Overclock.net › Forums › Industry News › Hardware News › [AMD] Get Your Rig Ready for the AMD FX Processor
New Posts  All Forums:Forum Nav:

[AMD] Get Your Rig Ready for the AMD FX Processor - Page 16

post #151 of 401
Quote:
Originally Posted by Brutuz;15218748 
Yep, 47.5% increase in PCMark05 iirc by changing it to GenuineIntel from CentuarHauls, see, Intel's going to fall back on the "We didn't know you'd have a CPU that supports feature x!" argument whenever someone asks, yet if they didn't use CPUIDs to work it out and instead just went by the flags...

As a business person, I understand Intels motive to optimize only for intel and to de-optimize for everyone else. What I find completely unacceptable, is that they obfuscated this fact.
post #152 of 401
Quote:
Originally Posted by Seronx;15218529 
Cinebench

AMD FX CPUs -> SSE2
AMD Phenom II -> SSE2
Intel Core 2(45nm) -> SSE4.1
Intel i7 1st gen(45nm/32nm) -> SSE4.2
Intel i7 2nd gen(32nm) -> AVX

It does matter what CPU is used when the compiler is racist
Intel Compiler check for Genuine Intel in all versions if it isn't an Intel CPU it forces the OS/CPU to use SSE2 instead which takes extra time to convert ISA instructions to legacy ISA instructions

True. We still need to wait till release for solid, definitive benchmarks on AMD FX.
Super Over 9000!
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 5820K @ 4.4GHz EVGA X99 Micro 2 ASUS GTX 1080 Ti ROG Strix OC 32GB Corsair Vengeance LPX 2666MHz 
Hard DriveHard DriveCoolingOS
960GB Sandisk Ultra II SSD 4x 4TB WD Green Corsair H110i AIO Windows 10 x64 
MonitorPowerCase
40" Samsung UE40JU6400 (4K Ultra HD) EVGA SuperNOVA 850 G2 Fractal Node 804 
  hide details  
Reply
Super Over 9000!
(13 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 5820K @ 4.4GHz EVGA X99 Micro 2 ASUS GTX 1080 Ti ROG Strix OC 32GB Corsair Vengeance LPX 2666MHz 
Hard DriveHard DriveCoolingOS
960GB Sandisk Ultra II SSD 4x 4TB WD Green Corsair H110i AIO Windows 10 x64 
MonitorPowerCase
40" Samsung UE40JU6400 (4K Ultra HD) EVGA SuperNOVA 850 G2 Fractal Node 804 
  hide details  
Reply
post #153 of 401
Here is a ton of info from the guy who figured it out.

Enough with te debate, we will have to rely on gcc built benchmarks to make accurate assements of cpu performance given that most of the popular benchmarks are built with the Intel Compiler.

http://www.agner.org/optimize/blog/read.php?i=49
post #154 of 401
Quote:
Originally Posted by SCollins;15218853 
Here is a ton of info from the guy who figured it out.

Enough with te debate, we will have to rely on gcc built benchmarks to make accurate assements of cpu performance given that most of the popular benchmarks are built with the Intel Compiler.

http://www.agner.org/optimize/blog/read.php?i=49

Open Source and Microsoft benchmarks

http://www.cpubenchmark.net/high_end_cpus.html and conveniently one of the Microsoft compiled benchmarks is here
(Performance Test 8 will be more Representative of Intel and AMD CPUs of this generation, but Performance Test 7 shows good legacy representation)

If someone quotes me on this....The Gigabyte bios is F5 the latest bios that will probably show Representative performance is F6E for the board that was benchmarked
Edited by Seronx - 10/7/11 at 6:12pm
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  
Reply
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  
Reply
post #155 of 401
Quote:
Originally Posted by Seronx;15218529 
Cinebench

AMD FX CPUs -> SSE2
AMD Phenom II -> SSE2
Intel Core 2(45nm) -> SSE4.1
Intel i7 1st gen(45nm/32nm) -> SSE4.2
Intel i7 2nd gen(32nm) -> AVX

