Jump to content

KSP1 Computer Building/Buying Megathread


Leonov

Recommended Posts

1 minute ago, Elthy said:

For some reason forbes doesnt let me past that inspirational quote, propably one of all the privacy addons i use. But i knew them as an economics magazine, didnt know they test hardware...

Spoiler

Nvidia GTX 1060 3DMark Time Spy Benchmark (DX12)Hitman (DX12) Benchmark for Nvidia GTX 1060

Gears of War Ultimate (DX12) Benchmark for Nvidia GTX 1060Rise of the Tomb Raider (DX12) Benchmark for Nvidia GTX 1060Ashes of the Singularity DX12 Benchmark for Nvidia GTX 1060

 

Link to comment
Share on other sites

11 minutes ago, Alphasus said:

Forbes Benchmarks
Tied or slightly slower in all tests(except for synthetics, which theoretically shows bad optimization), except for Tomb Raider, where the 480 got beaten by 20%.

I see two wins for the GTX1060, two for the RX480 and one tie. Are you looking at the same benchmarks?

Link to comment
Share on other sites

9 minutes ago, Camacha said:

I see two wins for the GTX1060, two for the RX480 and one tie. Are you looking at the same benchmarks?

Yes, I was. Let me just re quote something I said last post.

Tied or slightly slower in all tests(except for synthetics, which theoretically shows bad optimization), except for Tomb Raider, where the 480 got beaten by 20%(much more significant than the 11% that the 480 gets).

Edited by Alphasus
Link to comment
Share on other sites

Just now, Alphasus said:

Yes, I was. Let me just re quote something I said last post.

Tied or slightly slower in all tests(except for synthetics, which theoretically shows bad optimization), except for Tomb Raider, where the 480 got beaten by 20%.

Which was blatantly untrue. The RX480 does best in two actual games, with the GTX1060 edging ahead in another. Do note that Tomb Raider shows a ~20% advantage for the GTX1060, but that the RX480 also has a ~20% over its rival in the Hitman benchmark. The final game is tied and the other benchmark is artificial.

It is exactly these kinds of interpretations that lead to the feeling that your agenda is always to push the GTX cards.

Link to comment
Share on other sites

13 minutes ago, Camacha said:

Which was blatantly untrue. The RX480 does best in two actual games, with the GTX1060 edging ahead in another. Do note that Tomb Raider shows a ~20% advantage for the GTX1060, but that the RX480 also has a ~20% over its rival in the Hitman benchmark. The final game is tied and the other benchmark is artificial.

It is exactly these kinds of interpretations that lead to the feeling that your agenda is always to push the GTX cards.

I also think that lots of engines are still DX11, which pretty much means that DX11 will be meaningful until DX11 engines are common, or if you dont

I actually plan on buying the AMD cards(generation after this) when they release(2017/18 for reference)because I feel that Nvidia has no good card at the $300 price point, and their price points will keep going up. But I do love what Nvidia is doing with putting full desktop(?) cards in laptops, so any laptops with GPUs I buy will be Nvidia unless AMD gets much more efficient cards(50-75W matters in a laptop).

Also IIRC Hitman did the same thing Nvidia was doing with games where other manufacturer's cards were gimped(unintentionally in the case of Hitman) because of tech being used. I don't use Witcher Hairworks benchmarks for that reason. I mean that it was optimized for AMD cards, and I also won't use Project CARS benchmarks either(R9 390s can't beat 750 Ti's).

Edited by Alphasus
Link to comment
Share on other sites

Yeah, GTX 1060-1080 are great for gaminglaptops, the efficency is impressive. There is just the big issuse of the price, the cheapest pascal one in germany costs 1600€, where it becomes viable to buy 2 different devices. Those super expensive gaming laptops are quite a niche, notebooks below 1000€ are propably used way, way more. This is the place where Polaris 11 will shine (if it can make its way into notebooks before the small pascal chip), it just looks like the perfect chip for those laptops bought because schoolkids want to play games while parents want them to have a laptop.

Link to comment
Share on other sites

8 minutes ago, Elthy said:

Yeah, GTX 1060-1080 are great for gaminglaptops, the efficency is impressive. There is just the big issuse of the price, the cheapest pascal one in germany costs 1600€, where it becomes viable to buy 2 different devices. Those super expensive gaming laptops are quite a niche, notebooks below 1000€ are propably used way, way more. This is the place where Polaris 11 will shine (if it can make its way into notebooks before the small pascal chip), it just looks like the perfect chip for those laptops bought because schoolkids want to play games while parents want them to have a laptop.

