More Mantle Benchmarks. 290X beats 780 Ti With a 3970X @ 4.9Ghz. (32% boost on 1600p)

[deathjester]

pclab.pl have always posted biased reviews against AMD, it's a terrible website that spreads lies you should never trust it.

It always shows results which contradict all the other review sites.

Also, they benches a completely different map, with an unknown number of players.

For all we know they could have benched an empty BF4 map.

 

I don't trust either site, and I especially don't trust one that was too lazy to even overclock a last gen Intel past stock speeds with a 780Ti lol. 

 

RussianGPU? They lost all credibility with their 8350 benchmarks. No one was within 25-30 fps lows of their BS benchmarks on an AMD 8350 at stock. Sorry but welcome to the world of astroturfing.

 

Do you think if the 290x was this far ahead that sites like anandtech and Linus Tech tips wouldn't be singing Mantle's praises? If this was legit then the price of the 780Ti would be dropping. Is it?

 

AMD themselves said it wasn't going to help an I5/I7 much at all and you want me to buy these benchmarks? Do you see these benchmarks on AMD's site? Did you ever see them linking Tek Syndicate benchmarks? They knew they were BS and didn't want to associate themselves directly with the laughable benchmarks.

[MattDaemontools]

I don't trust either site, and I especially don't trust one that was too lazy to even overclock a last gen Intel past stock speeds with a 780Ti lol. 

 

RussianGPU? They lost all credibility with their 8350 benchmarks. No one was within 25-30 fps lows of their BS benchmarks on an AMD 8350 at stock. Sorry but welcome to the world of astroturfing.

 

Do you think if the 290x was this far ahead that sites like anandtech and Linus Tech tips wouldn't be singing Mantle's praises? If this was legit then the price of the 780Ti would be dropping. Is it?

 

AMD themselves said it wasn't going to help an I5/I7 much at all and you want me to buy these benchmarks? Do you see these benchmarks on AMD's site? Did you ever see them linking Tek Syndicate benchmarks? They knew they were BS and didn't want to associate themselves directly with the laughable benchmarks.

Why not bother actually reading the review ? they're definitely not the lazy ones in this discussion.

After we have tested all the cards, we still make a final measurement with our standard benchmark scene "Siege of Shanghai" to obtain comparative values. Note: We are using this time a test setup with a 3770K at default clock. Therefore, the benchmarks are not directly comparable to those of the 4770K @ 4,6 GHz comparable setups.

 

http://translate.google.com/translate?sl=de&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.pcgameshardware.de%2FBattlefield-4-PC-238749%2FSpecials%2FBattlefield-4-Second-Assault-Benchmarks-1109970%2F%3Futm_source%3Dyoutube.com%26utm_medium%3DsocialPCGH%26utm_campaign%3Dcomputecsocial%26utm_content%3DBattlefield-4-PC-238749SpecialsBattlefield-4-Second-Assault-Benchmarks-1109970

AMD themselves never said ANYTHING about Mantle on the 8350.

It was Oxide games that actually UNDERCLOCKED an 8350 to 2.0Ghz and it was on par with a 4770K at full clock speeds with Mantle in their Nitrous Engine demo.

http://forum.overclock3d.net/showthread.php?t=59004

GameGPU.ru's CPU benchmarks for the 8350 were confirmed by other websites, pclab.pl is actually the only website with out of the ordinary results.

On several occasions pclab.pl has shown results that contradicted all other tech review sites combined (without exaggeration this happened several times).

They have a track record of dubious benchies.. Back in bf3 days they actually claimed gpu bound scenario during CQ expansion,in a MP bench-all i5s reporting same averages regardless of clocks (ridiculous).

http://pclab.pl/zdjecia/artykuly/focus/vishera/oc/bf3_1920.png

The previous article from Eurogamer reporting game developers overwhelmingly choosing AMD FX multicores as the ideal basis for next gen games, means BF4 performance chart is likely a long term trend.

[Mindtrickz]

BF4 doesn't favor neither Nvidia nor AMD, if you look at the DX11 benchmarks you'll find the 280X & the 770 performing exactly on par, similarly with the 290X & the 780 and the 660 compared to the 270.

 

 

BF4 doesn't favor AMD, look at the damn benchmarks.

Both red and green are neck and neck at every single graphics tier.

 

That's silly. AMD used to give out huge bundles of games with their cards that weren't optimized specifically for them.

 

http://sites.amd.com/us/promo/never-settle/Pages/nsreloadedx.aspx

 

AMD is likely giving it out in promotion of Mantle.

 

This sort of thing is common, Nvidia bundled Metro with its cards and it ran faster on AMD GPUs.

Similar thing happened with Crysis 3 it was bundled with AMD cards but ran faster on Nvidia.

 

 

Battlefield 4 IS a AMD optimized game. It's apart of AMD's Gaming Evolved. Games specifically optimized for Radeon GPU's. Saying other wise is just nonsense. The fact that it is not optimized well enough for the money they received from AMD is one thing, but it is nonetheless AMD optimized game. 

 

Regarding Metro Last Nigh. It would appear that they have a foot in both camps; with them supporting propitiatory Nvidia and AMD technologies in their games. 

