AMD FX-8350 Faster Than i7-4930K For UHD

[Jack.EXE]

Before you label me as an AMD fanboy, read my wishlist.

 

I'm going to give my two cents

 

OP historically knows what he's talking about, and he's generally INTEL biased from my observations. I find this post credible from this standpoint.
 

Also:

This does prove that the 8350 wont bottleneck anything any more than a 4930K would, and while I agree this would also likely be true with an i5, it does mean that the lower clocked 8320 poses as an amazing value at $99-140 (depending on store)

 

Though this is at 4k, I do also have to admit that at much lower resolutions (MUCH LOWER, 640x480) intel would pull ahead, as games become more CPU bound.

HOWEVER, the majority of games today are GPU bound, and we all are using much higher resolutions.

 

For the sake of benchmarks, sure, use 640x480 or 640x360, but take it with a grain of salt, we game at 720p at least, I game at 3x900p, and most game at 1080, with some getting up to 1440p or even 4k.

 

so before you call BS on this, listen to yourself for a minute. It makes sense.

[clementk]

Before you label me as an AMD fanboy, read my wishlist.

 

I'm going to give my two cents

 

OP historically knows what he's talking about, and he's generally INTEL biased from my observations. I find this post credible from this standpoint.

 

Also:

This does prove that the 8350 wont bottleneck anything any more than a 4930K would, and while I agree this would also likely be true with an i5, it does mean that the lower clocked 8320 poses as an amazing value at $99-140 (depending on store)

 

Though this is at 4k, I do also have to admit that at much lower resolutions (MUCH LOWER, 640x480) intel would pull ahead, as games become more CPU bound.

HOWEVER, the majority of games today are GPU bound, and we all are using much higher resolutions.

 

For the sake of benchmarks, sure, use 640x480 or 640x360, but take it with a grain of salt, we game at 720p at least, I game at 3x900p, and most game at 1080, with some getting up to 1440p or even 4k.

 

so before you call BS on this, listen to yourself for a minute. It makes sense.

this is why i find this website BS https://www.youtube.com/watch?v=26UKz42uQ1Y. when i wanted to get a new PC i had to look up alot between the fx 8350 and the i5 4690k and the i5 was always better.

[Misanthrope]

this is why i find this website BS https://www.youtube.com/watch?v=26UKz42uQ1Y. when i wanted to get a new PC i had to look up alot between the fx 8350 and the i5 4690k and the i5 was always better.

Not always though: the 8350 is considerably cheaper and closer to low end i5 or even i3 prices for the 8320 that you can overclock to 8350 speeds anyway, possibly even with the stock cooler. The advantage however as I've recently found out is that AMD has been a bit too complacent and just keeps slashing prices without new chips. This also means that there are virtually no new mobo chipsets either, so you either go with really high end mobos for AM3+ or miss in some very very significant features like sata 3.0. Not to mention pci-e or m.2 and the like.

 

That's why I went with 1150 as my platform: feature parity overall just isn't there for AMD, even though relevant performance (as in relevant for most gamers except maybe mmo players) is up to par and enough for most games.

 

This also however, speaks volumes of the crap job most PC game developers are doing at not utilizing the extra power available on the newer intel chips: They all need to collectively smarten the fuck up and code their games for proper multithread chips.

[patrickjp93]

Not always though: the 8350 is considerably cheaper and closer to low end i5 or even i3 prices for the 8320 that you can overclock to 8350 speeds anyway, possibly even with the stock cooler. The advantage however as I've recently found out is that AMD has been a bit too complacent and just keeps slashing prices without new chips. This also means that there are virtually no new mobo chipsets either, so you either go with really high end mobos for AM3+ or miss in some very very significant features like sata 3.0. Not to mention pci-e or m.2 and the like.

 

That's why I went with 1150 as my platform: feature parity overall just isn't there for AMD, even though relevant performance (as in relevant for most gamers except maybe mmo players) is up to par and enough for most games.

 

This also however, speaks volumes of the crap job most PC game developers are doing at not utilizing the extra power available on the newer intel chips: They all need to collectively smarten the fuck up and code their games for proper multithread chips.

This is why if you're a really good parallel programmer and you have a passion for games, you should go to the video game industry. Most of the old guard missed taking high performance and parallel computing in college if they went, or if they didn't go, they never got to learn in a decent environment. It's hard to learn, especially on your own without a mentor who knows a few really good tricks to get you started, and the really advanced stuff with data race conditions and dependencies can be a nightmare for all but the brilliant.

