Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [WCCF] AMD Trinity A10-5800K APU Tested
New Posts  All Forums:Forum Nav:

[WCCF] AMD Trinity A10-5800K APU Tested - Page 5

post #41 of 75
Quote:
Originally Posted by sergionography View Post

the pictures posted say 2000ghz core frequency, isnt it supposed 2 be 3800?
http://cdn.overclock.net/1/13/600x418px-LL-135eb6c3_7906214953-orig.jpeg
if that is the case it would make trinity cores almost 40% faster than llano core for core, and thats just impossible -_-
im telling u man this is very fishy

Dude, it's CPUz during SuperPi. SuperPi doesn't show you full frequency; rarely while running SuperPi. Cool n Quiet and additional power saving modes are on.
Edited by Dmac73 - 3/21/12 at 12:29am
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
post #42 of 75
Quote:
Originally Posted by MadGoat View Post

I'm assuming this is one of the first glances into piledriver core performance?
If so, this is sad... IF it is indeed turbo clocking to 4.2 for super PI ... its even more horrid.
I can get under 17s 1m with 4ghz with my thuban...
Giving the Piledriver core the benifit of the doubt here, lets assume its running 3.8ghz for 23s 1m... at that speed it would need another 1.5 ghz (5.3ghz) to at least match my old slow PH2...
I'm really hoping (for AMD's sake) this is bogus, teething ES, or not a fully functional piledriver core at all...

Performance, clock for clock, Piledriver will be signifantly better than Trinity and not just because of L3 cache. Yes they have Piledriver based cores, there's several factors decreasing performance. A bigger difference between Llano and Phenom 2.



The scores seem right on the money, anyone freaking out thinking otherwise should slowdown and reconsider.

We can also assume turbo core isn't working. Let's say this is at 3.8.
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
post #43 of 75
So the overall score for the high end model is 9.3k in 3dmark06? I hope these are fake, because that's not that impressive. My 3 year old gatewayfx laptop gets that with a dual core clocked at 2.3

I am hoping these are a little more beefy, pulling 12k+ in 3dmark06
QUAD-METAL
(14 items)
 
  
CPUMotherboardGraphicsRAM
i5 2500k @ 4.2 GHZ Gigabyte z68x UD3P 2x Sapphire 7870 CrossfireX 16GB kingston ddr3 1600 
Hard DriveHard DriveOptical DriveOS
Crucial m4 64gb Intel 730 Series SSD Samsung DVD Burner Windows 7 64bit 
MonitorPowerCaseMouse
Asus 24" 1080p Rosewill 1000w PSU CM 690 Logitech G500 
  hide details  
Reply
QUAD-METAL
(14 items)
 
  
CPUMotherboardGraphicsRAM
i5 2500k @ 4.2 GHZ Gigabyte z68x UD3P 2x Sapphire 7870 CrossfireX 16GB kingston ddr3 1600 
Hard DriveHard DriveOptical DriveOS
Crucial m4 64gb Intel 730 Series SSD Samsung DVD Burner Windows 7 64bit 
MonitorPowerCaseMouse
Asus 24" 1080p Rosewill 1000w PSU CM 690 Logitech G500 
  hide details  
Reply
post #44 of 75

I am extremely doubtful of this claim that the CPU-only desktop Piledriver will be better and more powerful  than Trinity APU.

 

Piledriver will not have the integrated GPU and the ability to take advantage of it as an FPU through HSA (Heterogenuous System Architecture) which can bring in itself a huge performance boost.

 

Since HSA is only early on into the development process and the majority of apps that matter still make use of integer and not FPU performance, not all programs can take advantage of it.

 

This is why AMD is still releasing the lots-of-cores CPUs for desktops, servers and the high end segment.  But I don't doubt that one day we will see no more CPUs from AMD and just APUs.  They are so flexible.  You can use the GPU portion as either the FPU in assist with the CPU or of course as the GPU itself.  As AMD works on improvements including Resonant Clock Mesh and soon its own version of 3D transistors (although RSM is a pretty nice competitior to) along with power consumption improvements and die shrinks, we might soon be able to see APUs that can fit the 8 cores you'd normally expect in a desktop variant and a midrange GPU at acceptable power consumption levels, eradicating the need to even release CPUs.

 

So you might see the desktop-based Piledriver CPU being a better solution but as HSA is growing and so is the APU market, I just can't see things going to become the same way in the future.

