Overclock.net › Forums › Industry News › Technology and Science News › [CT] AMD: x86 Benchmarks Do Not Matter Anymore
New Posts  All Forums:Forum Nav:

[CT] AMD: x86 Benchmarks Do Not Matter Anymore - Page 2  

post #11 of 59
I wouldn't be at all surprised if the new AMD chips benchmark poorly. Previous AMD chips have benchmarked poorly, yet still had excellent real world performance, so I completely agree with this statement from them, benchmarks are not real world and so much weight should not be placed on them.
post #12 of 59
Quote:
Originally Posted by blackbalt89 View Post
*yawn*

Next AMD is going to say is that Bulldozer/Llano will divide by zero.
Technically you can divide by zero, if you're using floating points. The result is NaN (Not a number), or in some cases +/- Infinity. Integer math doesn't support this because all combinations of bits are used for legitimate numbers, so dividing an int by zero is almost always a recipe for disaster.
Edited by lordikon - 3/1/11 at 7:39pm
Foldatron
(17 items)
 
Mat
(10 items)
 
Work iMac
(9 items)
 
CPUMotherboardGraphicsGraphics
i7 950 EVGA x58 3-way SLI EVGA GTX 660ti GTX 275 
RAMHard DriveHard DriveHard Drive
3x2GB Corsair Dominator DDR3-1600 80GB Intel X25-M SSD 2TB WD Black 150GB WD Raptor 
Hard DriveOSMonitorKeyboard
2x 150GB WD V-raptor in RAID0 Win7 Home 64-bit OEM 55" LED 120hz 1080p Vizio MS Natural Ergonomic Keyboard 4000 
PowerCase
750W PC P&C Silencer CoolerMaster 690 
CPUGraphicsRAMHard Drive
Intel Core i5 2500S AMD 6770M 8GB (2x4GB) at 1333Mhz 1TB, 7200 rpm 
Optical DriveOSMonitorKeyboard
LG 8X Dual-Layer "SuperDrive" OS X Lion 27" iMac screen Mac wireless keyboard 
Mouse
Mac wireless mouse 
CPUGraphicsRAMHard Drive
i7-2600K AMD 6970M 1GB 16GB PC3-10600 DDR3 1TB 7200rpm 
Hard DriveOptical DriveOSMonitor
256GB SSD 8x DL "SuperDrive" OS X 10.7 Lion 27" 2560x1440 iMac display 
Monitor
27" Apple thunderbolt display 
  hide details  
Foldatron
(17 items)
 
Mat
(10 items)
 
Work iMac
(9 items)
 
CPUMotherboardGraphicsGraphics
i7 950 EVGA x58 3-way SLI EVGA GTX 660ti GTX 275 
RAMHard DriveHard DriveHard Drive
3x2GB Corsair Dominator DDR3-1600 80GB Intel X25-M SSD 2TB WD Black 150GB WD Raptor 
Hard DriveOSMonitorKeyboard
2x 150GB WD V-raptor in RAID0 Win7 Home 64-bit OEM 55" LED 120hz 1080p Vizio MS Natural Ergonomic Keyboard 4000 
PowerCase
750W PC P&C Silencer CoolerMaster 690 
CPUGraphicsRAMHard Drive
Intel Core i5 2500S AMD 6770M 8GB (2x4GB) at 1333Mhz 1TB, 7200 rpm 
Optical DriveOSMonitorKeyboard
LG 8X Dual-Layer "SuperDrive" OS X Lion 27" iMac screen Mac wireless keyboard 
Mouse
Mac wireless mouse 
CPUGraphicsRAMHard Drive
i7-2600K AMD 6970M 1GB 16GB PC3-10600 DDR3 1TB 7200rpm 
Hard DriveOptical DriveOSMonitor
256GB SSD 8x DL "SuperDrive" OS X 10.7 Lion 27" 2560x1440 iMac display 
Monitor
27" Apple thunderbolt display 
  hide details  
post #13 of 59
Quote:
Originally Posted by mushroomboy View Post
Except that type of power gives you nothing for performance in the laptop world. Not to mention the inability to have that type of power. Nor does it do it even for desktop, unless you just happen to have a server room. Yes, power IS everything, just not a specific type of power being everything.
Ok alright then, lets have a "friendly" competition, you can have a nuclear weapon and no gun. But I get a EMP and a pistol. EMP knocks out your computer and I come with pistol. Game over.

It's not about having the most powerful, but the most effective.

Having the most powerful thing is not effective unless you can use it to it's full potential. They are comparing the diffrences of a apu doing GPU and cpu calculations on a single die, that can't be benchmarked the same way as sandy bridge. To test the apu completely you need something like futuremark since it tests graphics and cpu. To test a sandy bridge all you need is wprime or super pi.

Woot at 256'th post!
Edited by Maelthras - 3/1/11 at 7:45pm
post #14 of 59
I don't know about you guys but I think AMD is doing it just right here

The integrated graphics in Llano will be called the HD 6620M according to another article and official AMD video. Going by naming scheme only, this is significantly better than most of the discrete HD 6XXXM parts already out. CPU performance isn't exactly big anymore unless you're doing apps like video editing or CPU folding, so they're right - which is why the difference between Phenom II and i7 in a gaming build is really really negligible. And guys, for the very least, they still are catering to us who may just be CPU hugs like me with my 2 1080p video renders at once or so. Bulldozer is on the horizon.
post #15 of 59
I pretty much agree with AMD regardless of their reasoning for making this statement.