[vm'N]

Can we please differentiate between parallel coding and multi-threading coding.

[clementk]

Can we please differentiate between parallel coding and multi-threading coding.

i am not 100% sure but as far as i know parallel codding is when you mutli threads to do the same thing at the same time. and multi-threaded coding is when you make each thread do something different. 

[HalGameGuru]

On a 100Mbps up and down connection, it's not that

THe problem is with a lot of network dependent games like MMOs YOUR speed isn't the bottleneck. It could be a node in between losing packets, traffic shaping, the servers not able to keep up, etc. A lot of MMO performance issues are game dependent. Look at a game like World of Tanks, the east coast server has a horrific ISP that throttles and traffic shapes causing low ping but massive packet loss, so your PC is capable of 120+ FPS but the network quality has you dropping into the TEENS regularly because the game is set to look for server response for its engine refresh, but WoT is a very CPU intensive game for being so "small" a footprint. You have to take any performance on a game with a server side tether with a grain of salt. Some people, especially in certain areas, no matter how fast their internet or their PCs will get horrible performance on some games because the nodes their packets have to go through are garbage.

 

The best benchmark of local hardware is a local only game. Problem is without a lot of I/O traffic over network like in BF4 or WoT you often don't get the same kind of CPU workload you want for that kind of benchmark. In MMOs ping and bandwidth are two very small pieces in a very large puzzle.

[patrickjp93]

Can we please differentiate between parallel coding and multi-threading coding.

No. Whether or not you use threads to do all the same thing with different data sets or different things is irrelevant. The old guard game programmers suck at both.

[deviant88]

bullshit strong with this one

[vm'N]

i am not 100% sure but as far as i know parallel codding is when you mutli threads to do the same thing at the same time. and multi-threaded coding is when you make each thread do something different.

Correct. Parallel code often run better on a GPU. GPU are made for parallel work.

[Majestic]

I think it has something to do with the terribly low framerates. I've seen the same thing happening in that Teksyndicate FX8350 vs 3770 video. He also ran framerates to the ground.

But in this article, even at 60+ fps, the AMD wins. This is just a BS test, and completely contradicts every other test conducted.

[nokturrduk84]

What a bullshit... Even 5 years old kid won't be deceived by this crap.

[Faa]

This is what I found from their 1440p review unigine heaven;






Look at the minimums alone, it's for a short period when transitioning quite CPU bound though, it's the same guy who made that 4K review. Just a diehard fanboy like Logan who likes fabricating benchmarks. Laughable part is that he didnt even use 8x AA for the 8350. I'll enable 8x AA and check what result I get.
 

 

One of my issues with benchmarking MMOs is that I'm not a networking guy, right but having so much happening in one area and having to have much information sent from the server to you may not result in a CPU bottleneck as much as it could be a bandwidth or latency bottleneck between the server and all of the players it needs to send information out to, so because I don't have a great understanding of it and I can't get all of the numbers from it I sort of take large groups with a grain of salt, but without large numbers of people the intel runs away with it.

Ping doesn't affect your FPS. See it from this perspective, pull your ethernet cable out, your frames remain the same, you have a few seconds to check it before you get disconnected. In some cases the NIC could bottleneck the CPU but that's never going to happen in a game.
 

 

The 8350 can still bottleneck a game, i've seen it in my benchmarks. MMORPGs don't have graphic so they remove the GPU bottleneck simply by not utilizing it fully and in this situation you do see the Intel shine much brighter than AMD but that isn't really bottlenecking the GPU but there are real world scenarios where intel can wipe the floor with AMD.

Yeah and they bottleneck 2Way SLI setups completely to the ground. Intel does it less or completely not. Besides even in mmo's GPU's are rendering the frames, have some magical CPU with a bunch of IPC, you'll easily push 2 780's each to 99% load.

[SeanBond]

GTX 780's in SLI aren't passing 60 fps average in Bioshock Infinite at 1440p?

 

Am I missing something? That seems really off to me.

What is even more shocking is sli 780's sitting at 50fps on dirt 3. I can recall similar frame rates when amd showed off dirt 3 running on two 290's... at 11520x2160..... all of these benchmarks are just...so wrong. 