 

If you read most of the news articles on OCN you will have seen the recent news article documenting the recent results with AMD's FPUs.  The HD 7970's FPU might be outperforming the GTX 680's by 2x at the same power consumption and even the 7870 is beating the GTX 680 in FPU performance.  I'm pretty sure this reach extends to lower-end cards and iGPU units.  AMD will have a huge advantage if and when HSA takes off.

post #45 of 75
Quote:
Originally Posted by Dmac73 View Post

Performance, clock for clock, Piledriver will be signifantly better than Trinity and not just because of L3 cache. Yes they have Piledriver based cores, there's several factors decreasing performance. A bigger difference between Llano and Phenom 2.
The scores seem right on the money, anyone freaking out thinking otherwise should slowdown and reconsider.
We can also assume turbo core isn't working. Let's say this is at 3.8.


http://www.phoronix.com/scan.php?page=article&item=amd_trinity_early&num=1
i came across this article, you might find it interesting, the engineering sample mentioned here has 2.5ghz speed, and radeon devastator graphics
i wonder if it could be the same one
Quote:
Originally Posted by xd_1771 View Post

I am extremely doubtful of this claim that the CPU-only desktop Piledriver will be better and more powerful  than Trinity APU.

Piledriver will not have the integrated GPU and the ability to take advantage of it as an FPU through HSA (Heterogenuous System Architecture) which can bring in itself a huge performance boost.

Since HSA is only early on into the development process and the majority of apps that matter still make use of integer and not FPU performance, not all programs can take advantage of it.

This is why AMD is still releasing the lots-of-cores CPUs for desktops, servers and the high end segment.  But I don't doubt that one day we will see no more CPUs from AMD and just APUs.  They are so flexible.  You can use the GPU portion as either the FPU in assist with the CPU or of course as the GPU itself.  As AMD works on improvements including Resonant Clock Mesh and soon its own version of 3D transistors (although RSM is a pretty nice competitior to) along with power consumption improvements and die shrinks, we might soon be able to see APUs that can fit the 8 cores you'd normally expect in a desktop variant and a midrange GPU at acceptable power consumption levels, eradicating the need to even release CPUs.

So you might see the desktop-based Piledriver CPU being a better solution but as HSA is growing and so is the APU market, I just can't see things going to become the same way in the future.

If you read most of the news articles on OCN you will have seen the recent news article documenting the recent results with AMD's FPUs.  The HD 7970's FPU might be outperforming the GTX 680's by 2x at the same power consumption and even the 7870 is beating the GTX 680 in FPU performance.  I'm pretty sure this reach extends to lower-end cards and iGPU units.  AMD will have a huge advantage if and when HSA takes off.

true HSA isnt until 2014 so dont expect too much now, at this point amd is pretty much pushing for gpu compute, something u can do on discrete gpus too
and yes piledriver on desktop will be better than apu, amd revlealed earlier that they will be adapting a tick tock strategy kinda like intel, however its a big different to were they release a core(piledriver) and even tho it will be out for production they will keep on working on it and enhancing it until its time to start baking piledriver desktop cpus, then by that time it would be a bit better(i wont expect too much tho but better than nothing)
and for now the cpu will still do with the 128bit fpu seen on bulldozer(or if enhanced and made bigger it will remain on the cpu for now)
post #46 of 75
Quote:
Originally Posted by xd_1771 View Post

You know what I'm willing to bet would be the problem here if it was real?

RAM.

APUs make use of system memory and are exceptionally bandwidth-intensive.  Anything less than DDR3-1600 starts to bottleneck even Llano A6-A8 APUs and I predict that the Trinity APU will require at least DDR3-1866-2000 to run without any bottlenecks on the top-end A10.  This may be why the entire APU is not seeing nearly as high performance here as it should be.  RAM is not specified in these tests.

If the tester dared try run these tests on generic DDR3-1333 memory than performance could be half of what it would be on a proper 1866+ kit.  Llano performance can be as much as (*depending on exact task) 10-20% less on DDR3-1333 and below kits, and 20% less & 40-50% less on a DDR3-1333 & DDR3-1066 kit respectively (both at CL7) vs a DDR3-1600 kit in gaming tasks.

Llano has between 2% less and 7% more clock-for-clock per-core performance than Phenom II depending on the test (cache sizes mean there are differences).

Cool'n'Quiet is probably bringing core frequency back down to 2Ghz after the testing, by the way.

This is true. With my A8 I gained a full 1 point in cinebench cpu test just by upgrading more ram with same speed and faster timings. First I used 4gb of corsair vengeance 1866mhz at 9-10-9-27 then went to 8gb gskill 1866mhz at 8-9-9-24 and the cinebench score increases 1.04 points. Also I get about a 10-15 fps extra in my games on average with the gskill ram.

Ram is defintely a major limiting factor for AMDs APUs. This should defintely be taken into consideration when testing these chips. Going from 4gb of 1333Mhz ram to 8gb of 1866 or 2000Mhz ram could be a much higher performance level for the APU. And since ram is so cheap nowadays you can easily provide you apu with hardware it needs to perform to its potential.
FUSION
(13 items)
 
  
CPUMotherboardRAMHard Drive
AMD Llano A8-3870k ASUS F1A75-M Pro FM1 uATX Corsair Vengeance Corsair Force 3 
Hard DriveOptical DriveCoolingOS
Hitachi ASUS DVD Burner Cooler Master Hyper TX3 Windows 7 Ultimate 64-bit 
MonitorKeyboardPowerCase
ASUS VE247H 24" 1080p Razer Lycosa Corsair CX500 Fractal Define Design R3 ATX 
Mouse
Razer DeathAdder 
  hide details  
Reply
FUSION
(13 items)
 