I've gone from a 3.8GHz Pentium 4 to a 1.6GHz Core2Duo to a 3.6GHz Athlon II X4 to a 3.8GHz X6.

From the P4 to the Core2Duo, I honestly didn't notice that much performance boost even though the C2D would slam the P4 in benches.

From the C2D to the Athlon II, I noticed a vast jump in everyday performance (web, application speed, etc.).

From the Athlon II to the X6, I gained huge amounts of number crunching ability, but once again, I noticed almost no real world benefit.

From the common consumer perspective, I would agree with AMD in that any modern CPU is pretty much 100% capable of everything a user will ask of it.

From an enthusiast perspective, obviously we want to have the fastest CPU in town, but it makes little to no difference in 90% of our computer usage. And on top of that, no true enthusiast is going to buy a CPU with a graphics core built into it, and AMD knows this. They aren't targeting this community with Llano.

In addition, I feel that AMD putting heavy emphasis on the graphics/parallel computing aspect of these new CPU's is going to win out in the end. Everything is becoming more GPU intensive (HD movies, Flash, games, etc.) Common consumers will notice increased GPU horsepower over increased CPU power any day of the week.

To illustrate this more effectively, take for instance two car companies:

Company A takes their 500HP land-rocket and improves it for next year by upping the output to 600WHP. (Intel, modern muscle cars... camaro, viper, etc.)

Company B takes their 400HP land-rocket and improves it for next year by adding suspension and an advanced braking system that allows the car to handle better in a wider variety of scenarios. (AMD, modern sports cars... R8, Lotus, Ferrari, Porsche, etc.)
Sandy Beach
(14 items)
 
  
CPUMotherboardGraphicsRAM
2500K @ 4.5GHz/1.32v (Tinkering) P8P67-PRO [BIOS 1704] MSI GTX780 TFIV G.Skill 4x4GB DDR3-1600 (8-8-8-24-1T) 
Hard DriveCoolingOSMonitor
256GB M4, WD Black 1TB & ES.2 1TB & Cuda 3TB XSPC Rasa w/ RS240 Windows 7 Ultimate x64 Dell SX2210 
KeyboardPowerCaseMouse
Filco Ninja w/ Browns SilverStone Strider Plus 750W Custom Stacker 832 Logitech G500 
Mouse Pad
Steelseries 4HD 
  hide details  
Sandy Beach
(14 items)
 
  
CPUMotherboardGraphicsRAM
2500K @ 4.5GHz/1.32v (Tinkering) P8P67-PRO [BIOS 1704] MSI GTX780 TFIV G.Skill 4x4GB DDR3-1600 (8-8-8-24-1T) 
Hard DriveCoolingOSMonitor
256GB M4, WD Black 1TB & ES.2 1TB & Cuda 3TB XSPC Rasa w/ RS240 Windows 7 Ultimate x64 Dell SX2210 
KeyboardPowerCaseMouse
Filco Ninja w/ Browns SilverStone Strider Plus 750W Custom Stacker 832 Logitech G500 
Mouse Pad
Steelseries 4HD 
  hide details  
post #16 of 59
Don't the benchmarks have an Intel advantage? What was that software that checks for intel genuine and if not, chooses poor optimization?
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  
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  
post #17 of 59
If you would read the article it clearly states that x86 benchmarks are still important, but now they are taking video, compute, graphics and x86 all in balance.

they aren't going to focus just solely on x86 anymore, and i am pretty sure hes just referring to llano; in case which they shouldn't just focus on x86.

Quote:
He stressed that AMD will continue to invest in x86, especially as Bulldozer is concerned, while looking for a “balance of graphics, video, compute and x86 capabilities.â€
READ THE ARTICLE YOU FOOLS
No Name
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i5 760 batch: L034B172 EVGA P55 LE EVGA GTX 460 EE@ 862Mhz Core (air) 4GB Dominator DDR3@ 1333Mhz 
Hard DriveOptical DriveOSMonitor
500GB Samsung Spinpoint F4 POS OEM Dell DVD burner Windows 7 Ultimate x64 Compaq S1922@ 1366x768 
KeyboardPowerCaseMouse
Saitek Eclipse II Antec 650w Truepower NZXT M59 Razer Deathadder 
Mouse Pad
Qck Mass 
  hide details  
No Name
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i5 760 batch: L034B172 EVGA P55 LE EVGA GTX 460 EE@ 862Mhz Core (air) 4GB Dominator DDR3@ 1333Mhz 
Hard DriveOptical DriveOSMonitor
500GB Samsung Spinpoint F4 POS OEM Dell DVD burner Windows 7 Ultimate x64 Compaq S1922@ 1366x768 
KeyboardPowerCaseMouse
Saitek Eclipse II Antec 650w Truepower NZXT M59 Razer Deathadder 
Mouse Pad
Qck Mass 
  hide details  
post #18 of 59
Quote:
Originally Posted by Z Overlord View Post
Don't the benchmarks have an Intel advantage? What was that software that checks for intel genuine and if not, chooses poor optimization?
Don't be silly it's not about synthetic benchmarks anymore.