 

 

[deathjester]

Why not bother actually reading the review ? they're definitely not the lazy ones in this discussion.

http://translate.google.com/translate?sl=de&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.pcgameshardware.de%2FBattlefield-4-PC-238749%2FSpecials%2FBattlefield-4-Second-Assault-Benchmarks-1109970%2F%3Futm_source%3Dyoutube.com%26utm_medium%3DsocialPCGH%26utm_campaign%3Dcomputecsocial%26utm_content%3DBattlefield-4-PC-238749SpecialsBattlefield-4-Second-Assault-Benchmarks-1109970

AMD themselves never said ANYTHING about Mantle on the 8350.

It was Oxide games that actually UNDERCLOCKED an 8350 to 2.0Ghz and it was on par with a 4770K at full clock speeds with Mantle in their Nitrous Engine demo.

http://forum.overclock3d.net/showthread.php?t=59004

GameGPU.ru's CPU benchmarks for the 8350 were confirmed by other websites, pclab.pl is actually the only website with out of the ordinary results.

On several occasions pclab.pl has shown results that contradicted all other tech review sites combined (without exaggeration this happened several times).

They have a track record of dubious benchies.. Back in bf3 days they actually claimed gpu bound scenario during CQ expansion,in a MP bench-all i5s reporting same averages regardless of clocks (ridiculous).

http://pclab.pl/zdjecia/artykuly/focus/vishera/oc/bf3_1920.png

The previous article from Eurogamer reporting game developers overwhelmingly choosing AMD FX multicores as the ideal basis for next gen games, means BF4 performance chart is likely a long term trend.

 

Oxide games provided ZERO benchmarks of a 4770k, and said "they think". The laughable thing is? Even in the BS benchmarks where they use the older intel series (this should scream to you that this is a AMD astroturfing site). the I7 is ahead.  If you "think" the 8350 is ever beating a 4770k with or without optimization you are buying into BS. Mantle is most likely going to fail. Steam OS doesn't run DirectX and Mantle runs DirectX shaders. AMD has already said they don't care if Mantle "or something like Mantle" takes over. Carmack has said OpenGL with extensions will do the same thing and not be limited to one silly GPU series. Mantle is doomed to fail like glide, with Nvidia not jumping on, and DirectX is NEVER going to be on SteamOS due to Microsoft. Add to that AMD, Intel, Nvidia are all partners in the group that owns OpenGL. The Kronos group. PS4? Runs OpenGl.

 

Eurogamer? That is like getting tech advice from IGN. It is a console magazine that wants to sell the consoles as having powerful hardware. How is a 8350 going to beat a 4770k? It gets slaughtered on single core and is slower then a I3. It gets its head kicked in, in multicore applications that aren't games. The only thing the 8350 can compete with and win over a 4770k is zipping files and it unzips them slower. Integer cores SUCK for gaming. The 8350 gets its head kicked in by the 4770k even in workloads, and the I5 is closer (and beats the 8350 in some workloads) to a "work chip" then the 8350 is to a 4770k.

 

http://www.anandtech.com/bench/product/836?vs=697

 

Here are multicore and single core benchmarks and if you think that AMD is running 5.3ghz 24/7 you are living in a dream world. It still gets it's head kicked in. Add to that? Singlecore is more important for GAMES. A higher clocked I5 can be faster then these 6 core 2011 chips in games that aren't 8 thread and their multicore DWARFS a AMD pseduo 8 core.

 

Multicore.

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=0

 

Singlecore.

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=1

 

The only AMD there that runs close to 24/7 speeds (just under 5 ghz) gets beat by a I3-2120 running at 3.44 ghz and that is a SANDY BRIDGE. Go ahead and add your own AMD chip to the cinebench LTT thread. 

 

http://linustechtips.com/main/topic/62476-post-your-cinebench-r15-scores/

 

Make sure to include temps as well like I did. AMD reccomends you don't go over 62.5 C, because AMD reports temps wrong. Have fun with the multicore/single core scores. Then tell me how an inferior chip is going to be magically faster in games when the single core sucks, and the multicore isn't all that impressive. 

 

This is what happens when someone compares the I5 and 83xx, when they don't have sponsors on even a GTX 780 (the gap will get bigger with the 88x series). The 832x even loses at streaming a "8 core" game like BF4.

 

 

In games that aren't CPU bound and are just GPU benchmarks (which is what makes the consoles recommending AMD a JOKE, because those games are primarily gpu bound) the AMD can keep up sometimes. In games where they are cpu bound? The AMD gets slaughtered. Oh and this guy was a former AMD fanboy, which he admits in his other videos lol. 

[BrightCandle]

There are two problems with Battlefield 4 testing and its important to understand you'll see a lot of variation because of it:

1) Everyone demands multiplayer tests with 64 players because otherwise its not a valid test.

2) The amount of variation that MP and 64 player introduces into the testing process makes it quite hard to get repeatable tests. They certainly can't be compared across reviews and odds are they can't really be compared within the same review very well.

 

The only repeatable tests are single player and low count multiplayer where the conditions are controlled. Once we understand this we can account for the big variances across the different sites and realise that in different circumstances different hardware can perform very differently. My BF4 can run anywhere from 100-200 fps and that is not an acceptable variance to be introducing into a CPU comparison. Even in single player different sites test in different places since there is not a standard benchmark so the results differ based on what the particular scene uses. I myself saw massive swings in fps throughout the campaign, some of it even got unplayable and yet its the same settings I run in multiplayer at above 100 fps consistently.

 

So before you all go on an ad hominen attack against some sites with no evidence take a breath and realise that the differences you see are due to the testing procedure that is being demanded and its highly unlikely to give you the high quality comparison you are looking for. There is a reason why Anandtech only does single player tests.