[SeanBond]

This is what I found from their 1440p review unigine heaven;

Look at the minimums alone, it's for a short period when transitioning quite CPU bound though, it's the same guy who made that 4K review. Just a diehard fanboy like Logan who likes fabricating benchmarks.

you....you don't have AA on. 

[Qwertious]

http://www.tweaktown.com/tweakipedia/58/core-i7-4770k-vs-amd-fx-8350-with-gtx-980-vs-gtx-780-sli-at-4k/index.html

8350 vs 4770k

[Faa]

you....you don't have AA on. 

Which was done with a 3960x; http://www.pcper.com/reviews/Graphics-Cards/NVIDIA-GeForce-GTX-980-and-GTX-970-GM204-Review-Power-and-Efficiency/3DMark-F

[Priller]

Here is a test done by a third party on Hardforum using evga gtx 780 6gb cards in sli.

 

 

fx 9370@5.0Ghz 16gb gskill 1600 - sabertooth 990fx 2.0 - samsung 840 pro 128gb

 

i7 3770k@4.5ghz - 16gb gskill 1600 - tpower tz77-xe3 - kingston ssdnow v300 128gb

 

Crisis was with 16X AA.

 

NOTE: I have picked up FPS by either running ram higher then 1600 and/or tightening the timings on FXs. 

 

http://hardforum.com/showpost.php?p=1041154467&postcount=25

[Faa]

Here is a test done by a third party on Hardforum using evga gtx 780 6gb cards in sli.

Whats up with people picking the highest score that the GPU could pull off for AMD? Run it 50 times with 50 different cpu's, i3 will top, next round a 4300 will top, next round a 5960x will top etc. Using the variation of the GPU's performance as an argument that AMD/Intel scored or is better is retarded. Owyeah you're expected to see your loads cranking up when you're pushing a 4K monitor. Ditch the resolution down, you'll see a difference like below here, the reason they choice for 4K to crank the loads up to minimize how bad AMD can perform.

http://vr-zone.com/articles/amd-fx-8350-vs-intel-core-i7-3770k-4-8ghz-multi-gpu-gaming-performance/17494.html - scales all the way up to 100%

http://www.anandtech.com/show/7189/choosing-a-gaming-cpu-september-2013 - scales all the way up to 80%

http://www.hardocp.com/article/2011/11/03/amd_fx8150_multigpu_gameplay_performance_review/#.VDhYpJ6sWuQ

http://www.bjorn3d.com/2011/06/sli-tests-intel-sandy-bridge-vs-amd-phenom-ii/ (quite outdated but a phenom x6 = 8350)

 

[thebennyboy]

Whats up with people picking the highest score that the GPU could pull off for AMD? Run it 50 times with 50 different cpu's, i3 will top, next round a 4300 will top, next round a 5960x will top etc. Using the variation of the GPU's performance as an argument that AMD/Intel scored or is better is retarded. Owyeah you're expected to see your loads cranking up when you're pushing a 4K monitor. Ditch the resolution down, you'll see a difference like below here, the reason they choice for 4K to crank the loads up to minimize how bad AMD can perform.

http://vr-zone.com/articles/amd-fx-8350-vs-intel-core-i7-3770k-4-8ghz-multi-gpu-gaming-performance/17494.html - scales all the way up to 100%

http://www.anandtech.com/show/7189/choosing-a-gaming-cpu-september-2013 - scales all the way up to 80%

http://www.hardocp.com/article/2011/11/03/amd_fx8150_multigpu_gameplay_performance_review/#.VDhYpJ6sWuQ

http://www.bjorn3d.com/2011/06/sli-tests-intel-sandy-bridge-vs-amd-phenom-ii/ (quite outdated but a phenom x6 = 8350)

 

 

What a stupid thing to say. Guys, we should all put our resolutions down to 720p to induce a bottleneck. Who the hell games at 720p??? Did you help develop FFXIII? 4K is the future so if the 8350 still performs well as 4K then doesn't that render the higher price of the Intel chip kinda unnecessary? I know that a lot of people may disagree, but still. 

[dalekphalm]

Sauce

 

-snip-

Hold on a second here...

 

Has anyone else noticed that the title of this article is 100% false?

 

UHD = Ultra High Definition - 3840 x 2160p

 

QHD = Quad High Definition - 2560 x 1440p

 

This article is about QHD (also known as 2.5K or 1440p). It has Nothing to do with UHD. Can you please fix the title?

