Overclock.net › Forums › Industry News › Video Game News › [Anand] Fable Legends DX12 Benchmark Analysis
New Posts  All Forums:Forum Nav:

[Anand] Fable Legends DX12 Benchmark Analysis - Page 43  

post #421 of 443
Clock speeds outside of factory overclocked to me don't mean much. What i mean is even though people say clock speed xxxx is stable chances are its not entirely stable. It's like all the cpu threads of overclocking, people are like look at me i run Xghz im top of the chart. Truth is stability is in question regardless of claims. There is not really a way to test stability 100% for end users. That's not counting good ol silicon lottery and all that jazz.
People tend to insist anyone can reach X clock speed on X gpu/cpu, but everyone has different usages and what may work for CSGO doesn't work properly in BF4. Overclocking is a crap shoot once in the hands of a user, too many variables interfere.

Clocks set from the factory are way more trustworthy as i am sure they have way better ways to test stability at the hardware level. Just my two cents.
post #422 of 443
Quote:
Originally Posted by dagget3450 View Post

Clock speeds outside of factory overclocked to me don't mean much. What i mean is even though people say clock speed xxxx is stable chances are its not entirely stable. It's like all the cpu threads of overclocking, people are like look at me i run Xghz im top of the chart. Truth is stability is in question regardless of claims. There is not really a way to test stability 100% for end users. That's not counting good ol silicon lottery and all that jazz.
People tend to insist anyone can reach X clock speed on X gpu/cpu, but everyone has different usages and what may work for CSGO doesn't work properly in BF4. Overclocking is a crap shoot once in the hands of a user, too many variables interfere.

Clocks set from the factory are way more trustworthy as i am sure they have way better ways to test stability at the hardware level. Just my two cents.

The CPU you can get pretty much 100% stable. GPU is a different story since temp/load/drivers effect OC. You can find a MAX OC of something like 1500MHz for GTX980 Ti and i am sure you could run 1450MHz 100% stable.
Ishimura
(21 items)
 
Silent Knight
(13 items)
 
 
CPUMotherboardGraphicsRAM
Intel Core i7 3770K @ 4.6GHz ASRock Z77E-ITX eVGA GTX 1080 Ti Hybrid AVEXIR Blitz 1.1 16GB DDR3-2400MHz CL10 
Hard DriveHard DriveCoolingCooling
SanDisk Ultra II 960GB Toshiba X300 5TB Corsair H100i GTX eVGA Hybrid Water Cooler  
CoolingOSMonitorKeyboard
4x GentleTyphoon AP-15 Windows 10 Pro 64-Bit Philips Brilliance BDM4065UC 4K Razer BlackWidow Chroma  
PowerCaseMouseMouse Pad
eVGA SuperNOVA 750 G3 Define Nano S Logitech G502 Proteus Core PECHAM Gaming Mouse Pad XX-Large 
AudioAudioAudioAudio
Audioengine D1 DAC Mackie CR Series CR3 Audio-Technica ATH-M50 Sennheiser HD 598 
Audio
Sony XB950BT 
CPUMotherboardGraphicsRAM
AMD Phenom II X4 955 @ 4.2GHz ASUS M4A79XTD EVO AMD Radeon HD 7970 3GB @ 1200/1500 2x 4GB G.SKILL Ripjaws X DDR3-1600 
Hard DriveHard DriveHard DriveCooling
OCZ Agility 3 60GB WD Caviar Green 1.5TB 2 x Seagate Barracuda 2TB XSPC Raystorm 
CoolingCoolingOSPower
EK-FC7970 XSPC RS360 Windows 10 Pro 64-Bit Corsair TX750 
Case
NZXT Switch 810  
  hide details  
Ishimura
(21 items)
 
Silent Knight
(13 items)
 
