Overclock.net › Forums › AMD › AMD CPUs › The Zambezi CPU Performance Estimate...
New Posts  All Forums:Forum Nav:

The Zambezi CPU Performance Estimate... - Page 3

post #21 of 354
Thread Starter 
Quote:
Originally Posted by amd-dude;12311629 
My only question is...why so many bulldozer haters...if it were not for competition between intel and amd we would be still be paying $1000+ for a entry level system. Obviously intel has performance card on amd, but anyone can buy a good amd chip lets say the 965BE and be able to play games relatively easy on their system.

Now don't get me wrong sandy bridge with the 2600k priced at $300 is a killer bargain...and thanks to its 5ghz+ ability this invites even more competition from amd...lets say that bulldozer is 50% faster than the i7 950...but what if they can offer that kind of performance for sub $200 prices...i would not mind that at all.

But back to the OP...i was looking at SIS soft sandra the other day and i saw how they rate cpus based on the GIPS i.e. 980x is 132GIPS if i'm not mistaken...now take into account that my sig rig specs get 62GIPS and i can run pretty much anything i want what does one do with an extra 70GIPS, benchmark all day? Honestly i think that everyone these days is talking wooo performance...but aren't you wasting money by having 70GIPS just waiting there for something to do? Is there anyone here who can say that they use 100% of their cpu all the time...my humble 555 only uses 50% while gaming and thats the hardest my system ever runs....but again to each his own...I just sayin'

You have a valid point IF all you do is gaming, video encoding, archive compression etc, not much of a gain. (what 3-10 seconds shaved off?)

When you start compiling programs then you start to see a significant gain in real world applications...

Not everyone can utilize an Uber CPU, it takes tools & you should have the NEED for such a monsters CPU, like a server etc..

AMD made it very clear that, it's not about the GHz and clock speed for a Desktop Computer, but how well it can multi-task, this is why they went for more cores & essentially more processes over a faster speed, nevertheless, they did make improvements on speed as well as power, but more importantly they clearly stated that it would be competing with Intel in price too! <<

I look at Intel's decision to speed up Ive Bridge as a sign that they FEAR THE BULLDOZER, lol, but I could be wrong.. biggrin.gif

Good post, thanks..
Edited by _GTech - 2/8/11 at 12:15pm
The Rock
(15 items)
 
  
Reply
The Rock
(15 items)
 
  
Reply
post #22 of 354
Yea i think and i honestly hope that bulldozer can level the playing field a bit cause i am tired of the intel boys (fan boys, not regular ones) saying 'oh well we beat you amd chip anytime'

I get what your saying about compiling and i have done some test for myself here as well..my chip just takes way to long to compile code in backtrack 4 or even if im using pyrat or cain and abel to crack hashes for demonstrations to clients the amd's just can't keep up even some of the slower i5's wax amd here.

Again its all about what the user wants and needs...for a gaming build I use amd only cause i can save serious cash and put it toward something else...but interms of doing heavy work...intel is the way...but again i hope bulldozer levels the playing field maybe they have a cat in a hat hence the limited information..they don't want intel knowing their plans for world domination biggrin.gif
X
(15 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1700 @ 3.9Ghz Gigabyte Aorus X-370 R9 Fury Strix. Corsair Vengence LPX 2x8Gb 
Hard DriveCoolingCoolingCooling
Crucial MX100 XSPC Raystorm XSPC EX 360 US Solar TDC 
OSMonitorKeyboardPower
Dual Boot (Win 7, Kali Linux) LG Ultrawide 29UM67 Gskill KM780 XFX XTR 650 
CaseMouse
Cooler Master HAF XB Evo Razer Naga Hex 
  hide details  
Reply
X
(15 items)
 
  
CPUMotherboardGraphicsRAM
Ryzen 1700 @ 3.9Ghz Gigabyte Aorus X-370 R9 Fury Strix. Corsair Vengence LPX 2x8Gb 
Hard DriveCoolingCoolingCooling
Crucial MX100 XSPC Raystorm XSPC EX 360 US Solar TDC 
OSMonitorKeyboardPower
Dual Boot (Win 7, Kali Linux) LG Ultrawide 29UM67 Gskill KM780 XFX XTR 650 
CaseMouse
Cooler Master HAF XB Evo Razer Naga Hex 
  hide details  
Reply
post #23 of 354
It is impossible to speculate accurately at this time. We can know everything about how the architecture is designed but without actual numbers this is nothing more than speculation about how people want it to beat intel.
3930k
(20 items)
 
  
Reply
3930k
(20 items)
 
  
Reply
post #24 of 354
Quote:
Originally Posted by _GTech;12311779 
You have a valid point IF all you do is gaming, video encoding, archive compression etc, not much of a gain. (what 3-10 seconds shaved off?)

