MSI GTX 960 4GB or MSI R9 380?

[Majestic]

High settings on a Q6600 + 960 - more or less similar. Where's the bottleneck you are so scared of?

 

Again with the crappy anecdotal, random-guy-of-youtube, evidence. No metrics on display showing GPU load, and no way of verifying he's actually running that hardware.

Plus I can't read fucking slavik so no idea what those settings are. And I'll bet if i youtube Q6600 Witcher 3 that's the first video I find. You always do this, just grab something random off google that mildly correlates your theory, don't explain it and then leave it to me to debunk. Having to type 4-5 times as many.

 

http://pclab.pl/zdjecia/artykuly/radek/2014/g1820/wykresy/c3_jungle_r.png 290X + G1820

 

http://pclab.pl/zdjecia/artykuly/chaostheory/2015/12/gfx_sapphire_fury/charts/c3w_1920vh.png 390 with i7-6700K @ 4.7ghz

 

There is your bottleneck.

[MEC-777]

But it's not nearly as bad as he's making it out to be. For fucks sake, he will lose 10% off of his 380 tops and it's still 40% faster than a 950 >.<

 

Some games are more GPU-bound and some are more CPU-bound. In this case with these components, the OP is going to run into far more CPU-bound gaming situations than otherwise. The 380 is only 40% faster than the 950 in *some* of the GPU-bound games. For the rest, it will be 0% faster.

[App4that]

But it's not nearly as bad as he's making it out to be. For fucks sake, he will lose 10% off of his 380 tops and it's still 40% faster than a 950 >.<

My 4790k bottlenecks my 980ti. Does it mean I should replace my 4790k? No. But the bottleneck is still there and I have even seen it in game. Getting a graphics card that is limited by the CPU should only be done knowingly.

[don_svetlio]

Some games are more GPU-bound and some are more CPU-bound. In this case with these components, the OP is going to run into far more CPU-bound gaming situations than otherwise. The 380 is only 40% faster than the 950 in *some* of the GPU-bound games. For the rest, it will be 0% faster.

 

My 4790k bottlenecks my 980ti. Does it mean I should replace my 4790k? No. But the bottleneck is still there and I have even seen it in game. Getting a graphics card that is limited by the CPU should only be done knowingly.

But in quite a few games the difference between a 950 and 380 will be felt far and wide. Not to mention his CPU is a true quad-core so it won't feature any of the stuttering issues associated with the Pentium and Celerons. I feel he would definitely see an improvement over a 950

[Majestic]

I'd comment on your benches if they were even slightly legitimate. @Prysin - here we go again

 

fml, ofcrouse you will just discard the evidence. Of course you fucking do. And not even bothering to explain why.

 

If you're going to outright dismiss results, you need to have a reason. Like the one I gave you on that youtube video. Not showing metrics of any kind, unintelligible menu and no way to verify hardware.

 

Plus, even it it was, that was a 25% OC. 

[don_svetlio]

fml, ofcrouse you will just discard the evidence. Of course you fucking do. And not even bothering to explain why.

 

If you're going to outright dismiss results, you need to have a reason.

If you don't give reasons why should anyone else? Explain why a Q6600 test on Witcher 3 via youtube video with an fps counter is not considered factual evidence

[App4that]

But in quite a few games the difference between a 950 and 380 will be felt far and wide. Not to mention his CPU is a true quad-core so it won't feature any of the stuttering issues associated with the Pentium and Celerons. I feel he would definitely see an improvement over a 950

I'd say that you could go ether way. Replace the CPU or graphics card first. Both will have limitations, so only the OP can decide the priority.

[MEC-777]

But in quite a few games the difference between a 950 and 380 will be felt far and wide. Not to mention his CPU is a true quad-core so it won't feature any of the stuttering issues associated with the Pentium and Celerons. I feel he would definitely see an improvement over a 950

 

Yes, the OP would see an improvement over a 950, but only in a few games and no improvement in others. Is it worth the upgrade if the improvements are not significant, or there is no improvement in a number of games the OP plays on a regular basis? That decision is up tot he OP, but IMO, it's not worth it. I would keep the 950 until I had the funds to upgrade the entire system and reap the benefits of a more powerful GPU in full. 

[don_svetlio]

-snip-

I never actually saw you explain WHY that result is invalid aside from it being from youtube (like 99% of benches)

[don_svetlio]

Yes, the OP would see an improvement over a 950, but only in a few games and no improvement in others. Is it worth the upgrade if the improvements are not significant, or there is no improvement in a number of games the OP plays on a regular basis? That decision is up tot he OP, but IMO, it's not worth it. I would keep the 950 until I had the funds to upgrade the entire system and reap the benefits of a more powerful GPU in full. 

While I do agree, I'd also like to point out some games benefit much more from a GPU than a CPU and those aren't as few as you might believe. Look at Dark Souls 3 for instance, it's requirements are a i3 2100 and a 6870/560. GPUs are important.

[Majestic]

I never actually saw you explain WHY that result is invalid aside from it being from youtube (like 99% of benches)

 

I did. Twice. You just don't understand, because you don't know jack shit about how computers work. You just continue to shitpost and all your "knowledge" is whatever is most popular. Like CPU not mattering, AMD unequivically being better value. All the doctrines you just repeat, like some LTT parrot.

 

He's not showing any metrics, other than absolute framerate. No CPU load, no GPU load, nothing. There is no way of telling how badly the GPU is being bottlenecked. And it's very easy to splice in footage going to the menu and showing ultra graphics, then returning back to medium setting gameplay.

