FX CPU?

[thebennyboy]

Doesnt prove nothing like i said , maybe they user ubersampling( or whatever is called in bf4) , maybe they used msaa x8 you dont dont neither do i.

What makes the benchmarks valid is that they used the same run as you do when benchmarking components for the same cpus.

 

Your argument that two different publications messed up at the same time doesnt convince me at all.

 

Perhaps if they provided more detail about what settings they used, it could be proven without reasonable doubt whether they were true or false. I still hold my doubts about the two benchmarks you post, in particular the first one.

[Faa]

Shows fuck all apart from how shit DX11 is.

DX11 is as good as Mantle in terms of removing CPU bottlenecks. It's just AMD's DirectX suffering from massive CPU overhead.

Was known for a long time AMD's DirectX drivers suffering from a large CPU overhead, PCLab.pl tested this;

Not sure why you would say DirectX11 is crap, AMD is here to blame.

Also

 

Your minimums is the same as Anandtech got, bottleneck proved

[thebennyboy]

DX11 is as good as Mantle in terms of removing CPU bottlenecks. It's just AMD's DirectX suffering from massive CPU overhead.

Was known for a long time AMD's DirectX drivers suffering from a large CPU overhead, PCLab.pl tested this;

Not sure why you would say DirectX11 is crap, AMD is here to blame.

 

Maybe because DX11 is crap? And listing a load of benches using 720p proves pretty much nothing.

 

By moving the benchmark to Direct3D 12, Microsoft and Futuremark were able to significantly reduce the single-threaded bottlenecking, distributing more of the User Mode Driver workload across multiple threads.

 

Source: http://www.anandtech.com/show/7889/microsoft-announces-directx-12-low-level-graphics-programming-comes-to-directx/4

[harrynowl]

Maybe because DX11 is crap? And listing a load of benches using 720p proves pretty much nothing.  Source: http://www.anandtech.com/show/7889/microsoft-announces-directx-12-low-level-graphics-programming-comes-to-directx/4

720p benchmarking is to remove any GPU bottlenecking

[tnlich]

Wow Zappian . I was just explaining that.. He should wait till the next gen of Intel and or AMD comes out. LOL what a concept... Yea?

[Faa]

Maybe because DX11 is crap? And listing a load of benches using 720p proves pretty much nothing.

 

 

Source: http://www.anandtech.com/show/7889/microsoft-announces-directx-12-low-level-graphics-programming-comes-to-directx/4

You aren't getting it; Mantle's main niche was to remove CPU bottlenecks. At 1080p with all ultra settings at 4x msaa and a 4670K at 4.4GHz on an empty server you aren't CPU bottlenecked at all so that's why we took it down to 720p/low settings to test how much CPU bottlenecking were removed and as you can see from the GPU usage we were heavily bottlenecked.

This is the difference between Mantle & DirectX when being GPU limited;

Difference is almost neglible. And when you're CPU bound; see my 290x at DX vs 290x at Mantle.

That 290x only managed reaching 34% usage at 150 fps with DX, With Mantle it went up to 60% at 230 fps, my 970 managed getting 60% usage as well and 230 fps. If Nvidia managed to achieve the same results without a new API that definitely means that DirectX11 isn't crap. If AMD's DirectX11 performs dramatically worse than Nvidia's DirectX11 there's definitely an issue on their side.

