Overclock.net › Forums › Industry News › Technology and Science News › [AMD] Exposing the Phantom x86 Bottleneck
New Posts  All Forums:Forum Nav:

[AMD] Exposing the Phantom x86 Bottleneck - Page 4

post #31 of 42
Quote:
Originally Posted by almighty15 View Post
Crysis, Warhead, STALKER series......80% of the games I play are only dual threaded and while for the most part performance is really good. When the on screen action heats up or things explode the frames drop down below 30, sometimes way below 30...
No, Crysis use 4 threads.

Quote:
Originally Posted by Starbuck5000 View Post
Ewwwwwwwwwwwww transcoding on a GPU
What's wrong with transcoding on a GPU? TBH, I don't care if apps use my cpu or gpu, either chip is overkill for most cpu/gpu using apps.
Edited by Z Overlord - 3/3/11 at 11:16am
Ice Wings
(14 items)
 
  
CPUMotherboardGraphicsRAM
i5 3570K @ 4.5GHz Asus Sabertooth Z77 Asus Nvidia Geforce GTX 980 Samsung Green MV-3V4G3D/US 8GB 2x4GB DDR3 1600MHZ 
Hard DriveHard DriveCoolingOS
OCZ Agility 3 240GB  Samsung F4 HD204UI Noctua NH-D14 Windows 10 Pro 64 bit 
MonitorKeyboardPowerCase
Asus RoG Swift PG278Q Ducky One (DKON1687) MX Reds w/ PBT keycaps. SeaSonic X Series X650 Gold Fractal Design Define R4 
MouseMouse Pad
Logitech G303 Puretrak Talent  
  hide details  
Reply
Ice Wings
(14 items)
 
  
CPUMotherboardGraphicsRAM
i5 3570K @ 4.5GHz Asus Sabertooth Z77 Asus Nvidia Geforce GTX 980 Samsung Green MV-3V4G3D/US 8GB 2x4GB DDR3 1600MHZ 
Hard DriveHard DriveCoolingOS
OCZ Agility 3 240GB  Samsung F4 HD204UI Noctua NH-D14 Windows 10 Pro 64 bit 
MonitorKeyboardPowerCase
Asus RoG Swift PG278Q Ducky One (DKON1687) MX Reds w/ PBT keycaps. SeaSonic X Series X650 Gold Fractal Design Define R4 
MouseMouse Pad
Logitech G303 Puretrak Talent  
  hide details  
Reply
post #32 of 42
normally transcoding with a gpu degrade the quality compared to just cpu transcoding
    
CPUMotherboardGraphicsRAM
i7 2600k 4.6ghz 1.376v Msi Z68A-G43 (cheap board!!) MSI 390 Gaming 16gb G.skill Ripjaws X 1866 
Hard DriveOptical DriveCoolingOS
Multiple HHD and SSD over 12tb LG GGC-H20L Custom cpu water loop Windows 8 pro 
MonitorPowerCase
AOC u2879vf Corsair tx-750 HAF 912 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i7 2600k 4.6ghz 1.376v Msi Z68A-G43 (cheap board!!) MSI 390 Gaming 16gb G.skill Ripjaws X 1866 
Hard DriveOptical DriveCoolingOS
Multiple HHD and SSD over 12tb LG GGC-H20L Custom cpu water loop Windows 8 pro 
MonitorPowerCase
AOC u2879vf Corsair tx-750 HAF 912 
  hide details  
Reply
post #33 of 42
Quote:
Originally Posted by DuckieHo View Post
Interesting, I would think HPC workloads would most be FP and DP-FP?
http://www.anandtech.com/show/2881

80% is what they claimed is pure interger work. As a matter of optimizing you try to avoid floating points at every turn so it doesn't surprise me in any way whatsoever. If you think about what an average server does in business it's pretty easy to see why floating points are relatively unimportant. If you consider what work must be involved in domain controllers, file servers, phone servers, etc anyways. Which is the type of thing I work with. There's no need for any floating points in most of the things a server would do in the process of tending to those tasks.
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
post #34 of 42
Quote:
Originally Posted by Z Overlord View Post
What's wrong with transcoding on a GPU? TBH, I don't care if apps use my cpu or gpu, either chip is overkill for most cpu/gpu using apps.
What isn't wrong with transcoding on a GPU...:
http://www.anandtech.com/show/4083/t...-2100-tested/9

especially one from nvidia
DARTH
(19 items)
 