 
CPUMotherboardGraphicsRAM
Intel Core i7 3770K @ 4.6GHz ASRock Z77E-ITX eVGA GTX 1080 Ti Hybrid AVEXIR Blitz 1.1 16GB DDR3-2400MHz CL10 
Hard DriveHard DriveCoolingCooling
SanDisk Ultra II 960GB Toshiba X300 5TB Corsair H100i GTX eVGA Hybrid Water Cooler  
CoolingOSMonitorKeyboard
4x GentleTyphoon AP-15 Windows 10 Pro 64-Bit Philips Brilliance BDM4065UC 4K Razer BlackWidow Chroma  
PowerCaseMouseMouse Pad
eVGA SuperNOVA 750 G3 Define Nano S Logitech G502 Proteus Core PECHAM Gaming Mouse Pad XX-Large 
AudioAudioAudioAudio
Audioengine D1 DAC Mackie CR Series CR3 Audio-Technica ATH-M50 Sennheiser HD 598 
Audio
Sony XB950BT 
CPUMotherboardGraphicsRAM
AMD Phenom II X4 955 @ 4.2GHz ASUS M4A79XTD EVO AMD Radeon HD 7970 3GB @ 1200/1500 2x 4GB G.SKILL Ripjaws X DDR3-1600 
Hard DriveHard DriveHard DriveCooling
OCZ Agility 3 60GB WD Caviar Green 1.5TB 2 x Seagate Barracuda 2TB XSPC Raystorm 
CoolingCoolingOSPower
EK-FC7970 XSPC RS360 Windows 10 Pro 64-Bit Corsair TX750 
Case
NZXT Switch 810  
  hide details  
post #423 of 443
Quote:
Originally Posted by prjindigo View Post

Someone at another site pointed out that Anand used a 1.25GHz 980ti against a stock clock Fury X. Such a claim seems consistent with the massive spanking we see the 390x give the stock 980.

That's really hard to believe.

For one, I would hope that AT would have stated as such.

Two, AT has personally tested only one 980 Ti to this point which was the stock version. It boosted between 1101-1189 mhz depending on the game. So they haven't even had their hands on a 980 Ti that would boost that high out of the box.
Upstairs Rig
(11 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus VI Hero evga 1080 Ti with Hybrid mod Corsair Vengeance Pro 2133 mhz 
Hard DriveHard DriveCoolingOS
Samsung 850 EVO 500gb WD Caviar Black Corsair h100i GTX Windows 8.1 64bit 
MonitorPowerCase
xb280hk EVGA Supernova 1000 G2 Corsair Carbide Air 540 
  hide details  
Upstairs Rig
(11 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus VI Hero evga 1080 Ti with Hybrid mod Corsair Vengeance Pro 2133 mhz 
Hard DriveHard DriveCoolingOS
Samsung 850 EVO 500gb WD Caviar Black Corsair h100i GTX Windows 8.1 64bit 
MonitorPowerCase
xb280hk EVGA Supernova 1000 G2 Corsair Carbide Air 540 
  hide details  
post #424 of 443
Quote:
Originally Posted by raghu78 View Post

Its the same over and over again. Any performance benchmark which paints Nvidia in a slightly unfavourable light is useless as is asynchronous compute. These are the same people who bought a GK104 in 2013 over a HD 7970/R9 280X. We know starting from 2014 how that turned out. Anyway as usual GCN owners will see their cards age very well. Right now the R9 290 / R9 390 / GTX 980 Ti are great buys. I will recommend nothing else. btw there are many who still feel every R9 290 will perform like a R9 390 which is not entirely true. The memory controller on R9 390 is tweaked to consistently run GDDR5 at 6 Ghz stock and can overclock to 6.8 Ghz speeds. R9 290 memory overclocks are a hit and miss depending on the particular AIB model along with the memory brand. The extra bandwidth is making a significant difference between R9 290 and R9 390 perf.

"The extra bandwidth is making a significant difference between R9 290 and R9 390 perf".

1-3% is "significant" performance difference? We already have tons of data here alone comparing 290/290x to 390/390x at same clock speeds. The conclusion is the difference is negligible.
My Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7-4770k @ 4.1Ghz  Asus z87-Plus XFX Double D r9 290X G Skill Sniper 8GB @ 2133 
OSMonitorKeyboardMouse
Windows 10 Pro 27 Inch 1440P PLS @ 110Hz Mechanical Keyboard MX Red Steelseries Kana 
Mouse PadAudio
Steelseries Mouse pad Samson SR850 Headphones/Asus Xonar Sound card 
  hide details  
My Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7-4770k @ 4.1Ghz  Asus z87-Plus XFX Double D r9 290X G Skill Sniper 8GB @ 2133 
OSMonitorKeyboardMouse
Windows 10 Pro 27 Inch 1440P PLS @ 110Hz Mechanical Keyboard MX Red Steelseries Kana 
Mouse PadAudio
Steelseries Mouse pad Samson SR850 Headphones/Asus Xonar Sound card 
  hide details  
post #425 of 443
Quote:
Originally Posted by mcg75 View Post

That's really hard to believe.