  
CPUMotherboardRAMHard Drive
AMD Llano A8-3870k ASUS F1A75-M Pro FM1 uATX Corsair Vengeance Corsair Force 3 
Hard DriveOptical DriveCoolingOS
Hitachi ASUS DVD Burner Cooler Master Hyper TX3 Windows 7 Ultimate 64-bit 
MonitorKeyboardPowerCase
ASUS VE247H 24" 1080p Razer Lycosa Corsair CX500 Fractal Define Design R3 ATX 
Mouse
Razer DeathAdder 
  hide details  
Reply
post #47 of 75
Quote:
Originally Posted by sergionography View Post

this looks very fake, 12% increase in overall performance with clocks that are 25% higher at the same tdp, thats already what bulldozer does, piledriver is better and enhanced. and ive read much more legit rumors saying trinity will perform 30% better in cpu tasks


12% increase with 25% higher clocks.......... how is that surprising? It already started with a 10% ipc deficit.

I own a bunch of amd k10 rigs, but I have to say, you can't really polish a turd.

You expected amd to perform a miracle and squeeze 30% better performance out of the crap bulldozer?
How? by 30% better ipc, or 30% higher clocks ?
ipc is very hard to fix, and reasonably, would only increase by single digit percentage.

these benchmarks are in line with what most people expect.

remember the FUD surrounding bulldozer prelaunch? and how lots of people cried fake?
actually turned out to be a turd, right ?
post #48 of 75
Quote:
Originally Posted by mikezachlowe2004 View Post

This is true. With my A8 I gained a full 1 point in cinebench cpu test just by upgrading more ram with same speed and faster timings. First I used 4gb of corsair vengeance 1866mhz at 9-10-9-27 then went to 8gb gskill 1866mhz at 8-9-9-24 and the cinebench score increases 1.04 points. Also I get about a 10-15 fps extra in my games on average with the gskill ram.

Your full of it
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
Heat
(14 items)
 
Asus G73JH
(9 items)
 
 
CPUMotherboardGraphicsGraphics
i7 3770K @ 4600mhz Gigabyte GA-Z77X-UD5H eVGA GTX 980 eVGA GTX 980 
RAMHard DriveOptical DriveCooling
Team 8gb @ 2400mhz  SSDs LiteOn H70Core/ Ultra Kaze PP 
OSPowerCase
W7 Ultimate X64 PCP&C MKII 950W CM Storm Scout 
CPUMotherboardGraphicsRAM
Intel i7 720 Asus HD5870M 6gb Triple Channel 
Hard DriveCoolingOSMonitor
500gb Asus Dual Fans W7 17.3" 
Mouse
Razer Orochi 
  hide details  
Reply
post #49 of 75
Quote:
Originally Posted by Dmac73 View Post

Your full of it

i agree he is, has no proof to back up the claim ethier... so until proof is given he's full of it :v
JunkoXan's Build
(15 items)
 
   
CPUMotherboardGraphicsRAM
Intel 2700k (Engineering Sample) Asus P67 Sabertooth Sapphire Dual-X 280x G.Skill Sniper (2x4gb) 
Hard DriveCoolingOSMonitor
Samsung F3 Spinpoint Cooler Masters 616+ Windows 7 Upstar 20" 1080p TV 
MonitorKeyboardPowerCase
Dell Monitor 21" 1680x1050 Dell Keyboard Antec High Current Gamer 900W Cooler Master HAF-XB 
MouseMouse Pad
E-3lue Mazer II 2600DPI Optical Mouse Custom Made 
  hide details  
Reply
JunkoXan's Build
(15 items)
 
   
CPUMotherboardGraphicsRAM
Intel 2700k (Engineering Sample) Asus P67 Sabertooth Sapphire Dual-X 280x G.Skill Sniper (2x4gb) 
Hard DriveCoolingOSMonitor
Samsung F3 Spinpoint Cooler Masters 616+ Windows 7 Upstar 20" 1080p TV 
MonitorKeyboardPowerCase
Dell Monitor 21" 1680x1050 Dell Keyboard Antec High Current Gamer 900W Cooler Master HAF-XB 
MouseMouse Pad
E-3lue Mazer II 2600DPI Optical Mouse Custom Made 
  hide details  
Reply
post #50 of 75
I would like to see this tested with more RAM. According to this: http://www.guru3d.com/article/amd-fx-8150--8120-6100-and-4100-performance-review/6 the FX-4100 gets a 4102 CPU score in 3dmark06. I really want to see Piledriver come through on this one, a 5% boost isn't going to cut it at all unless the power consumption goes down drastically.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Rumors and Unconfirmed Articles
Overclock.net › Forums › Industry News › Rumors and Unconfirmed Articles › [WCCF] AMD Trinity A10-5800K APU Tested