===========================================================

Oxide Games

Star Swarm Stress Test - ©2013

C:\Users\NJ\Documents\Star Swarm\Output_14_12_20_1300.txt

Version 1.10

12/20/2014 13:00

===========================================================

== Hardware Configuration =================================

GPU: AMD Radeon R9 200 Series

CPU: GenuineIntel

Intel® Core i5-4670K CPU @ 3.40GHz

Physical Cores: 4

Logical Cores: 4

Physical Memory: 8472465408

Allocatable Memory: 140737488224256

===========================================================

== Configuration ==========================================

API: DirectX

Scenario: ScenarioFollow.csv

User Input: Disabled

Resolution: 1920x1080

Fullscreen: True

GameCore Update: 16.6 ms

Bloom Quality: High

PointLight Quality: High

ToneCurve Quality: High

Glare Overdraw: 16

Shading Samples: 64

Shade Quality: Mid

Deferred Contexts (D3D11): Disabled

Small Batch Optimized (Mantle): Enabled

Temporal AA Duration: 16

Temporal AA Time Slice: 2

Detailed Frame Info: Off

===========================================================

== Results ================================================

Test Duration: 360 Seconds

Total Frames: 13661

Average FPS: 37.94

Average Unit Count: 4182

Maximum Unit Count: 5547

Average Batches/MS: 672.54

Maximum Batches/MS: 1534.21

Average Batch Count: 20389

Maximum Batch Count: 125874

===========================================================

===========================================================

Oxide Games

Star Swarm Stress Test - ©2013

C:\Users\NJ\Documents\Star Swarm\Output_14_12_20_1359.txt

Version 1.10

12/20/2014 13:59

===========================================================

== Hardware Configuration =================================

GPU: AMD Radeon R9 200 Series

CPU: GenuineIntel

Intel® Core i5-4670K CPU @ 3.40GHz

Physical Cores: 4

Logical Cores: 4

Physical Memory: 8472465408

Allocatable Memory: 140737488224256

===========================================================

== Configuration ==========================================

API: Mantle

Scenario: ScenarioFollow.csv

User Input: Disabled

Resolution: 1920x1080

Fullscreen: True

GameCore Update: 16.6 ms

Bloom Quality: High

PointLight Quality: High

ToneCurve Quality: High

Glare Overdraw: 16

Shading Samples: 64

Shade Quality: Mid

Deferred Contexts (D3D11): Disabled

Small Batch Optimized (Mantle): Enabled

Temporal AA Duration: 16

Temporal AA Time Slice: 2

Detailed Frame Info: Off

===========================================================

== Results ================================================

Test Duration: 360 Seconds

Total Frames: 22123

Average FPS: 61.45

Average Unit Count: 4554

Maximum Unit Count: 5937

Average Batches/MS: 1271.53

Maximum Batches/MS: 4380.75

Average Batch Count: 21258

Maximum Batch Count: 108987

===========================================================

The difference is clear, the GPU usages were nearly all the time at 99% providing overall a 50% boost, feel free to compare AMD's DX11 GPU usage with my 970/3930K at stock (56 fps);

Besides you still have worse minimums with Mantle than a i3 with two 770's using an API that's crap. 

[i_build_nanosuits]

Dude is getting there like 50% more, averaging at what your maximums were. Write AMD a mail for a refund, what a scam.

 

This is the lightest map built upon frostbyte 3 for a battlefield game EVER...you can't use this map as reference man it's just desert sand and 2 freaking tents really...come on! just sayin

and the guy won't even show his settings for all i'm concerned they could all be on low...you have to pick better than this man.

[zappian]

This is the lightest map built upon frostbyte 3 for a battlefield game EVER...you can't use this map as reference man it's just sand and 2 tents really...come on!   just sayin

 

Yup , anyone who believes that a FX cpu doenst bottleneck two hign end cards like two 290x is a moron.

[Faa]

 

This is the lightest map built upon frostbyte 3 for a battlefield game EVER...you can't use this map as reference man it's just sand and 2 tents really...come on!   just sayin

and the guy won't even show is settings for all i'm concerned they could all be on low...you have to pick better than this man.

Whatever, 51 minimums for a 4.5GHz monsterproof CPU is awful for two 280x's in CF.





The 4690K at stock just suffers a 10% degradation where as that 290x optimized CPU suffers from a 50% loss.

[WaitForMeImOnDialUp]

 

Prove it with http://technet.microsoft.com/en-us/sysinternals/bb896653.aspx

 

 

 

okay first, that program will get you kicked by punkbuster in less than 30 seconds in bf4. i tried 5 times, all 5 times i got kicked quickly.

i will however admit you are mostly correct...mostly

bf4 does start out at 6 threads and jumps up to 8 as needed, i tried to get a screen shot of 8 being used but punkbuster kept kicking me out in less than 30 seconds so all i could get is showing 7 threads being used. it does use all 8 but it does mostly use 6.

 

 

with dragonage inquisition surprisingly it mostly uses only 2-4 threads (3-4 avg) but will use all 8 but only for a second randomly.as you can see the last 4 threads only popped in use breifly

 

 

neat little program and i admit i may not be using it correctly? i am too scared to use it during any steam game as steam likes to give permanant ban for any program punkbuster doesnt like.