For one, I would hope that AT would have stated as such.

Two, AT has personally tested only one 980 Ti to this point which was the stock version. It boosted between 1101-1189 mhz depending on the game. So they haven't even had their hands on a 980 Ti that would boost that high out of the box.

What Anandtech did was that they did not test with the latest AMD driver whereas they tested with an older nVIDIA driver (355.82).
Anandtech: Warning: Spoiler! (Click to show)
Source.

What ExtremeTech did was use the latest drivers from both manufacturers (355.98 for nVIDIA).
ExtremeTech: Warning: Spoiler! (Click to show)
Source

What PCper did was use the older nVIDIA driver (355.82) as well as the latest AMD driver.
PCper: Warning: Spoiler! (Click to show)
Source

What PCGamesHardware did was use the latest nVIDIA driver (355.98) as well as the latest AMD driver but with factory overclocked cards.
PcGamesHardware: Warning: Spoiler! (Click to show)
Source



Another difference is in the CPUs used.
Anandtech used an overclocked Core i7-4960X @ 4.2GHz
Extremetech used a Core i7-5960X @ 3GHz
PCper used a Skylake Core i7-6700K @ 4GHz
PCGamesHardware used a Skylake Core i7-6700K @ 4.5GHz

With the older AMD driver, GCN Fiji cards actually lose FPS as the CPU performance increases as seen here: Warning: Spoiler! (Click to show)
Thus the Anandtech numbers suffered from an AMD driver bug. We can ignore these results as they're bugged.

It also appears that nVIDIA requires a high clock rate on the first core of a CPU. Likely in order to allow for driver shader swaps (Shader compiler/Software side Scheduling). This is why the GTX 980 Ti performed poorly in the ExtremeTech tests using a Core i7-5960X @ Stock or 3GHz.

The Skylake CPUs offer the best IPC, therefore the nVIDIA GPUs really like the Skylake platform due to the Software Scheduling aspect of the nVIDIA Maxwell 2 architecture.

This is all in keeping with my original theory on the matter. You can see the extra CPU overhead of the Maxwell 2 software side scheduling on the Corei3 results here: Warning: Spoiler! (Click to show)


Therefore the conclusion is that nVIDIAs Maxwell 2 requires a CPU of a certain clockspeed in order to attain its full potential. This may not be a problem at present but could be if games become very demanding on the CPU. Some DX12 titles may exhibit this issue if they stress the CPU to a high degree due to the Multi-Threading nature of DX12. A simple overclock should remedy this issue. On the AMD front, the issue with Fiji appears to have been rectified by a newer driver which Anandtech omitted to use. Therefore I think this nullifies Anandtech's figures from consideration as the driver bug has been fixed, every other site used the fix except for Anandtech.
Edited by Mahigan - 9/26/15 at 9:53pm
Kn0wledge
(20 items)
 
Pati3nce
(14 items)
 
Wisd0m
(10 items)
 
Kn0wledge
(20 items)
 
Pati3nce
(14 items)
 
Wisd0m
(10 items)
 
post #426 of 443
Quote:
Originally Posted by TopicClocker View Post

Why do you say that? There are no DX12 games at the moment, so the CPU overhead for AMD GPUs in DX11 games are still prominent in the majority of games of today.

The 390 and 390X have 8GB of VRAM, and the 390X comes close to the 980 in performance at times, but if you factor in the overclocking capabilities of the 980 it easily wins against the 390X, and sometimes the Fury.

The 4GB VRAM isn't a problem at the moment for the 980 or the Fury, but we're unsure of what future games behold.
Quote:
Originally Posted by ZealotKi11er View Post

80% of people thing otherwise. GM204 crushed anything AMD had to offer in sales. People don't care about performance/$.
Quote:
Originally Posted by Mahigan View Post

What Anandtech did was that they did not test with the latest AMD driver whereas they tested with an older nVIDIA driver which offers nVIDIA better performance (355.82).
Anandtech: Warning: Spoiler! (Click to show)
Source.

