Overclock.net › Forums › AMD › AMD CPUs › AMD FX 8150 all the good !
New Posts  All Forums:Forum Nav:

AMD FX 8150 all the good ! - Page 42  

post #411 of 513
Oh yeah their is no question that I at least think amd 8 and 16 cores direction is more plausible then intels stupid 3d transistor technology. So if you ask me if intel delays their 8 and 10 core cpu's I can easily see another pentium D era for intel.
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
post #412 of 513
Quote:
Originally Posted by Malcom28 View Post


Full Reviews:
ASUS Crosshair V Formula 990FX 9901 BIOS
http://www.legitreviews.com/article/1741/1/

ASUS Crosshair V Formula 990FX 9901 BIOS
http://www.hitechlegion.com/reviews/processors/13752

9901 BIOS



BenchMark Result

http://www.overclockzone.com/tor_za/...nc/far1280.png

http://www.overclockzone.com/tor_za/...nc/far1920.png

http://www.overclockzone.com/tor_za/...nc/coj1280.png

http://www.overclockzone.com/tor_za/...nc/coj1920.png

http://www.overclockzone.com/tor_za/.../dirt31280.png

http://www.overclockzone.com/tor_za/.../dirt31920.png

http://www.overclockzone.com/tor_za/...150/index9.htm
Edited by cold2010 - 10/16/11 at 3:36am
post #413 of 513
Quote:
Originally Posted by BlackOmega View Post
You're right and wrong all at the same time. AMD and Intel both benefit greatly from more memory frequency and tighter timings. Well ever since Intel got AMD inside (integrated memory controller).
You are correct that AMD's do benefit greatly from tighter timings. However, since the introduction of the i CPUs' for intel, they also benefit greatly from tighter timings. For instance, 1066 CAS 5, will be noticeably faster than 1600 CAS9. This applies to both intel and AMD.
Your statement of "Intel has always been the faster RAM speed, and latency never has really effected them" is outdated. That applies to LGA 775 and before.

As for timings affecting results, I can say that fine tuning my memory frequency and timings on my C2 AM3 increased my 3Dm06 score by almost 2500 points. Approximately a 14.5% increase in performance.




Look I really don't give a crap, I've read the same reviews that you did. I was merely pointing out that you always bring this crappy Civilization BM into it --where BD did the worst. The reason why it's stupid is because they're basing it off of gameplay. Which is ridiculous because the test for each run is NOT going to be IDENTICAL. For a benchmark to be called as such, it has to be the exact same every time. That's why the BF3 beta BM wasn't a good one either.

I understand what you're trying to say with the single threaded performance. However, mutlicore CPUs' have been out for quite a long time. It's time for software dev's to step up. Although, I did read through some other stuff on the net, specifically, about the scheduler in win7. It does work pretty badly with BD, no doubt about it. As a matter of fact, it could also be the reason that BD is so power hungry. Instead of the scheduler sending it to a couple of cores that could easily do the work, it sends it out across the modules. So the CPU is utilized in an extremely inefficient manner, hogging power, and adding cycles, hence slowing it down.

As I mentioned earlier, I think it's still too early to tell. I'll probably get one though. I want to mess around with it for kicks and giggles.
For the memory latency issue, I have a 775 motherboard and I have heavily downclocked and tightened my memory just to fix my cpu bottleneck. If I ran my memory at stock 1200 cas 5,5,5,15 2T my overclock would have huge bottleneck at 3.72 ghz. When I run it at 930 cas 4,4,4,12 1T I get nearly a 10-15% boost in all cpu applications. So yeah even on the 775 sockets tighter timing is excruciatingly effective. The only problem is its not well noticed all the time because the northbridge to fsb 1:1 ratio theory is also true on 775 motherboards and if I do something else like 3:2 I loose a lot of the latency benefits, so very few with the 775 ever experience the latency improvements they do on newer boards.
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
post #414 of 513
Quote:
Originally Posted by Darkpriest667 View Post
Its funny 15 years ago software developers were practically begging hardware makers to keep up. Now the tides have turned and instead of stepping up you are whining? Really?