It does matter what CPU is used when the compiler is racist

Intel Compiler check for Genuine Intel in all versions if it isn't an Intel CPU it forces the OS/CPU to use SSE2 instead which takes extra time to convert ISA instructions to legacy ISA instructions
Yeah I knew that the Intel Compiler check look for a Intel CPU,if a VIA/AMD CPU is present,the CPU has to "take the long way around" so to speak.
Quote:
Originally Posted by SCollins;15218623 
on cinebench 10.5 this was certainly the case as cinebecnh uses the ICC. I can't speak for the new cinebench 11.x something, but you might want to look up Agner and intel compiler on google for more info.
I looked it up,it is really interesting because Intel was supposed to stop using the ICC in their bentmarks after the Intel vs. AMD suit.
Quote:
Originally Posted by SCollins;15218705 
No, it generates less optimized binarys. I have no idea on the compile time. I imagine that the performance difference in the binary may be more or less tied to the type of instructions in use. GCC is very brand agnostic and it will allow you to compile multiple code paths IIRC. I use it in a very novice way so it could be far more capable then I am aware of.

What interesting is that benchmarks built with gcc vrs Intel compiler, generate very different results.

Really though, google (agner, Intel Compiler" and see for yourself.

IIRC ARS technica did a article about spoofing cpuid on a via nano a few years back, the results were shocking !!!
Wasn't it Intel Atom vs. the VIA Nano?
Quote:
Originally Posted by Brutuz;15218748 
Yep, 47.5% increase in PCMark05 iirc by changing it to GenuineIntel from CentuarHauls, see, Intel's going to fall back on the "We didn't know you'd have a CPU that supports feature x!" argument whenever someone asks, yet if they didn't use CPUIDs to work it out and instead just went by the flags...
47.5%? eek.gif If this could be done with a Intel vs. AMD CPU the results would be very surprising.
Edited by Heavy MG - 10/7/11 at 6:12pm
Moar cores!
(16 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX 8320 @ 4.0Ghz/4.2Ghz turbo gigabyte 990FXA-UD3 rev. 4.0 Gigabyte Nvida GTX 770OC 2GB/ Gigabyte 970 G1 G... Gskill Sniper 1866Mhz 8GB (4GB x 2) 
Hard DriveHard DriveOptical DriveCooling
Corsair Neutron GTX 120GB WD Black 1TB Asus 24X CM Hyper 212Evo 
OSMonitorKeyboardPower
Windows 7 Home X64 HP 22BW 21.5" LED-IPS CM Storm Trigger/MX Brown Corsair TX 650V2 
CaseMouseMouse PadAudio
CM HAF922 Logitech G500S Steelseries QCK Mini Asus Xonar DX 7.1 
  hide details  
Reply
Moar cores!
(16 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX 8320 @ 4.0Ghz/4.2Ghz turbo gigabyte 990FXA-UD3 rev. 4.0 Gigabyte Nvida GTX 770OC 2GB/ Gigabyte 970 G1 G... Gskill Sniper 1866Mhz 8GB (4GB x 2) 
Hard DriveHard DriveOptical DriveCooling
Corsair Neutron GTX 120GB WD Black 1TB Asus 24X CM Hyper 212Evo 
OSMonitorKeyboardPower
Windows 7 Home X64 HP 22BW 21.5" LED-IPS CM Storm Trigger/MX Brown Corsair TX 650V2 
CaseMouseMouse PadAudio
CM HAF922 Logitech G500S Steelseries QCK Mini Asus Xonar DX 7.1 
  hide details  
Reply
post #156 of 401
When BD is released:

party-hard-3.gif
post #157 of 401
Quote:
Originally Posted by Heavy MG;15218901 
Yeah I knew that the Intel Compiler check look for a Intel CPU,if a VIA/AMD CPU is present,the CPU has to "take the long way around" so to speak.

