Overclock.net › Forums › Industry News › Hardware News › [VC] 290Xvs780 In Non Quiet Mode At 2160P
New Posts  All Forums:Forum Nav:

[VC] 290Xvs780 In Non Quiet Mode At 2160P - Page 48

post #471 of 475
My Sapphire R9 280X Toxics (2 of them) arrive on Saturday, best price per bang out there, going to clock them both up by another 10%, and they'll then be faster than a 780 (oh and they cost the same as a single RIP-OFF 780) Been with Nvidia as long as I can remember, and I'm sick of being ripped-off!

Good bye Nvidia, you're not going to be laughing all the way to bank at my expense again!
Edited by Mombasa69 - 10/21/13 at 2:51am
post #472 of 475
Well looks like AMD gets my hard earned cash again.
Frankenputer
(15 items)
 
  
CPUMotherboardGraphicsGraphics
I7 6700 skylake 1151 msi z170a sli plus MSI GTX 980ti EVGA GTX 980ti 
RAMHard DriveOptical DriveCooling
32 gb Corsair Vengance samsung EVO 840 Sony DVD-+RW DL + Sony bdu Custom Loop 
OSMonitorPowerCase
Windows 10 home BenQ VW2424H LCD Monitor EVGA 1000 watt corsair 400r 
  hide details  
Reply
Frankenputer
(15 items)
 
  
CPUMotherboardGraphicsGraphics
I7 6700 skylake 1151 msi z170a sli plus MSI GTX 980ti EVGA GTX 980ti 
RAMHard DriveOptical DriveCooling
32 gb Corsair Vengance samsung EVO 840 Sony DVD-+RW DL + Sony bdu Custom Loop 
OSMonitorPowerCase
Windows 10 home BenQ VW2424H LCD Monitor EVGA 1000 watt corsair 400r 
  hide details  
Reply
post #473 of 475
Quote:
Originally Posted by HeadlessKnight View Post

Scaling most of the time is not linear with Clock increase.

Only because there is usually some sort of bottleneck elsewhere.
Quote:
Originally Posted by HeadlessKnight View Post

If you get a 47% increase in fps from a 47% increase in core clock and 30% increase in memory clock then the scaling is not linear.

Sure it is. In this case, the core was the bottleneck, but memory bandwidth was not, nor was anything off the card.
Quote:
Originally Posted by HeadlessKnight View Post

And I got about 24% increase in Metro Last Light and about ~14.8% in Metro 2033. GPU usage was 99% all the time, In Skyrim (ENB) I got 25% increase at those clocks. So your point is moot and null.
So performance vs clock is not scaling linearly, it depends on other factors too such as the architecture itself & the game engine.

Using commonly reported GPU utilization percentages is a serious flaw in this argument.

If you want to prove you are not CPU (or system memory, or PCI-E or, HT/QPI, or whatever) limited, you need to change CPU (or other) speed. If performance doesn't change, then it's not a limiting factor, but if it does, it likely is, no matter what anything else reports.

Surely the game engine can be limited by various factors, some of which may not be apparent without extensive testing, but if it's only limited by the GPU, increasing GPU performance cannot possibly result in anything but near linear scaling. To imply otherwise is to say that a GPU's IPC varies with clock speed, which is certainly not the case.
Primary
(15 items)
 
Secondary
(13 items)
 
Vishera Testbed
(11 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3GHz, 1.225v Gigabyte X99 SOC Champion (F4m) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-12-12-28-T1, 1.35v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 CM Storm Spawn Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 1600MT/s 7-8-8-19-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 1600MT/s 7-8-8-1... OCZ (Toshiba) Trion 150 120GB Hitachi Deskstar 7k1000.C 1TB 2x Seagate 7200.10 RAID 0 
CoolingOSPowerCase
Prolimatech Genesis + 2x140mm Cougar 1200rpm Windows Server 2008 R2 Antec TP-750 Antec P182 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
FX-9590 @ 5GHz, 1.55v ASUS Sabertooth 990FX R2.0 Reference NVIDIA GTX 780 2x8GiB G.Skill DDR3-1866 10-11-10-30-T1, 1.5v 
Hard DriveHard DriveCoolingOS
Crucial M500 480GB 2x Samsung Spinpoint F1 1TB XSPC RX360 + X20 750 + Raystorm Windows 7 Pro SP1 x64 
MonitorPowerCase
Dell S2740L Seasonic SS-860XP2 Coolermaster HAF-932 
  hide details  
Reply
Primary
(15 items)
 
Secondary
(13 items)
 
Vishera Testbed
(11 items)
 