  
CPUMotherboardGraphicsGraphics
Intel 4770K Gigabyte Z87X-UD4H Diamond R9 290X Asus R9 290 Reference 
RAMHard DriveHard DriveHard Drive
Corsair Vengeance CMZ8GX3M2A1866C9 Intel 320 Series SSD Seagate 1 Tb 7.2k WD Blue 
Optical DriveCoolingCoolingCooling
2x Liteon 52x 32x 52x XSPC RASA 240RS kit XSPC RX480 v2 EK Full Cover R9 290X waterblocks 
OSMonitorKeyboardPower
Windows 7 64bit Ultimate Dell P2715Q Ducky Zero DK2108 (Red Switches) Corsair RM1000 
CaseMouseMouse Pad
Thermaltake Core x9 Micro$oft X5 Desk 
  hide details  
Reply
DARTH
(19 items)
 
  
CPUMotherboardGraphicsGraphics
Intel 4770K Gigabyte Z87X-UD4H Diamond R9 290X Asus R9 290 Reference 
RAMHard DriveHard DriveHard Drive
Corsair Vengeance CMZ8GX3M2A1866C9 Intel 320 Series SSD Seagate 1 Tb 7.2k WD Blue 
Optical DriveCoolingCoolingCooling
2x Liteon 52x 32x 52x XSPC RASA 240RS kit XSPC RX480 v2 EK Full Cover R9 290X waterblocks 
OSMonitorKeyboardPower
Windows 7 64bit Ultimate Dell P2715Q Ducky Zero DK2108 (Red Switches) Corsair RM1000 
CaseMouseMouse Pad
Thermaltake Core x9 Micro$oft X5 Desk 
  hide details  
Reply
post #35 of 42
Quote:
Originally Posted by AMDMAXX View Post
What isn't wrong with transcoding on a GPU...:
http://www.anandtech.com/show/4083/t...-2100-tested/9

especially one from nvidia
Biased article, they use the 460 and not the 560 to compare with a 6879. They should compare a 5XXX to the 4XX cards, instead of intentionally making Nvidia look bad.

They also don't go into detail whether or not they had options for better precision at a lower cost of frame rates. They used the same software, which doesn't tell us if there is a bias for the programmer either. One program could transcode terrible for Nvidia and amazing for AMD, the other could do different. It all depends on what they decided to optimize more. I'm not saying GPU transcoding is amazing, rather that this article doesn't give it justice.
Current Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
FX-8350 4.6GHz@1.44v GA-990FXA-UD3 R4.0 HD 7950 (1100/1450) 8G Muskin DDR3 1866@8CLS 
Hard DriveOptical DriveOSMonitor
1TB WD LiteOn DVD-RW DL Linux/Windows 19" Phillips TV 1080p 
PowerCaseMouseMouse Pad
OCZ 600W Generic Junk Logitech MX400 Generic Junk 
Audio
SBL 5.1 
  hide details  
Reply
Current Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
FX-8350 4.6GHz@1.44v GA-990FXA-UD3 R4.0 HD 7950 (1100/1450) 8G Muskin DDR3 1866@8CLS 
Hard DriveOptical DriveOSMonitor
1TB WD LiteOn DVD-RW DL Linux/Windows 19" Phillips TV 1080p 
PowerCaseMouseMouse Pad
OCZ 600W Generic Junk Logitech MX400 Generic Junk 
Audio
SBL 5.1 
  hide details  
Reply
post #36 of 42
Quote:
Originally Posted by mushroomboy View Post
Biased article, they use the 460 and not the 560 to compare with a 6879. They should compare a 5XXX to the 4XX cards, instead of intentionally making Nvidia look bad.
The 560 wasn't out at the time of the article.
GE60
(13 items)
 
  
CPUGraphicsHard DriveOS
Intel Core i5 4200M Geforce GTX760M 750GB WD Scorpio Black Windows 7 Home Premium 
MonitorKeyboardMouseMouse Pad
15.6" CM Quickfire Rapid CM Xornet Steelseries QcK+ 
  hide details  
Reply
GE60
(13 items)
 