I think that the high end laptop market should drop in price soon like how it did last generation(Dell Inspiron 7--- at $800, 960M).

Small Pascal might not even exist. If it does, I'd expect a 1050 to compete with the 470.

Also if anyone wants to do VR, Forbes(and Steam Hardware Survey) agree that 1060s are better(can't tell why).

Edited by Alphasus
Link to comment
Share on other sites

I remember a review which showed framedrops on the 480, the stuttering apparently got noticeable. Well, with the price of an Oculus Rift or a HTC Vive the 480/1060 are outside of the price range anyway, if you spend that much on an hardwareupgrade you can also spend 10%-20% more for way higher graphic power.

39 minutes ago, Alphasus said:

Small Pascal might not even exist. If it does, I'd expect a 1050 to compete with the 470.

The other pascal chips also came out of the blue, Nvidia could suprise us again...

Link to comment
Share on other sites

I have to say AMD have lost this round on the GPU front. The RX 480 is their best card and it's getting matched or beaten by nVidia's mid-range one. They're also losing on power efficiency to nVidia again despite AMD having talked up the efficiency of Polaris before launch. I thought AMD and nVidia both moved onto the same process, but AMD are at a huge clock speed disadvantage for some reason. Despite that the RX 480 has theoretically higher performance than the GTX 1060 but isn't converting that into higher gaming framerates.

I'm not saying the RX 480 is bad. I'm sure it's a good card at the right price. But considering that right price is similar to or a bit less than the GTX 1060, when the RX 480 is a physically larger chip thus costing more to make ... yeah, not so good for AMD.

Edited by cantab
Link to comment
Share on other sites

Just now, cantab said:

I have to say AMD have lost this round on the GPU front. The RX 480 is their best card and it's getting matched or beaten by nVidia's mid-range one.

Traditionally, the high-end market has been a prestigious one, but not very profitable. Hitting the lower high-end to mid-range market is much more important and AMD seems to do well there. Having the performance crown is mostly a chest beating affair, even though it does seem to help sales a bit too.

That being said, the top performance and efficiency of Nvidia in their latest generation is pretty impressive. Even their own high-end cards from just a generation before seem rather uninteresting by comparison.

Link to comment
Share on other sites

The lower clocks are part of the concept, AMD has a higher density on its chips. But yeah, efficency is lower by quite a bit, although its not as bad as the 970 vs. 390.

But i wouldnt say they have lost, they take a different approach than Nvidia. Choosing between the 1060 and 480 isnt easy and depends on how you want to use the card, add in Gsync vs. Freesync and it gets even more complicated.

Edit: Also noone (of us) knows how expensive the chips realy are, maybe the Polaris 10 is cheaper to produce than the GP 106. Chip area isnt everything.

Edited by Elthy
Link to comment
Share on other sites

1 hour ago, Elthy said:

Edit: Also noone (of us) knows how expensive the chips realy are, maybe the Polaris 10 is cheaper to produce than the GP 106. Chip area isnt everything.

We have seen that a while back with the HD4770. Even though that was, at that point, a small chip, AMD managed to achieve great yields on a new process by using a few smart tricks. Nvidia struggled to do the same, which meant their yields were much lower and the chips became much more expensive in return. Yields, binning and a few other factors are as important as raw chip area. You might be better off with a process you control well.

Link to comment
Share on other sites

23 hours ago, Alphasus said:

Also if anyone wants to do VR, Forbes(and Steam Hardware Survey) agree that 1060s are better(can't tell why).

Could you post quotes and sources? Considering the earlier inaccurate representation of the results, it is invaluable for us to have a look ourselves.

Why is Forbes being pushed as an authority? I am not saying their results are worthless, but a business magazine doing hardware reviews certainly gives me pause. As far as I am aware the Steam Survey only tells us about popularity and not actual fitness for a certain purpose. This can also not be deduced from popularity in any meaningful way.

23 hours ago, Alphasus said:

Also IIRC Hitman did the same thing Nvidia was doing with games where other manufacturer's cards were gimped(unintentionally in the case of Hitman) because of tech being used. I don't use Witcher Hairworks benchmarks for that reason. I mean that it was optimized for AMD cards, and I also won't use Project CARS benchmarks either(R9 390s can't beat 750 Ti's).

Can you substantiate your claim? Nvidia was caught employing some very nasty business practices. Saying AMD is doing the same is a rather grave accusation.