[79wjd]

Dear lord youall are so funny. GO Intel !!! GO AMD !!! He has a AMD he wants to NOT I REPEAT NOT SWITCH OUT THE MOTHERBOARD!!!!!!!!!! If you feel sluggish playing man and running 1 vid card. get another to X fire or SLI or even a faster card and make sure your ram is 1600 or better and around 16 gigs worth. The main problem is the FX days are at a end and intel done there 4th gen and new stuff is coming soon. Save up.. Relax  and don't let these ego maniacs tell you this or that. Back to basics always.

8gb of Ram is MORE than enough for any modern game, and ram speeds are literally meaningless unless an APU is being used. 

 

Wow Zappian . I was just explaining that.. He should wait till the next gen of Intel and or AMD comes out. LOL what a concept... Yea?

Broadwell is virtually identical in performance to Haswell, and Skylake isn't likely to be such a massive improvement (in terms of cpu power) either to warrant waiting for it. 

[tnlich]

LOL too funny

[harrynowl]

LOL too funny

... Are you capable of forming a logical argument?

[Faa]

okay first, that program will get you kicked by punkbuster in less than 30 seconds in bf4. i tried 5 times, all 5 times i got kicked quickly.

i will however admit you are mostly correct...mostly

bf4 does start out at 6 threads and jumps up to 8 as needed, i tried to get a screen shot of 8 being used but punkbuster kept kicking me out in less than 30 seconds so all i could get is showing 7 threads being used. it does use all 8 but it does mostly use 6.

 

 

 

with dragonage inquisition surprisingly it mostly uses only 2-4 threads (3-4 avg) but will use all 8 but only for a second randomly.as you can see the last 4 threads only popped in use breifly

 

 

 

neat little program and i admit i may not be using it correctly? i am too scared to use it during any steam game as steam likes to give permanant ban for any program punkbuster doesnt like.

Looking at the cycles delta (for a given time eg 3 secs, it's not how many cycles the CPU has gone through in total) there are only 5 threads, the 6th one is like 10% of what the other threads were requiring from the CPU and anything else is so extremely minor we just ignore. The maximum load for each thread on a 8 core (100/8) is 12.5% because a thread can only use one core at a time, 5 of your threads are reaching that cap. So no, just 5 threads (even lower than what I got - I had 6) can't be executed on more than 5 cores/threads. 

This is what you would call 8 threaded;

11 threads with each 5% usage (my max is 8.33% because 100/12=8.33)

In case you are wondering, those are software threads (a stream of instructions, google it up), processexplorer shows how much CPU a thread is using, it's far more accurate than just windows taskmanager.

[WaitForMeImOnDialUp]

ahhh okay..well i guess i learned something new, i admit i dont know how to use that program or really what i was suppossed to be looking for.

[tnlich]

The point is that you are arguing in the 1st place. There is no need. The thread is dead with all the graphs and crap and bickering over what the guy should do. Be a sport and just not type anymore and ill do the same.. Yea?

[iAveex]

Wow guys we have went a long way, but this is what Linus had to say yesterday. 

http://gyazo.com/c1420df9fea5dcbe8e2f0567fc583536

[Dabombinable]

Wow guys we have went a long way, but this is what Linus had to say yesterday. 

http://gyazo.com/c1420df9fea5dcbe8e2f0567fc583536

Its capable yes, but it is showing its age.

[bcredeur97]

Get a Nahelem CPU and motherboard if you can find one.. they are still great. Just overclock it.

 

Sandy Bridge might be easier to find if its in your budget.

[79wjd]

Wow guys we have went a long way, but this is what Linus had to say yesterday. 

http://gyazo.com/c1420df9fea5dcbe8e2f0567fc583536

The thing you have to realize, is even a 5 year old core 2 quad/duo can game very well, you just have to realize if you pair it with anything high end it will hold you back quite a bit. The same is true with the FX6300 -- An fx chip will severely hold back any high end GPU in any game that isn't well threaded. 

[tnlich]

iAveex man if the 6300 is still working well. Stick with it. Your going hear all sorts of things. If the Games start creeping to a halt or unplayable then be looking to upgrade. Don't do it to egotistical hype that you see a lot on here. Save up some cash and about May be looking for the New Intel and or AMD's or look for links to upcoming stuff. I myself went from a X58 build to what I have now.  Save up... Watch Linus, Paul or them funny ....guys at Tech Syndicate on reviews and make a decision then. All my opinion.

[tnlich]

er around May that is