  
CPUGraphicsHard DriveOS
Intel Core i5 4200M Geforce GTX760M 750GB WD Scorpio Black Windows 7 Home Premium 
MonitorKeyboardMouseMouse Pad
15.6" CM Quickfire Rapid CM Xornet Steelseries QcK+ 
  hide details  
Reply
post #37 of 42
Quote:
Originally Posted by sLowEnd View Post
The 560 wasn't out at the time of the article.
So he couldn't have waited less than a month to write an article?

http://www.tomshardware.com/reviews/...nc,2839-7.html

Still was biased, and I'm guessing he wasn't doing a fair comparison for CPU vs GPU either.
Current Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
FX-8350 4.6GHz@1.44v GA-990FXA-UD3 R4.0 HD 7950 (1100/1450) 8G Muskin DDR3 1866@8CLS 
Hard DriveOptical DriveOSMonitor
1TB WD LiteOn DVD-RW DL Linux/Windows 19" Phillips TV 1080p 
PowerCaseMouseMouse Pad
OCZ 600W Generic Junk Logitech MX400 Generic Junk 
Audio
SBL 5.1 
  hide details  
Reply
Current Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
FX-8350 4.6GHz@1.44v GA-990FXA-UD3 R4.0 HD 7950 (1100/1450) 8G Muskin DDR3 1866@8CLS 
Hard DriveOptical DriveOSMonitor
1TB WD LiteOn DVD-RW DL Linux/Windows 19" Phillips TV 1080p 
PowerCaseMouseMouse Pad
OCZ 600W Generic Junk Logitech MX400 Generic Junk 
Audio
SBL 5.1 
  hide details  
Reply
post #38 of 42
AnandTech is about as unbiased as they come, usually. I don't see any side-taking going on in that article, really.
post #39 of 42
Quote:
Originally Posted by mushroomboy View Post
Biased article, they use the 460 and not the 560 to compare with a 6879. They should compare a 5XXX to the 4XX cards, instead of intentionally making Nvidia look bad.

They also don't go into detail whether or not they had options for better precision at a lower cost of frame rates. They used the same software, which doesn't tell us if there is a bias for the programmer either. One program could transcode terrible for Nvidia and amazing for AMD, the other could do different. It all depends on what they decided to optimize more. I'm not saying GPU transcoding is amazing, rather that this article doesn't give it justice.
anandtech is hardly biased.
Both programs they use are the only one that supports the quicksync intel has, and the article is about the i5 2500k.
Theres no bias in that.
    
CPUMotherboardGraphicsRAM
i7 2600k 4.6ghz 1.376v Msi Z68A-G43 (cheap board!!) MSI 390 Gaming 16gb G.skill Ripjaws X 1866 
Hard DriveOptical DriveCoolingOS
Multiple HHD and SSD over 12tb LG GGC-H20L Custom cpu water loop Windows 8 pro 
MonitorPowerCase
AOC u2879vf Corsair tx-750 HAF 912 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i7 2600k 4.6ghz 1.376v Msi Z68A-G43 (cheap board!!) MSI 390 Gaming 16gb G.skill Ripjaws X 1866 
Hard DriveOptical DriveCoolingOS
Multiple HHD and SSD over 12tb LG GGC-H20L Custom cpu water loop Windows 8 pro 
MonitorPowerCase
AOC u2879vf Corsair tx-750 HAF 912 
  hide details  
Reply
post #40 of 42
Quote:
Originally Posted by Krusher33 View Post
More "Oh my noez the framerates dropped so it must be my cpu!". Why? Why does one always come to this conclusion???
Easy, you can watch the gpu usage drop in cpu intensive areas of games whereas a higher end cpu would keep the gpu usage at 99%.(I'm talking modern games here, I know there are older games that won't hit 99% on modern gpus) It obviously all depends on the games being tested as well. Games that actually utilize 4+ threads have far less bottlenecking issues. It's when the game only utilizes 1-3 threads that bottlenecking becomes apparent on phenom II's with high end video card setups. The serious IPC deficit causes the cpu not to send instructions fast enough to the gpu to keep the gpu running at its optimal speed. that being said bottlenecks do not mean unplayable, it simply means a component is preventing another from running at its optimal speed. It's people here that gave it ridiculous meanings to try and claim bottlenecks don't exist.
Edited by scyy - 3/3/11 at 1:57pm
3930k
(20 items)
 
  
Reply
3930k
(20 items)
 
  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Technology and Science News
Overclock.net › Forums › Industry News › Technology and Science News › [AMD] Exposing the Phantom x86 Bottleneck