It is time for software developers to step up. Especially in the gaming field, which most have. This isnt 1999. Everyone that owns a computer owns a dual core or better. Heck my parents who are almost 70 have a quad core i7. This is not a hard concept to understand that almost everyone runs multiple core computers now. So it seems only logical that software developers should make almost every program multithreaded so that they run more efficiently on these multiple cores.
So people want developers to optimize the game engines so that they scale with core count and you can get 500FPS in CPU limited scenarios like 1024x760 and no filters, instead of 300FPS right?
It's not so easy anyway. It's not all about your code, there are also limitations in the programing platform you work on.

You don't get an octo core-only for gaming, it's ridiculous to go blaming developers for not taking advantage of that many cores. Gaming developers focus is on making things work good and look good in average computers. That's good coding, not trying to please us bragging right junkies

AMD knows their current lineup is more than capable of running today's and tomorrows games, ok, but I don't think they can get away with their current motto of giving up on IPC and single threaded performance by adding cores for multithreaded performance. It's not efficient transistor count/die size/power consumption wise
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
I7 2600K @ 4.8GHz. HT off Asus P8P67 Pro Gigabyte GTX480 Windforce 3 860MHz - 2000MHz 8GB Mushkin / Exceleram 2133MHz CL8 
Hard DriveOSPowerCase
WD Velociraptor 300GB / WD Caviar Green 1TB Win 7 64b Corsair HX850 Cooler Master CM 690 
  hide details  
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
I7 2600K @ 4.8GHz. HT off Asus P8P67 Pro Gigabyte GTX480 Windforce 3 860MHz - 2000MHz 8GB Mushkin / Exceleram 2133MHz CL8 
Hard DriveOSPowerCase
WD Velociraptor 300GB / WD Caviar Green 1TB Win 7 64b Corsair HX850 Cooler Master CM 690 
  hide details  
post #415 of 513
Quote:
Originally Posted by PvtHudson View Post

...


Is this benchmark even real? I7 2500k??
post #416 of 513
Umm when has software development ever followed the gamers wishes. Its not about gamers practicality, its about the average consumers practicality and quad core is becoming the laptop and desktop mainstream standard. So if everyone just gives up then we are standing still instead of moving forward for what will be practically reasonable for the future. Like I said before, one time quad core was excessive, today, dual-cores are the norm for netbooks, and almost all laptops today include quad-cores, with many being even more powerful then the cpu in my current desktop rig while also being cheaper. If you think its reasonable to say oh because today it feels unnecessary so we shouldn't try to start learning how to develop for it then how do you think we move forward in our innovation ever?
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
my system
(13 items)
 
  
CPUMotherboardGraphicsRAM
intel q9450 @ 3.72 ghz evga 750i a1 ftw two evga superclocked gtx 470's in sli gskill 4gb ddr2 1200 5-5-5-15-2T 
Hard DriveOptical DriveOSMonitor
2x western digital black 1tb raid 0 random dvd-rw/cd-rw 22x which somehow uses ata windows 7 64 + linux ubuntu spectre 24inch 
PowerCase
silverstone op1000-e (1000 watts) Silverstone Raven 2 
  hide details  
post #417 of 513
Quote:
Originally Posted by julabask View Post
Umm when has software development ever followed the gamers wishes. Its not about gamers practicality, its about the average consumers practicality and quad core is becoming the laptop and desktop mainstream standard. So if everyone just gives up then we are standing still instead of moving forward for what will be practically reasonable for the future. Like I said before, one time quad core was excessive, today, dual-cores are the norm for netbooks, and almost all laptops today include quad-cores, with many being even more powerful then the cpu in my current desktop rig while also being cheaper. If you think its reasonable to say oh because today it feels unnecessary so we shouldn't try to start learning how to develop for it then how do you think we move forward in our innovation ever?
I don't think you get my point.
Games are GPU limited, you can get 200 FPS in CPU bound configs even if just a couple of cores are being used, so there's little to be optimized most of the times. You don't want a game to have 6 or 8 cores at 100% cause you would be complaining your $500 top of the line GPU is being underused