CPUMotherboardGraphicsRAM
5820K @ 4.3GHz, 1.225v Gigabyte X99 SOC Champion (F4m) 2x Sapphire R9 290X Tri-X OC New Edition (10036... 4x4GiB Crucial @ 2667, 12-12-12-28-T1, 1.35v 
Hard DriveHard DriveHard DriveCooling
Plextor M6e 128GB (fw 1.05) M.2 (PCI-E 2.0 2x) 2x Crucial M4 256GB 4x WD Scorpio Black 500GB Cooler Master Nepton 280L 
OSMonitorKeyboardPower
Windows 7 Professional x64 SP1 BenQ BL3200PT Filco Majestouch Tenkeyless (MX Brown) Corsair RM1000x 
CaseMouseAudio
Fractal Design Define R4 CM Storm Spawn Realtek ALC1150 + M-Audio AV40 
CPUMotherboardGraphicsRAM
X5670 Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS Reference R9 290X w/Stilt's MLU 1000e / 1375m E... 2x Samsung MV-3V4G3D/US @ 1600MT/s 7-8-8-19-T1,... 
RAMHard DriveHard DriveHard Drive
1x Crucial BLT4G3D1608ET3LX0 @ 1600MT/s 7-8-8-1... OCZ (Toshiba) Trion 150 120GB Hitachi Deskstar 7k1000.C 1TB 2x Seagate 7200.10 RAID 0 
CoolingOSPowerCase
Prolimatech Genesis + 2x140mm Cougar 1200rpm Windows Server 2008 R2 Antec TP-750 Antec P182 
Audio
ASUS Xonar DS 
CPUMotherboardGraphicsRAM
FX-9590 @ 5GHz, 1.55v ASUS Sabertooth 990FX R2.0 Reference NVIDIA GTX 780 2x8GiB G.Skill DDR3-1866 10-11-10-30-T1, 1.5v 
Hard DriveHard DriveCoolingOS
Crucial M500 480GB 2x Samsung Spinpoint F1 1TB XSPC RX360 + X20 750 + Raystorm Windows 7 Pro SP1 x64 
MonitorPowerCase
Dell S2740L Seasonic SS-860XP2 Coolermaster HAF-932 
  hide details  
Reply
post #474 of 475
Quote:
Originally Posted by Mombasa69 View Post

My Sapphire R9 280X Toxics (2 of them) arrive on Saturday, best price per bang out there, going to clock them both up by another 10%, and they'll then be faster than a 780 (oh and they cost the same as a single RIP-OFF 780) Been with Nvidia as long as I can remember, and I'm sick of being ripped-off!

Good bye Nvidia, you're not going to be laughing all the way to bank at my expense again!


Id rather not have to deal with the headaches of multi gpu's


Also: im not sure why you feel like you were being ripped off, top end gpus have never been good bang for your buck, only amplified by the fact that amd had no answer for it so nvidia could charge whatever they want
Edited by Blindsay - 10/22/13 at 6:47am
Server/Gaming Rig
(22 items)
 
  
CPUCPUMotherboardGraphics
Intel E5-2670 intel E5-2670 Asus Z9PE-D8 WS Visiontek R9 390 
RAMHard DriveHard DriveHard Drive
8x4GB DDR3 1333 ECC Registered Toshiba 5TB X300 Seagate 4TB HGST 4TB 
Hard DriveHard DriveHard DriveHard Drive
Sandisk 240GB Sandisk 480GB Samsung 830 128GB Toshiba 6TB X300 
Hard DriveHard DriveOptical DriveOS
Toshiba 5TB X300 Toshiba 5TB X300 WH16NS40 Windows 10 Pro 
MonitorMonitorKeyboardPower
Dell S1715H Dell u2311h blackwidow ultimate Kingwin Lazer platinum 1000w 
CaseMouse
Corsair 800D Deathadder Chroma 
  hide details  
Reply
Server/Gaming Rig
(22 items)
 
  
CPUCPUMotherboardGraphics
Intel E5-2670 intel E5-2670 Asus Z9PE-D8 WS Visiontek R9 390 
RAMHard DriveHard DriveHard Drive
8x4GB DDR3 1333 ECC Registered Toshiba 5TB X300 Seagate 4TB HGST 4TB 
Hard DriveHard DriveHard DriveHard Drive
Sandisk 240GB Sandisk 480GB Samsung 830 128GB Toshiba 6TB X300 
Hard DriveHard DriveOptical DriveOS
Toshiba 5TB X300 Toshiba 5TB X300 WH16NS40 Windows 10 Pro 
MonitorMonitorKeyboardPower
Dell S1715H Dell u2311h blackwidow ultimate Kingwin Lazer platinum 1000w 
CaseMouse
Corsair 800D Deathadder Chroma 
  hide details  
Reply
post #475 of 475
Quote:
Originally Posted by Blameless View Post


Sure it is. In this case, the core was the bottleneck, but memory bandwidth was not, nor was anything off the card.