What ExtremeTech did was use the latest drivers from both manufacturers (355.98 for nVIDIA).
ExtremeTech: Warning: Spoiler! (Click to show)
Source

What PCper did was use the older nVIDIA driver (355.82) as well as the latest AMD driver. this is all fun to read. The truth is it dosent matter to the gpu buying consumers. they wiil wait for pascal and buy nvidia.
PCper: Warning: Spoiler! (Click to show)
Source

What PCGamesHardware did was use the latest nVIDIA driver (355.98) as well as the latest AMD driver but with factory overclocked cards.
PcGamesHardware: Warning: Spoiler! (Click to show)
Source



Another difference is in the CPUs used.
Anandtech used an overclocked Core i7-4960X @ 4.2GHz
Extremetech used a Core i7-5960X @ 3GHz
PCper used a Skylake Core i7-6700K @ 4GHz
The thing is this. people will wait for pascal and buy it. Nvidia has mindshare of the market.AMD has ruined their name with terrible marketing and value branding.People recognise the brands and will choose who they believe is better. its no different from the athlon 64 days. people see the brand they know and buy.
post #427 of 443
Gotta love all this DX12 bs... there's always some reason why x is worse than y and why y is better than x.

CPU this, driver that, oc'd vs stock... nothing has been fair so far apparently.

This whole thing has been shameful up to this point. Just give us the damn benchmark so we can do our own testing.
My main PC
(8 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 6700k Asus ROG Maximus VIII Gene Nvidia GTX 1080Ti G.Skill Ripjaws 
Hard DriveOSKeyboardPower
Samsung 850 EVO  Windows 10 Razer Blackwidow Chroma EVGA Supernova 1300w 
  hide details  
My main PC
(8 items)
 
  
CPUMotherboardGraphicsRAM
Intel i7 6700k Asus ROG Maximus VIII Gene Nvidia GTX 1080Ti G.Skill Ripjaws 
Hard DriveOSKeyboardPower
Samsung 850 EVO  Windows 10 Razer Blackwidow Chroma EVGA Supernova 1300w 
  hide details  
post #428 of 443
Quote:
Originally Posted by Mahigan View Post

What Anandtech did was that they did not test with the latest AMD driver whereas they tested with an older nVIDIA driver (355.82).
Anandtech: Warning: Spoiler! (Click to show)
Source.

What ExtremeTech did was use the latest drivers from both manufacturers (355.98 for nVIDIA).
ExtremeTech: Warning: Spoiler! (Click to show)
Source

What PCper did was use the older nVIDIA driver (355.82) as well as the latest AMD driver.
PCper: Warning: Spoiler! (Click to show)
Source

What PCGamesHardware did was use the latest nVIDIA driver (355.98) as well as the latest AMD driver but with factory overclocked cards.
PcGamesHardware: Warning: Spoiler! (Click to show)
Source



Another difference is in the CPUs used.
Anandtech used an overclocked Core i7-4960X @ 4.2GHz
Extremetech used a Core i7-5960X @ 3GHz
PCper used a Skylake Core i7-6700K @ 4GHz
PCGamesHardware used a Skylake Core i7-6700K @ 4.5GHz

With the older AMD driver, GCN Fiji cards actually lose FPS as the CPU performance increases as seen here: Warning: Spoiler! (Click to show)
Thus the Anandtech numbers suffered from an AMD driver bug. We can ignore these results as they're bugged.

It also appears that nVIDIA requires a high clock rate on the first core of a CPU. Likely in order to allow for driver shader swaps (Shader compiler/Software side Scheduling). This is why the GTX 980 Ti performed poorly in the ExtremeTech tests using a Core i7-5960X @ Stock or 3GHz.

The Skylake CPUs offer the best IPC, therefore the nVIDIA GPUs really like the Skylake platform due to the Software Scheduling aspect of the nVIDIA Maxwell 2 architecture.

