twist your words when you flat out say things like:

11 hours ago, Prysin said:

i5s, are not enough for high end gaming. But people refuse to acknowledge it, because they bought an i5 and hate the idea that they were screwed over.

10 hours ago, Prysin said:

digital foundry proved its not just a issue tomorrow, but today.

10 hours ago, Prysin said:

4 cores/4 threads isnt enough for modern AAA titles. It just isn't. You may claim it is, but you'll be hurting the PC Gaming scene by doing so.

 

The way I read your posts, and the way I think most people read your posts, is that you're saying you should not get an i5 for gaming, because they will not be able to give you satisfying performance. Judging by your comment about how other people "refuse to acknowledge it", and how people are giving me thumbs up and not you, it seems like other people are interpreting it that way too.

You might not mean "i5s are unsuitable for gaming because they aren't powerful enough", but that's what your posts are saying, and it's bullshit. You're also using edge cases to justify your extreme generalization.

I am not the one using hyperbole, you are. You are quite frankly acting quite hostile towards me too (the way I interpreter your posts) while I think I have been calm, asked for more evidence in a polite and unbiased manner and even added a bunch of "you might mean it differently" to try and be as polite and calm as possible. But your reaction to this is to call me emotional and ignoring facts. Maybe, just maybe, it is you who is emotionally invested and don't look at the facts with an objective eye?

[Prysin]

44 minutes ago, LAwLz said:

No, I actually think it's you who are twisting words around (from the Digital Foundry video).

Or do you mean I am twisting your words around? Because I think everyone here will have interpreted your posts the way I did. I mean, I don't really have to twist your words when you flat out say things like:

 

 

The way I read your posts, and the way I think most people read your posts, is that you're saying you should not get an i5 for gaming, because they will not be able to give you satisfying performance. Judging by your comment about how other people "refuse to acknowledge it", and how people are giving me thumbs up and not you, it seems like other people are interpreting it that way too.

 

You might not mean "i5s are unsuitable for gaming because they aren't powerful enough", but that's what your posts are saying, and it's bullshit. You're also using edge cases to justify your extreme generalization.

I am not the one using hyperbole, you are. You are quite frankly acting quite hostile towards me too (the way I interpreter your posts) while I think I have been calm, asked for more evidence in a polite and unbiased manner and even added a bunch of "you might mean it differently" to try and be as polite and calm as possible. But your reaction to this is to call me emotional and ignoring facts. Maybe, just maybe, it is you who is emotionally invested and don't look at the facts with an objective eye?

yes, all those thumbs up, grand total of what? 5 in 4-6 hours?

read it as you want, the results are clear. The i5's, all of them, is moving towards being unfit for mainstream gaming. They are getting closer and closer to no longer backing XX70 GPUs from Nvidia and X90 GPUs from AMD. It will not power them properly going forward, and as such, one shouldn't consider them a "good purchase". Not to mention, you say "who cares if there is a little dip here and there".... wait what? Smooth frame delivery is the most important thing. Screw 60FPS, if your frame delivery is stuttering, regardless of FPS (which is the whole point of DF doing frame timing, in which case we DO see the i5 struggling time and time again), then your experience will be bad. What sane version would advocate that a product not delivering a smooth experience is ok, just because the next step up IS quite pricy?! Doesn't that mean that you're setting the bar too low for a good experience in the first place?

[LAwLz]

1 hour ago, Prysin said:

The i5's, all of them, is moving towards being unfit for mainstream gaming.

By that logic, so are the i7s. I guess we shouldn't recommend the RX480 either, right? Because it is moving towards being "unfit for mainstream gaming".

Do you even know what mainstream gaming is? Mainstream gaming is not 4K, never dipping below 60 FPS. Mainstream gaming is Minecraft, LoL, CS:GO. Those types of games. Even my really old 2500K can play those games more than well enough. Some dips below 60 FPS in some of the most graphic intense games current out is more than fine for 99.999% of people. Hell, most people are used to 30 FPS (thanks consoles).

You claim that I speak in hyperbole, but it is actually you who do. You are saying i5s are unfit for gaming, and your source is a video of an i5 doing about 100 FPS in multiple games, with the rare dip below 60 in some. If you think that's unplayable, damaging to PC gaming, and all the other things you have said then you're out of your mind. You must be trolling. Seriously, you are complaining and saying that a setup that gets ~140 FPS in Battlefield 4 is unfit for gaming, and harmful to the entire gaming community.

 

I think you're acting completely irrationally. The i5 is more than enough for gaming. You have not provided any evidence for it not being enough. Not even your own links agree with you.

Again, you're making a generalization based on a test deliberately made to not represent "real world" usage. Even the video itself says that it is deliberately made to bottleneck the CPU, while other setups and games will not see the same bottleneck.

 

1 hour ago, Prysin said:

It will not power them properly going forward, and as such, one shouldn't consider them a "good purchase".

Digital Foundry disagrees with you, I disagree with you, most people will disagree with you. You can't and shouldn't base your purchases on what might happen in the future. Just buy whatever suits you right now, and if your needs change in the future then just upgrade. If history has taught us anything, then it's that future proofing pretty much never works.

 

1 hour ago, Prysin said:

Smooth frame delivery is the most important thing.

I disagree. I think it is important but not always important, and other things are important too. I'd rather have my FPS fluctuating between 80 and 100, than to have it stable at 40 for example.

 

1 hour ago, Prysin said:

Doesn't that mean that you're setting the bar too low for a good experience in the first place?

Nope

You're setting it unreasonably high, making generalizations, and basing your opinion on a test deliberately meant to represent a worse case scenario which does not reflect real world usage.

 

 

Let me ask you this: Do you think we should ever recommend the RX480 (for gaming)? Should we ever recommend anything below an i7 (for gaming)? You're saying that they are unfit for gaming, "disturbing" and "hurting the gaming community", and then you accuse me of hyperbole (which I don't think I've used once in this entire thread)? Come on dude... Nobody has ever died from digging below 60 FPS for a brief moment.

If you think that an average FPS of like 80, and a minimum of like 50 is unacceptable then too bad for you, but most people will disagree. By saying that it's not good enough, without providing people with the evidence to make their own minds up, you are being very misleading. I don't know if you're being misleading on purpose (I got a feeling that you are, but I might be wrong), but I think that there is a large disconnect between your generalized and hyperbole statements, and what most people think.

 

Instead of saying that i5s are harmful to the gaming scene and they won't provide a good experience, maybe you should say that in some games you might get the occasional dip down to ~50 FPS, but your average will be well above 60 in most games? That will give people enough info to make their own minds up whether or not it is "harmful" and a "bad experience"? It is almost always better to give people information and let them decide what to think, than to try and push your own opinions onto others, because a lot of times people don't share the same opinions as you do.