Edited by Camacha
Link to comment
Share on other sites

2 minutes ago, Camacha said:

Could you post quotes and sources? Considering the earlier inaccurate representation of the results, it is invaluable for us to have a look ourselves.

Why is Forbes being pushed as an authority? I am not saying their results are worthless, but a business magazine doing hardware reviews certainly gives me pause. As far as I am aware the Steam Survey only tells us about popularity and not actual fitness for a certain purpose. This can also not be deduced from popularity in any meaningful way.

Can you substantiate your claim? Nvidia was caught employing some very nasty business practices. Saying AMD is doing the same is a rather grave accusation.

I said that the game was optimized for AMD cards unintentionally. Considering that consistently equal cards from Nvidia and AMD have performed very differently in Hitman, with extreme performance differences, I think its fair to say that the game was unintentionally optimized for AMD.

SteamVR Performance Test Results That's from the Forbes article, and was used to make this claim:"Where Nvidia’s GTX 1060 — at least the Founder’s Edition — brings in a decisive victory is thermals and VR performance."

 

Link to comment
Share on other sites

2 minutes ago, Alphasus said:

I said that the game was optimized for AMD cards unintentionally. Considering that consistently equal cards from Nvidia and AMD have performed very differently in Hitman, with extreme performance differences, I think its fair to say that the game was unintentionally optimized for AMD.

 

This has been the case for decades. Through intentional sponsoring, or simply game developers developing on one brand of cards instead of the other, certain games have performed better with certain models and brands. That is the very reason you do benchmarks per game. A certain brand or card edging ahead is exactly what you want to test and see.

 

2 minutes ago, Alphasus said:

That's from the Forbes article, and was used to make this claim:"Where Nvidia’s GTX 1060 — at least the Founder’s Edition — brings in a decisive victory is thermals and VR performance."

So, again: why Forbes and why a benchmark that another technology reviewing site calls rather opaque? Admittedly, proper benchmarks seem surprisingly hard to come by.

Link to comment
Share on other sites

Damn, i realy cant find the VR test i mentioned earlier. They didnt test only syntetic benchmarks, but also real games, where the RX 480 showed worse performance than the 1060 (and 390/970) due to framedrops/stuttering. It looked quite credible. If anyone wants to invest more time into this you should search the AMD or PCMR subreddit, it was linked in one of those.

Link to comment
Share on other sites

13 minutes ago, Elthy said:

Damn, i realy cant find the VR test i mentioned earlier. They didnt test only syntetic benchmarks, but also real games, where the RX 480 showed worse performance than the 1060 (and 390/970) due to framedrops/stuttering. It looked quite credible. If anyone wants to invest more time into this you should search the AMD or PCMR subreddit, it was linked in one of those.

Even though I have not searched very extensively, I have not come up with any credible and significant results either.

Link to comment
Share on other sites

Regarding "VR" benchmarks, don't the Pascal cards have new features and capabilities targeting VR, in particular the ability to much more efficiently render one scene from multiple viewpoints, such as one for each eye in a VR headset. Great if you're doing VR but it means the VR-specific benchmarks aren't representative of general performance.

Link to comment
Share on other sites

I now ordered a Gainward GTX 1070 Phoenix and a Dell S2716DG. Im still not sure if that was the right decision, but im kinda tired of waiting for Vega. My guess is that ill regret it in 3-4 years when i want a new GPU but am stuck with G-Sync, i totaly depend on Nvidia to futher support it.

Link to comment
Share on other sites

On 28.8.2016 at 10:32 PM, Alphasus said:

Thanks, I'll take a look at that case!

So, computer is running, and im kinda amazed by the case. Ive never seen such a well made case for this price, it felt more like costing 80$. Its completly painted inside, has good material strenght and seems well fabricated (no sharp edges). Also i liked its interior design with the PSU in the front.

But im not sure if its good for a GPU, maybe for DHE designs. The case got filled with cables, there wasnt much of an airflow remaining (but enough for an i3). I didnt know ITX could be that cramped, i dont want to test what happens if you try to stuff an ATX PSU and a large GPU in there...

Link to comment
Share on other sites

1 hour ago, Elthy said:

So, computer is running, and im kinda amazed by the case. Ive never seen such a well made case for this price, it felt more like costing 80$. Its completly painted inside, has good material strenght and seems well fabricated (no sharp edges). Also i liked its interior design with the PSU in the front.