When you start compiling programs then you start to see a significant gain in real world applications...

Not everyone can utilize an Uber CPU, it takes tools & you should have the NEED for such a monsters CPU, like a server etc..

AMD made it very clear that, it's not about the GHz and clock speed for a Desktop Computer, but how well it can multi-task, this is why they went for more cores & essentially more processes over a faster speed, nevertheless, they did make improvements on speed as well as power, but more importantly they clearly stated that it would be competing with Intel in price too! <<

I look at Intel's decision to speed up Ive Bridge as a sign that they FEAR THE BULLDOZER, lol, but I could be wrong.. biggrin.gif


Good post, thanks..

Or they chose to speed it up because that's typically what companies try to do each cpu generation...rolleyes.gif
3930k
(20 items)
 
  
Reply
3930k
(20 items)
 
  
Reply
post #25 of 354
the best answer you will get is that AMD has benchmarked it as ~= to the 980X at stock, which is similar to the 2600K at stock.
    
CPUMotherboardGraphicsRAM
Intel Core i5 2500k @ 4.5GHZ (for now) MSI P67A-GD65 BFG GTX 260 OC SLI 2x2GB G.Skill Ripjaw X 2133 CL9 
Hard DriveOptical DriveOSMonitor
Barracuda 320GB, Spinpoint F3 1TB, Barracuda 1.5TB Sony Optiarc BD-ROM Windows 7 Ultimate x64 HP 2311x 
KeyboardPowerCaseMouse
Dell L100 In Win Commander 1200W RV02B-EW Trackman Wheel 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Core i5 2500k @ 4.5GHZ (for now) MSI P67A-GD65 BFG GTX 260 OC SLI 2x2GB G.Skill Ripjaw X 2133 CL9 
Hard DriveOptical DriveOSMonitor
Barracuda 320GB, Spinpoint F3 1TB, Barracuda 1.5TB Sony Optiarc BD-ROM Windows 7 Ultimate x64 HP 2311x 
KeyboardPowerCaseMouse
Dell L100 In Win Commander 1200W RV02B-EW Trackman Wheel 
  hide details  
Reply
post #26 of 354
Quote:
Originally Posted by amd-dude;12311916 
Yea i think and i honestly hope that bulldozer can level the playing field a bit cause i am tired of the intel boys (fan boys, not regular ones) saying 'oh well we beat you amd chip anytime'

I get what your saying about compiling and i have done some test for myself here as well..my chip just takes way to long to compile code in backtrack 4 or even if im using pyrat or cain and abel to crack hashes for demonstrations to clients the amd's just can't keep up even some of the slower i5's wax amd here.

biggrin.gif

Use the GPU within your built to speed up the hash cracking. It is relative easy to do and speeds up cracking significantly.
Gunslinger
(13 items)
 
  
CPUMotherboardGraphicsRAM
1090 MSI 890FXA GD70 5870 Corsair 
Hard DriveOSMonitorPower
C300 Windows I-INC Thermaltake 850 
Case
HAF 932 
  hide details  
Reply
Gunslinger
(13 items)
 
  
CPUMotherboardGraphicsRAM
1090 MSI 890FXA GD70 5870 Corsair 
Hard DriveOSMonitorPower
C300 Windows I-INC Thermaltake 850 
Case
HAF 932 
  hide details  
Reply
post #27 of 354
Trying to divine the future by forcing your head deep into the nether regions of your back side is not useful in the slightest. Until something official gets released, there is simply no way to know what's going to happen.

That said, here's to hoping AMD pulls off something spectacular. Competition is good for prices.
Jane
(11 items)
 
  
CPUMotherboardGraphicsRAM
i5 750 Zotac H55ITX-C-E Sapphire 7850 4GB Corsair XMS 1333Mhz 
Hard DriveCoolingOSMonitor
Samsung F3 1TB Antec 620 Windows 7 Samsung BX2331 23" LED 
PowerCaseAudio
Silverstone 45SF 450W Silverstone SG05 Swan M10 
  hide details  
Reply
Jane
(11 items)
 
  
CPUMotherboardGraphicsRAM
i5 750 Zotac H55ITX-C-E Sapphire 7850 4GB Corsair XMS 1333Mhz 
Hard DriveCoolingOSMonitor
Samsung F3 1TB Antec 620 Windows 7 Samsung BX2331 23" LED 
PowerCaseAudio
Silverstone 45SF 450W Silverstone SG05 Swan M10 
  hide details  
Reply
post #28 of 354
As it's been said you can't calculate 4GHz x 4 = 16GHz, let alone compare that "16GHz" to say 3GHz x 8 = "24GHz" because programs just don't linearly scale like that when you add threads, let alone applications like games that aren't well multi-threaded which cannot just divide the tasks to additional cores as you slap them on the system. Cache sizes and a plenty of other factors I don't know anything about matter a huge lot, and the architecture which we don't know much about makes all guesses until release and benchmarks futile.