I looked it up,it is really interesting because Intel was supposed to stop using the ICC in their bentmarks after the Intel vs. AMD suit.

Wasn't it Intel Atom vs. the VIA Nano?

Yes on the long way around comment, also the slow way around

I don't know whats in that settlement.

I belive you are correct, Atom vrs Nano.
post #158 of 401
Quote:
Originally Posted by Seronx;15218898 
Open Source and Microsoft benchmarks

http://www.cpubenchmark.net/high_end_cpus.html and conveniently one of the Microsoft compiled benchmarks is here
(Performance Test 8 will be more Representative of Intel and AMD CPUs of this generation, but Performance Test 7 shows good legacy representation)

If someone quotes me on this....The Gigabyte bios is F5 the latest bios that will probably show Representative performance is F6E for the board that was benchmarked


What exactly does this page tells us ? what tests, how are they weighted ? What is the statistical significance and confidence ?
How do they even arrive at this measurement ?

I mean, how does each cpu stack up in avx/sse/etc etc etc.

I can't make any sense out of that chart and why are server CPU's in there ?
post #159 of 401
Quote:
Originally Posted by SCollins;15218976 
Yes on the long way around comment, also the slow way around

I don't know whats in that settlement.

I belive you are correct, Atom vrs Nano.

The Agner blog post dated 2010-9-22 11:37 said something about a new Intel compiler version,despite the legal battles,the CPU optimizations are unchanged.
Moar cores!
(16 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX 8320 @ 4.0Ghz/4.2Ghz turbo gigabyte 990FXA-UD3 rev. 4.0 Gigabyte Nvida GTX 770OC 2GB/ Gigabyte 970 G1 G... Gskill Sniper 1866Mhz 8GB (4GB x 2) 
Hard DriveHard DriveOptical DriveCooling
Corsair Neutron GTX 120GB WD Black 1TB Asus 24X CM Hyper 212Evo 
OSMonitorKeyboardPower
Windows 7 Home X64 HP 22BW 21.5" LED-IPS CM Storm Trigger/MX Brown Corsair TX 650V2 
CaseMouseMouse PadAudio
CM HAF922 Logitech G500S Steelseries QCK Mini Asus Xonar DX 7.1 
  hide details  
Reply
Moar cores!
(16 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX 8320 @ 4.0Ghz/4.2Ghz turbo gigabyte 990FXA-UD3 rev. 4.0 Gigabyte Nvida GTX 770OC 2GB/ Gigabyte 970 G1 G... Gskill Sniper 1866Mhz 8GB (4GB x 2) 
Hard DriveHard DriveOptical DriveCooling
Corsair Neutron GTX 120GB WD Black 1TB Asus 24X CM Hyper 212Evo 
OSMonitorKeyboardPower
Windows 7 Home X64 HP 22BW 21.5" LED-IPS CM Storm Trigger/MX Brown Corsair TX 650V2 
CaseMouseMouse PadAudio
CM HAF922 Logitech G500S Steelseries QCK Mini Asus Xonar DX 7.1 
  hide details  
Reply
post #160 of 401
Quote:
Originally Posted by SCollins;15219015 
What exactly does this page tells us ? what tests, how are they weighted ? What is the statistical significance and confidence ?
How do they even arrive at this measurement ?

The higher the score the higher the Multitasking/Multi-threaded potential
Quote:
Originally Posted by SCollins;15219015 
I mean, how does each cpu stack up in avx/sse/etc etc etc.

The benchmark pretty much uses hand coded instructions and I do not know how high in the ISA level it goes up to(probably just SSE4.2)
Quote:
Originally Posted by SCollins;15219015 
I can't make any sense out of that chart and why are server CPU's in there ?

It is a benchmark...so of course all types of CPUs are tested
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  
Reply
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  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [AMD] Get Your Rig Ready for the AMD FX Processor