AMD is evolving and Intel is not. Everything Intel has done up to this point is brute force computing whereas AMD is taking a more integrated approach.

What matters at the end of the day is real world performance for the end user, not a bunch of stupid people arguing about which configuration is faster based on a bunch of incoherent benchmarks.
Shhhhh!
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 930 Gigabyte GA-X58A-UD3R XFX 5870 1GB 6GB G.SKILL PI 1600MHz 
Hard DriveOptical DriveOSMonitor
120GB OCZ Vertex 2, 2x Hitachi 2TB (7K2000) Pioneer Blu-ray BDR-205BKS, Samsung SH-S243N Windows 7 Ultimate x64/WinXP Pro SP3 Samsung P2570 
KeyboardPowerCaseMouse
Logitech Elite Keyboard Corsair 850HX Lian-Li PC-B70 Logitech MX510 
Mouse Pad
Gel 
  hide details  
Shhhhh!
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 930 Gigabyte GA-X58A-UD3R XFX 5870 1GB 6GB G.SKILL PI 1600MHz 
Hard DriveOptical DriveOSMonitor
120GB OCZ Vertex 2, 2x Hitachi 2TB (7K2000) Pioneer Blu-ray BDR-205BKS, Samsung SH-S243N Windows 7 Ultimate x64/WinXP Pro SP3 Samsung P2570 
KeyboardPowerCaseMouse
Logitech Elite Keyboard Corsair 850HX Lian-Li PC-B70 Logitech MX510 
Mouse Pad
Gel 
  hide details  
post #19 of 59
Quote:
Originally Posted by Maelthras View Post
Ok alright then, lets have a "friendly" competition, you can have a nuclear weapon and no gun. But I get a EMP and a pistol. EMP knocks out your computer and I come with pistol. Game over.

It's not about having the most powerful, but the most effective.

Having the most powerful thing is not effective unless you can use it to it's full potential. They are comparing the diffrences of a apu doing GPU and cpu calculations on a single die, that can't be benchmarked the same way as sandy bridge. To test the apu completely you need something like futuremark since it tests graphics and cpu. To test a sandy bridge all you need is wprime or super pi.

Woot at 256'th post!
Ummm, is your first language english? That's what I was saying, or at least inferring...

But if you want to get technical, my nuke can still go off whether or not my computer works. In fact you can make a nuke without the use of any real electronics, you can use a fully mechanical detonator. I drop the nuke out of a plane and you still die, but you know lets make this friendly.

Just to let you know how this works:

http://www.zamandayolculuk.com/cetin...mbombalari.htm

That was little boy (had to look it up, didn't remember which one). I can actually list the parts required and how to make them. =) A bit of engineering smarts I have picked up, amongst other things. So sure, give me the atomic bomb and I'll make sure you don't use that gun.

[edit]

Quote:
Originally Posted by tosh.0 View Post
If you would read the article it clearly states that x86 benchmarks are still important, but now they are taking video, compute, graphics and x86 all in balance.

they aren't going to focus just solely on x86 anymore, and i am pretty sure hes just referring to llano; in case which they shouldn't just focus on x86.



READ THE ARTICLE YOU FOOLS
Except you can now use the GPU to do those benchmarks, since it's not really a GPU it is part of the APU now. It's not even a separate die either, it's literally the same chip. If you can code a benchmark program that can use that to it's advantage yet still compute the same numerical equations I'm betting we know who wins. I'm sure we will soon see a new mix of coding styles emerge for these chips.
Edited by mushroomboy - 3/1/11 at 8:54pm
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  
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  
post #20 of 59
Quote:
Originally Posted by lordikon View Post
Technically you can divide by zero, if you're using floating points. The result is NaN (Not a number), or in some cases +/- Infinity. Integer math doesn't support this because all combinations of bits are used for legitimate numbers, so dividing an int by zero is almost always a recipe for disaster.
Divide by zero is undetermined.

Software implementation of the result is what you are talking about... but that does not negate the fact that the result is undetermined regardless of how a computer represents this.
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
Once again...
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 920 [4.28GHz, HT] Asus P6T + Broadcom NetXtreme II VisionTek HD5850 [900/1200] + Galaxy GT240 2x4GB G.Skill Ripjaw X [1632 MHz] 
Hard DriveOSMonitorKeyboard
Intel X25-M 160GB + 3xRAID0 500GB 7200.12 Window 7 Pro 64 Acer H243H + Samsung 226BW XARMOR-U9BL  
PowerCaseMouseMouse Pad
Antec Truepower New 750W Li Lian PC-V2100 [10x120mm fans] Logitech G9 X-Trac Pro 
  hide details  
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Technology and Science News
This thread is locked  
Overclock.net › Forums › Industry News › Technology and Science News › [CT] AMD: x86 Benchmarks Do Not Matter Anymore