AFAIK it'd be a lot easier to theoretically pit them against each other if we could benchmark one i5 2500K core against one Zambezi core with exactly the same amount of cache, then we'd have some sort of "per hertz" performance. Even then you couldn't calculate the performance differential between quad and an octo unless you know the program you're going to use linearly scales with additional threads, and even in a best-case scenario like this the real-life performance might be quite different due to for example, differences in the said cache size.
    
CPUMotherboardGraphicsRAM
Phenom II X6 1055T 3.3GHz 1.29V Asus M4A88TD-V EVO Gigabyte GTX 460 1GB OC 12GB Kingston HyperX blu DDR3 1600MHz CL9 
Hard DriveOptical DriveOSMonitor
OCZ Agility 3 60GB + Samsung Spinpoint F3 1TB some Samsung USB DVD-RW Windows 7 Professional 64-bit Acer P236H 
KeyboardPowerCaseMouse
Prodige Revolution, horrible.. Corsair VX 550W Fractal Design Define R3 Logitech Nano V550 
Mouse Pad
Sharp Extinctor 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Phenom II X6 1055T 3.3GHz 1.29V Asus M4A88TD-V EVO Gigabyte GTX 460 1GB OC 12GB Kingston HyperX blu DDR3 1600MHz CL9 
Hard DriveOptical DriveOSMonitor
OCZ Agility 3 60GB + Samsung Spinpoint F3 1TB some Samsung USB DVD-RW Windows 7 Professional 64-bit Acer P236H 
KeyboardPowerCaseMouse
Prodige Revolution, horrible.. Corsair VX 550W Fractal Design Define R3 Logitech Nano V550 
Mouse Pad
Sharp Extinctor 
  hide details  
Reply
post #29 of 354
This math cannot be done.

Performance is not additive and is barely relative inside of a given series.

For Example:
A Q9450 at 3.20ghz (8x400.0mhz) 45nm
A Q9650 at 3.20ghz (9x355.5mhz) 45nm

They are within the same series, but the performance of the two setups will be slightly off (Off of my head guess is that the Q9450 will come out ahead). Reason being is that the CPU speed is so inherently linked with other hardware performance. With this example the 44mhz FSB difference is enough to skew the ram speed by 88mhz or more. The motherboard running a higher FSB will also increase overall performance.

Just this one example is enough to say that 3.2ghz doesn't mean 3.2ghz; it means 3.2ghz on a given cpu with a given setup. Take into consideration that efficencies of cores are not equal even though they theoretically should be. 3.2ghz on Core 0 is not performing the same workload exactly the same way at 3.2ghz on Core 1. The only way to make the cores speed an additive (or multiplicative if you are feeling fancy) effect is for them to run 100% workload (not power) efficency and to run the exact same code across all cores.

This alone shows that a speed is not the same from core to core even in the same series let alone between architectures or manufacters.

Then to compare between architectures:

3.2ghz Core 2 Quad Q9450 (8x400mhz) 45nm
3.2ghz Core i7 960 (24x133) 45nm

The new architecture gives an average of ~12% faster than the Core 2 45nm Source, but realize again that this is an average and it's really a non-linear trend. The only way to find a true "clock-for-clock" gain is to use one benchmark under the exact same conditions with the exact same hardware (impossible) and with all other hardware performance being exactly the same (impossible...especially on the ram as the base clock vs. fsb is inherently different and the imc on the Core i7 is leaps and bounds beyond the best memory controller on an socket 775 mobo). You would get a plot of performance increase per clock speed that again would not be linear.


Then to take it across manufacter boundaries:
3.2ghz Intel =/= 3.2ghz AMD

Again, the clock speeds vs. performance is not linear as AMD has in the past had some scaling issues with higher clock speed.

To give a better example (and this is inherently arbitrary), let's say that at 3.2ghz both the AMD and Intel CPUs perform exactly the same, but you might find that at 4.0ghz the AMD CPU performs 10% under the Intel CPU...but then at 5.0ghz maybe the AMD CPU performs 10% over the Intel CPU....It's not linear.