 

Most of these video's are people rationalizing to themselves their hardware is still fast. So they have every invested interest to make it appear faster than it is. Same as people making screenshots of performance, staring at a wall with 60 fps.

[don_svetlio]

I did. Twice. You just don't understand, because you don't know jack shit about how computers work. You just continue to shitpost and all your "knowledge" is whatever is most popular. Like CPU not mattering, AMD unequivically being better value. All the doctrines you just repeat, like some LTT parrot.

 

He's not showing any metrics, other than absolute framerate. No CPU load, no GPU load, nothing. There is no way of telling how badly the GPU is being bottlenecked. And it's very easy to splice in footage going to the menu and showing ultra graphics, then returning back to medium setting gameplay.

What about this?

[App4that]

While I do agree, I'd also like to point out some games benefit much more from a GPU than a CPU and those aren't as few as you might believe. Look at Dark Souls 3 for instance, it's requirements are a i3 2100 and a 6870/560. GPUs are important.

Upgrading the CPU does improve just about everything done on your computer. Not saying that is most games the GPU isn't more important, but there is an argument to going after the CPU first. I can speak from experience that when you do finally upgrade the GPU you get the full benefit, and a generous swelling of pride.

[Majestic]

What about this?

 

30-ish fps dips and 70% GPU load? Bottleneck extravaganza.

[App4that]

What about this?

 

I love you Don but I have to stop you right there. Fallout is my native land and as such I know her well.

 

If that's ultra I'm Donald Trump.

[Majestic]

I love you Don but I have to stop you right there. Fallout is my native land and as such I know her well.

 

If that's ultra I'm Donald Trump.

 

Are you going to make LTT great again? 

[KeltonDSMer]

It totally depends on the games you play and at what resolution/settings as well whether or not you will see gains from going to a 380. I have a q6600 @3.0 and a 270. In DOTA 2 for example, I get the same FPS @ 1080p minimum settings/no AA as I do @ 1440p max settings/with AA. 100% cpu bottlenck.

 

I could test it with my 7950, but I'm sure I'd see exactly the same FPS with that card because I am so far below 100% usage on the 270 and the driver overhead would be exactly the same (both GCN 1.0). FWIW, I run my 270 @800/1200mhz/-20% power because it bottlenecks so hard.

 

 

OP, check your usage with the 950 in a bunch of different titles. If you are seeing full GPU usage in some of them, you might see better performance from a 380 in those titles (I say might because there is still the driver overhead issue which shifts the bottleneck even further on the cpu with an AMD card). If you aren't getting 100% (99%) usage in the games you like playing there is no reason to change cards without upgrading the cpu IMO.

[don_svetlio]

30-ish fps dips and 70% GPU load? Bottleneck extravaganza.

I see 80% GPU load with 90% CPU-load.

[Majestic]

I see 80% GPU load with 90% CPU-load.

 

I know you do, because you see what you want to see. That is not ultra, it has massive dips and It also does not sustain 80% GPU load, it drops to 70% frequently. On an area that doesn't particularly seem populated. I bet if he walks around town with those settings, it's a slideshow.

 

Q6600 is a CPU from 2007. Nbody expects a 9 year old CPU to run Ultra settings, and noone would advice anyone to pair a high-end GPU with a 9 year old CPU except for you. Which is why you're giving out bad advice, and you need to stop it.

[App4that]

Are you going to make LTT great again? 

My first executive action will be to place all console peasants on a list so we can track them.

[don_svetlio]

My first executive action will be to place all console peasants on a list so we can track them.

So you're the CIA now? Great.

[App4that]

So you're the CIA now? Great.

Don't shy away Don, that Fallout video was not ultra. Did you catch the shadow distance and object fade? The blocks were modded to probably 3 too. 5 is stock, I run mine at 7, 13 is the max.

 

You should know not to use Fallout with me in the room Don

[don_svetlio]

I know you do, because you see what you want to see. That is not ultra, it has massive dips and It also does not sustain 80% GPU load, it drops to 70% frequently. On an area that doesn't particularly seem populated. I bet if he walks around town with those settings, it's a slideshow.

 

Q6600 is a CPU from 2007. Nbody expects a 9 year old CPU to run Ultra settings, and noone would advice anyone to pair a high-end GPU with a 9 year old CPU except for you. Which is why you're giving out bad advice, and you need to stop it.

Tell me, have you ever ran a weaker CPU with a better GPU?

[don_svetlio]

Don't shy away Don, that Fallout video was not ultra. Did you catch the shadow distance and object fade? The blocks were modded to probably 3 too. 5 is stock, I run mine at 7, 13 is the max.

 

You should know not to use Fallout with me in the room Don

To be honest I couldn't care less about the actual game - it just goes to show you can always tweak the settings to your liking.

What I want to say is upgrading for some people is not as simple and I am definitely able to relate to the OP as that's how I've been doing it for quite a while now. I always see a performance increase, even when the CPU is bottlenecking you still notice a tangible increase in fps as well as the ability to just ramp up settings so as to stress the GPU more.

[App4that]

To be honest I couldn't care less about the actual game - it just goes to show you can always tweak the settings to your liking.

What I want to say is upgrading for some people is not as simple and I am definitely able to relate to the OP as that's how I've been doing it for quite a while now. I always see a performance increase, even when the CPU is bottlenecking you still notice a tangible increase in fps as well as the ability to just ramp up settings so as to stress the GPU more.

But the game was being run at console settings at best. Video sucked so I couldn't tell that much. It's after the Godrays and shadow distance cross medium you'll see a massive bottleneck. That and the lower CPU load because of the Nvidia drivers also make the video a poor example.