But im not sure if its good for a GPU, maybe for DHE designs. The case got filled with cables, there wasnt much of an airflow remaining (but enough for an i3). I didnt know ITX could be that cramped, i dont want to test what happens if you try to stuff an ATX PSU and a large GPU in there...

I might go with the Fractal case then, because that should have more GPU space.

Link to comment
Share on other sites

On 31.8.2016 at 0:11 AM, Elthy said:

I now ordered a Gainward GTX 1070 Phoenix and a Dell S2716DG. Im still not sure if that was the right decision, but im kinda tired of waiting for Vega. My guess is that ill regret it in 3-4 years when i want a new GPU but am stuck with G-Sync, i totaly depend on Nvidia to futher support it.

And im sending them back. My feedback:

GTX 1070 Phantom:

The cooler is amazing, even under full load (112% Powerlimit) its still almost unadible. I never thought its possible to have such a high TDP cooled that silent. Ill sure miss that, but i hope the MSI 480 Gaming X ive ordered too wont be much louder. The performance is awesome, too, but kinda overkill for 1080p where i will stay for now. Sadly the card chirped a bit under load, with that silent cooler it was its main sound source.

Dell S2716DG:

Im kinda disappointed by it. Reviews seemed decent and i was told TN colors werent as bad as a few years back, but no matter how much i tinkered with the settings everything just seemed kinda gray. Its colors are way, way worse than those of my Dell UltraSharp U2312HM, noticeable for everyone i asked. I was propably to used to better colors on all my other hardware.
The 1440p are kinda nice in games, but pose problems when it comes to scaling in Windows/Firefox. Windows upscaling isnt realy an option, e.g. Steam looks horrible with it.
144hz/Gsync are also kinda nice, the effect is noticeable, but not enough to justify the drawbacks and the costs. I dont play lots of fast shooters, only Overwatch, so its not realy worth it.

For now ill wait for better Monitors, AMD/Freesync seems to be the better option in the long run there since they have way more to choose from...

Link to comment
Share on other sites

1 hour ago, Elthy said:

And im sending them back. My feedback:

GTX 1070 Phantom:

The cooler is amazing, even under full load (112% Powerlimit) its still almost unadible. I never thought its possible to have such a high TDP cooled that silent. Ill sure miss that, but i hope the MSI 480 Gaming X ive ordered too wont be much louder. The performance is awesome, too, but kinda overkill for 1080p where i will stay for now. Sadly the card chirped a bit under load, with that silent cooler it was its main sound source.

480s are also overkill for 1080p, unless you plan on upgrading in ~4 years. But if games get more demanding with new console releases I see this changing.

Dell S2716DG:

Im kinda disappointed by it. Reviews seemed decent and i was told TN colors werent as bad as a few years back, but no matter how much i tinkered with the settings everything just seemed kinda gray. Its colors are way, way worse than those of my Dell UltraSharp U2312HM, noticeable for everyone i asked. I was propably to used to better colors on all my other hardware.
The 1440p are kinda nice in games, but pose problems when it comes to scaling in Windows/Firefox. Windows upscaling isnt realy an option, e.g. Steam looks horrible with it.

Exactly why I likely won't upgrade. Resoultion scaling still seems horrible in my experience(tried it with a friend's 4k laptop).
144hz/Gsync are also kinda nice, the effect is noticeable, but not enough to justify the drawbacks and the costs. I dont play lots of fast shooters, only Overwatch, so its not realy worth it.

For now ill wait for better Monitors, AMD/Freesync seems to be the better option in the long run there since they have way more to choose from...

It would be nice if Nvidia used open source stuff so that G-Sync isn't so highly priced. You can't get 1440p cheaply though, as 4k is cheaper for some reason.

Edited by Alphasus
Link to comment
Share on other sites

2 hours ago, Alphasus said:

480s are also overkill for 1080p, unless you plan on upgrading in ~4 years. But if games get more demanding with new console releases I see this changing.

Exactly my plan, my 7970 also lasted me over 4 years. Similar with all my friends getting an 480 now, all have GPUs from the first 28nm generation. The only reason i would upgrade sooner would be a great pricedrop of VR or an extremly good game, maybe Cyberpunk 2077.

Link to comment
Share on other sites

22 hours ago, Alphasus said:

Exactly why I likely won't upgrade. Resoultion scaling still seems horrible in my experience(tried it with a friend's 4k laptop).

Really? You can't even scale it back to 1080 (or exactly half the native resolution) without it looking awful? That's a bummer.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...