There have always been an issue with people saying clock-for-clock it's this much faster...it's just hogwash but the real truth is so much more complicated and leaves no definitive answer that most people propogate the lie.

Bulldozer/Zambezi isn't out yet. There aren't real benchmarks yet. No comparrision can be made. Internal testing always makes the CPU look better than it is. (Intel is infamous for their internal testing). They run the test under the best possible conditions and with a cherry picked chip until it tells them what they want it to (Both companies do this).

For those tl;dr - No performance estimate can be made.
Gaming PC / HTPC
(17 items)
 
Bedroom PC
(13 items)
 
 
CPUMotherboardGraphicsRAM
i7 7700k Asus Prime Z270-A GB Auros-X 1080 Ti G-Skill (2x8gb) 3000mhz  
Hard DriveHard DriveHard DriveOptical Drive
Samsung 850 Western Digital 5900rpm Seagate 7200rpm LG 22x DVD Burner 
CoolingOSMonitorKeyboard
Noctua NH-D14 Win 10 x64 Pro LG 55" 4k B6 OLED Microsoft Desktop 400 
PowerCaseMouseMouse Pad
Corsair TX750 V2 NZXT Source 210 Microsoft Desktop 400 Black 
Audio
Custom Surround 5.1 
CPUMotherboardGraphicsRAM
Intel 2120T Biostar H61MH Intel HD 2000 G-Skill F3-12800CL9D-8GBXL  
Hard DriveOptical DriveCoolingOS
Samsung Evo 128gb Asus 24x DVD Burner Stock Win 10 x64 
MonitorKeyboardPowerCase
LG 40" 1080p HP Wireless Keyboard/Mouse Silverstone 400W ST40F-ES Silverstone ML03B 
Other
Linksys WMP600N Wireless Card 
  hide details  
Reply
Gaming PC / HTPC
(17 items)
 
Bedroom PC
(13 items)
 
 
CPUMotherboardGraphicsRAM
i7 7700k Asus Prime Z270-A GB Auros-X 1080 Ti G-Skill (2x8gb) 3000mhz  
Hard DriveHard DriveHard DriveOptical Drive
Samsung 850 Western Digital 5900rpm Seagate 7200rpm LG 22x DVD Burner 
CoolingOSMonitorKeyboard
Noctua NH-D14 Win 10 x64 Pro LG 55" 4k B6 OLED Microsoft Desktop 400 
PowerCaseMouseMouse Pad
Corsair TX750 V2 NZXT Source 210 Microsoft Desktop 400 Black 
Audio
Custom Surround 5.1 
CPUMotherboardGraphicsRAM
Intel 2120T Biostar H61MH Intel HD 2000 G-Skill F3-12800CL9D-8GBXL  
Hard DriveOptical DriveCoolingOS
Samsung Evo 128gb Asus 24x DVD Burner Stock Win 10 x64 
MonitorKeyboardPowerCase
LG 40" 1080p HP Wireless Keyboard/Mouse Silverstone 400W ST40F-ES Silverstone ML03B 
Other
Linksys WMP600N Wireless Card 
  hide details  
Reply
post #30 of 354
All I was trying to say is exactly what has been said in the post against mine. They're comparing Bulldozer to a Nehalem processor clocked at a measly 3.06ghz. Of course newer tech will prevail.

In essence I was trying to point out that Bulldozer needs to be compared not to tech from 2008 but tech from 2011 (Sandy Bridge) in which case I would be much more amazed that their 8 core can beat a Sandy quad with HT which is not even equated to 4 full cores, but rather something like half a core each.

If Bulldozer > Sandy Bridge then I will retract my previous statement.
Edited by blackbalt89 - 2/8/11 at 2:38pm
Black Sand
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k @ 4.8GHz ASUS Maximus IV Extreme-Z EVGA GTX 580 SLI 8GB Corsair Dominator GT @ 1885 MHz 
Hard DriveOSMonitorPower
60GB Mushkin Callisto Deluxe +1TB WD Caviar Black Windows 7 Home Premium ASUS VW246 Corsair HX-850 
Case
HAF 932 
  hide details  
Reply
Black Sand
(13 items)
 
  
CPUMotherboardGraphicsRAM
i7 2600k @ 4.8GHz ASUS Maximus IV Extreme-Z EVGA GTX 580 SLI 8GB Corsair Dominator GT @ 1885 MHz 
Hard DriveOSMonitorPower
60GB Mushkin Callisto Deluxe +1TB WD Caviar Black Windows 7 Home Premium ASUS VW246 Corsair HX-850 
Case
HAF 932 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD CPUs
Overclock.net › Forums › AMD › AMD CPUs › The Zambezi CPU Performance Estimate...