With a lousy IPC that forces you to add a 50% - 100% more cores to match the competition you're killing scalability (more cores means more transistors and more power draw), and relying on software to utilize that many cores or some new instruction set that again, may not be suitable for every application.
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
I7 2600K @ 4.8GHz. HT off Asus P8P67 Pro Gigabyte GTX480 Windforce 3 860MHz - 2000MHz 8GB Mushkin / Exceleram 2133MHz CL8 
Hard DriveOSPowerCase
WD Velociraptor 300GB / WD Caviar Green 1TB Win 7 64b Corsair HX850 Cooler Master CM 690 
  hide details  
My System
(13 items)
 
  
CPUMotherboardGraphicsRAM
I7 2600K @ 4.8GHz. HT off Asus P8P67 Pro Gigabyte GTX480 Windforce 3 860MHz - 2000MHz 8GB Mushkin / Exceleram 2133MHz CL8 
Hard DriveOSPowerCase
WD Velociraptor 300GB / WD Caviar Green 1TB Win 7 64b Corsair HX850 Cooler Master CM 690 
  hide details  
post #418 of 513
Quote:
Originally Posted by Blackops_2 View Post
Not all algorithms can be parallelized. Much game code could like be split up though. Problem is that the threading API sucks on windows.
post #419 of 513
Quote:
Originally Posted by SCollins View Post
Not all algorithms can be parallelized. Much game code could like be split up though. Problem is that the threading API sucks on windows.
All code can be parallelized to a point.



Instead of trying to place the same problem on more cores

Increase the size of the problem so it can be parralleized more and provide better features

http://en.wikipedia.org/wiki/Gustafson's_law
Edited by Seronx - 10/16/11 at 11:08pm
AMD FX ~Seronx
(16 items)
 
  
CPUMotherboardGraphicsRAM
FX-9800P Acer Wasp R7 M440 SK Hynix HMA41GS6AFR8N-TF 
Hard DriveHard DriveOptical DriveCooling
KINGSTON RBU-SNS8152S3128GG2 TOSHIBA MQ01ABD100 HL-DT-ST DVDRAM GUE1N Stock 
OSMonitorKeyboardPower
Microsoft Windows 10 Home Build 14393 Viewsonic XG2401 24 Hz-144 Hz Ducky Channel Shine 3 Stock 65W 
CaseMouseMouse PadAudio
Acer Exoskeleton Steelseries Rival 300 Razer Megasoma AMD-Realtek ALC255 
  hide details  
AMD FX ~Seronx
(16 items)
 
  
CPUMotherboardGraphicsRAM
FX-9800P Acer Wasp R7 M440 SK Hynix HMA41GS6AFR8N-TF 
Hard DriveHard DriveOptical DriveCooling
KINGSTON RBU-SNS8152S3128GG2 TOSHIBA MQ01ABD100 HL-DT-ST DVDRAM GUE1N Stock 
OSMonitorKeyboardPower
Microsoft Windows 10 Home Build 14393 Viewsonic XG2401 24 Hz-144 Hz Ducky Channel Shine 3 Stock 65W 
CaseMouseMouse PadAudio
Acer Exoskeleton Steelseries Rival 300 Razer Megasoma AMD-Realtek ALC255 
  hide details  
post #420 of 513
Quote:
Originally Posted by Seronx View Post
All code can be parallelized to a point.



Instead of trying to place the same problem on more cores

Increase the size of the problem so it can be parralleized more and have better features

http://en.wikipedia.org/wiki/Gustafson's_law
This is a common thought problem I think many programmer face. In windows you have a primary thread, everything gets lumped in it. It would be better to have threads for each task. I know its moving that way but 40 years of programming for a crappy API has really hindered much teaching on the subject.

Its really going to come down to a better model to do this type of stuff. Be kind of started on the problem, Apple has GCD which is a move into a more easy to use model, but no one really has a good design yet that deals with the problem extensively.

First company to crack that problem, wins the race for the next decade.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD CPUs
This thread is locked  
Overclock.net › Forums › AMD › AMD CPUs › AMD FX 8150 all the good !