 

Back on topic: This is extremely fishy. I don't trust this article at all, until others can replicate/confirm the results. AT BEST it simply means that the games are GPU bound, and the CPU doesn't matter at that point. Even if games were multithreaded enough to take advantage of 8 cores vs 6 cores, the 4930K is just so much more powerful that it would negate the 2 "core" advantage that AMD has here. Not to mention that each "core" is quite a bit less powerful on the AMD side (Hoping Zen changes that).

[thebennyboy]

Hold on a second here...

 

Has anyone else noticed that the title of this article is 100% false?

 

UHD = Ultra High Definition - 3840 x 2160p

 

QHD = Quad High Definition - 2560 x 1440p

 

This article is about QHD (also known as 2.5K or 1440p). It has Nothing to do with UHD. Can you please fix the title?

 

Back on topic: This is extremely fishy. I don't trust this article at all, until others can replicate/confirm the results. AT BEST it simply means that the games are GPU bound, and the CPU doesn't matter at that point. Even if games were multithreaded enough to take advantage of 8 cores vs 6 cores, the 4930K is just so much more powerful that it would negate the 2 "core" advantage that AMD has here. Not to mention that each "core" is quite a bit less powerful on the AMD side (Hoping Zen changes that).

 

The article is correct, the only 1440p benchmarks in there are to show the difference with Intel & AMD between UHD and QHD where the Intel chip has the upper hand.

[dalekphalm]

The article is correct, the only 1440p benchmarks in there are to show the difference with Intel & AMD between UHD and QHD where the Intel chip has the upper hand.

My mistake, I searched the article for "UHD", but did not search it for "4K". However, my main confusion resulted in the quote the OP made:

 

 

Earlier on in the week, we pitted our AMD and Intel systems against each other, in a battle at 2560x1440. The GPUs in question were NVIDIA's reference GeForce GTX 780s in SLI, and the new GTX 980s in SLI. For $300, the AMD represents super value, and what most would think, the loser out of the two setups.

No reference to 4K in that at all. In fact, that quote is incredibly misleading, as it implies that the article is about 1440p (hence my own confusion).

[Faa]

What a stupid thing to say. Guys, we should all put our resolutions down to 720p to induce a bottleneck. Who the hell games at 720p??? Did you help develop FFXIII? 4K is the future so if the 8350 still performs well as 4K then doesn't that render the higher price of the Intel chip kinda unnecessary? I know that a lot of people may disagree, but still. 

Reading isnt your strongest point, isn't it? Ask your teacher what the performance variation of a GPU means. His benchmarks are showing the highest result the GPU could pull out for AMD. He probably ran it like 50 times on AMD and just once on Intel. Or probably AMD was lucky getting the best out but that's not luck anymore when it did better in every test.

Let's look at their 1440p BS;

33% difference between their result & PCper's (used a 3960x) on Intel. Which gives us a 7 fps difference meaning the 8350 was bottlenecking slightly. Apparently 980's in SLI just add 5 fps more, where as its adding 15 fps more on AMD. Apparently the minimums were cranked by 100% which is BS. Few moments you have a CPU bottleneck still in unigine heaven, which are the minimums, a 980 upgrade isn't going to do anything about that. Heaven is still quite shitty multithreaded, http://i.imgur.com/EbiYNIm.jpg - I'd like to see AMD doing better.

You benchmark CPU's when theyre holding the GPU back, not the other way around and abusing it like Logan and this idiotic fanboy did. Once we get cards that are like 4 times faster than a 980, we'll see AMD bottlenecking again in 4K this time as usual.

 

 

The article is correct, the only 1440p benchmarks in there are to show the difference with Intel & AMD between UHD and QHD where the Intel chip has the upper hand.

There wasn't any Intel vs AMD test. 

 

 

Back on topic: This is extremely fishy. I don't trust this article at all, until others can replicate/confirm the results. AT BEST it simply means that the games are GPU bound, and the CPU doesn't matter at that point. Even if games were multithreaded enough to take advantage of 8 cores vs 6 cores, the 4930K is just so much more powerful that it would negate the 2 "core" advantage that AMD has here. Not to mention that each "core" is quite a bit less powerful on the AMD side (Hoping Zen changes that).

The review just shows us the performance variation of the GPU, he just abused that variation to show AMD being better. 

[biFanBoy]

AMD cpus just lame in front of any Intel, even in non K CPU it can perform less, but the true is that they came up with 8 cores, ages a go, and Intel just now.. which is sad. Again showing they have more tech ideas than intel,