This is all in keeping with my original theory on the matter. You can see the extra CPU overhead of the Maxwell 2 software side scheduling on the Corei3 results here: Warning: Spoiler! (Click to show)


Therefore the conclusion is that nVIDIAs Maxwell 2 requires a CPU of a certain clockspeed in order to attain its full potential. This may not be a problem at present but could be if games become very demanding on the CPU. Some DX12 titles may exhibit this issue if they stress the CPU to a high degree due to the Multi-Threading nature of DX12. A simple overclock should remedy this issue. On the AMD front, the issue with Fiji appears to have been rectified by a newer driver which Anandtech omitted to use. Therefore I think this nullifies Anandtech's figures from consideration as the driver bug has been fixed, every other site used the fix except for Anandtech.
So Maxwell 2.0 is already doing software sheduling even without a proper driver, i don't think another driver will help nvidia in that mater .
Number 3
(16 items)
 
Dust Catcher
(9 items)
 
CPUMotherboardGraphicsRAM
i7 6700k 4.5 ghz MSI Gaming M7  1080ti aorus xtreme 2x8 go Gskill 3200mhz C15 
Hard DriveHard DriveHard DriveHard Drive
crucial ssd 128 Go 2x Samsung 840 evo 250 Go Hitachi HDD 2To Toshiba HDD 2To 
Optical DriveCoolingOSMonitor
Old DVD-RW KRAKEN X61 Windows 10 x64 Swift 2k 165hz 
KeyboardPowerCaseMouse
g910 CM V1000 CM Trooper modded g502 proteus 
CPUMotherboardGraphicsRAM
Phenom ii x6 1075t Asus Sabertooth 990fx CF 2x Asus 6950 directcu ii  2x4Gb Gskill 1333 
Hard DriveCoolingOSPower
80go Intel x25 ssd series + 500 go Samsung  cooler master V6 gt Win 7 64bit ultimate edition cooler master silent pro 850w 
Case
HAF 912 advanced 
  hide details  
Number 3
(16 items)
 
Dust Catcher
(9 items)
 
CPUMotherboardGraphicsRAM
i7 6700k 4.5 ghz MSI Gaming M7  1080ti aorus xtreme 2x8 go Gskill 3200mhz C15 
Hard DriveHard DriveHard DriveHard Drive
crucial ssd 128 Go 2x Samsung 840 evo 250 Go Hitachi HDD 2To Toshiba HDD 2To 
Optical DriveCoolingOSMonitor
Old DVD-RW KRAKEN X61 Windows 10 x64 Swift 2k 165hz 
KeyboardPowerCaseMouse
g910 CM V1000 CM Trooper modded g502 proteus 
CPUMotherboardGraphicsRAM
Phenom ii x6 1075t Asus Sabertooth 990fx CF 2x Asus 6950 directcu ii  2x4Gb Gskill 1333 
Hard DriveCoolingOSPower
80go Intel x25 ssd series + 500 go Samsung  cooler master V6 gt Win 7 64bit ultimate edition cooler master silent pro 850w 
Case
HAF 912 advanced 
  hide details  
post #429 of 443
Quote:
Originally Posted by Noufel View Post

So Maxwell 2.0 is already doing software sheduling even without a proper driver, i don't think another driver will help nvidia in that mater .
it dosent matter what maxwell2 is doing. dx12 and the gpus that use half assed non defined standard feature sets are in its infancy along with windows 10 and vulcan. all this chest thumping over a few fps is meaningless.pascal and artic islands with a muture software package is the real battle. sad part is maxwell 2 is performing well in both dx11 and dx12. even more of a concern is the fiji cards barely matching maxwell 2 while the hawaii cards are strong. if there was any concern it would be for arctic islands. lets be real its way to early to tell. odds are the consumer dosent care about a few fps in some early tests. they wiil buy Nvidia like the apple people. Brand recognition sells. both nvidia and apple are proof of that regardless of the better performer.
post #430 of 443
DX12 does seem to remove the CPU overhead that has been hobbling AMD cards in dx11.

If not for Fury scaling badly, this would be quite a sweep for AMD. A 390X is like ~15% off a 980Ti, if Fury X would to better it by about 30% it would be faster than the OC'ed models(~10%) of 980Ti.
The biggest problem for AMD going ahead would be lack of clockspeed on GCN compared to nvidia GPUs.
    
CPUMotherboardGraphicsRAM
e2140@3.2Ghz abit IP35-E HIS IceQ4 4850 4GB 667@800 5-5-5-15 
OSPower
win xp 32 bit Corsair 450VX@stock 
  hide details  
    
CPUMotherboardGraphicsRAM
e2140@3.2Ghz abit IP35-E HIS IceQ4 4850 4GB 667@800 5-5-5-15 
OSPower
win xp 32 bit Corsair 450VX@stock 
  hide details  
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Video Game News
This thread is locked  
Overclock.net › Forums › Industry News › Video Game News › [Anand] Fable Legends DX12 Benchmark Analysis