You can't decide based on that, Most of the realistic situations scaling won't be 100% vs core clock/ memory clock, unless there are other problems such as a severe memory bandwidth bottleneck.
Quote:
Originally Posted by Blameless View Post


Using commonly reported GPU utilization percentages is a serious flaw in this argument.

If you want to prove you are not CPU (or system memory, or PCI-E or, HT/QPI, or whatever) limited, you need to change CPU (or other) speed. If performance doesn't change, then it's not a limiting factor, but if it does, it likely is, no matter what anything else reports.

OK. Probably you are right smile.gif and I did exactly what you asked me for.

Here are the testings I've done :-

Gigabyte HD 7950 Clocks : @ 800/1250 (Base Clocks) ,
1100/1250 (+37.50% OC on Core, Stock Memory),
1100/1625 (+37.50% OC on Core, +30.0% OC on Memory),
1210/1250 (+51.25% OC on Core, Stock Memory) ,
1210/1625 (+51.25% OC on Core, +30.0% OC on Memory)

i7 870 : 3300 MHz, PCI-E: 100 MHz, Memory : DDR3-1322

Test 1 : Metro 2033 (All Maxed including 4x MSAA +DOF -PhysX)

34.40 fps (800/1250)
40.11 fps (1100/1250) +16.59% over stock
41.44 fps (1210/1250) +20.47% over stock
43.92 fps (1100/1625) +27.67% over stock
45.87 fps (1210/1625) +33.34% over stock

Test 2 : Metro Last Light (All Maxed including SSAA -PhysX)

27.03 fps (800/1250)
33.22 fps (1100/1250) +22.90% over stock
34.97 fps (1210/1250) +27.67% over stock
35.52 fps (1100/1625) +29.37% over stock
37.56 fps (1210/1625) +38.96% over stock

Test 3 : Heaven Benchmark 4.0 (1920x1080 Max 0xAA)

36.20 fps (800/1250)
46.10 fps (1100/1250) +27.34% over stock
47.60 fps (1100/1650) +31.49% over stock
49.40 fps (1210/1250) +36.46% over stock
51.30 fps (1210/1625) +41.71% over stock

i7 870 : 4340 MHz, PCI-E: 105 MHz, Memory : DDR3-1651

Test 4 : Metro 2033

34.45 fps (800/1250)
40.23 fps (1100/1250) +16.78% over stock
41.75 fps (1210/1250) +21.19% over stock
43.91 fps (1100/1625) +27.46% over stock
45.91 fps (1210/1625) +33.27% over stock

Test 5 : Metro Last Light (All Maxed including SSAA -PhysX)

27.00 fps (800/1250)
33.25 fps (1100/1250) +23.15% over stock
34.96 fps (1210/1250) +29.48% over stock
35.56 fps (1100/1625) +31.70% over stock
37.55 fps (1210/1625) +39.00% over stock

Test 6 : Heaven Benchmark 4.0 (1920x1080 Max 0xAA)

36.20 fps (800/1250)
46.30 fps (1100/1250) +27.90% over stock
47.80 fps (1100/1650) +32.00% over stock
49.30 fps (1210/1250) +36.19% over stock
51.40 fps (1210/1625) +41.99% over stock

In the first test for example (1210/1625) gave me 45.87 from 34.4 on stock clocks that's 33.34% increase from 51.25% core oc and that without even taking into consideration the 30% memory overclock. Using Heaven Benchmark 4 as an example same thing can be said but the scaling is a bit higher, 41.71%. still not perfect scaling and far from close. Also notice 7950 @ 1100/1625 is faster than itself @ 1210/1250 in Metro 2033/ Last Light.

Conclusion : Graphics cards overclocks act differently based on game engine & the efficiency of the architecture itself and how balanced it is. CPU doesn't matter very much as long as it is up to the task.
Edited by HeadlessKnight - 10/22/13 at 10:17am
Haswell-E
(11 items)
 
   
CPUMotherboardGraphicsRAM
Intel Core i7-5820K @ 4.4 GHz Asus X99-A/USB 3.1 (3402) MSI GTX 980 Ti Gaming 6 GB  Corsair Vengeance LPX 16 GB (4x4) DDR4-3000 
Hard DriveHard DriveCoolingOS
Samsung 850 Pro 1 TB Samsung 950 Pro NVMe 512 GB Noctua NH-D15 Windows 10 
MonitorPowerCase
Asus VG236H 23" 120 ZM1000-HP CoolerMaster HAF 922 
CPUMotherboardGraphicsRAM
Intel Core i7 4720HQ @ 2.60 GHz Asus G751Y Asus GTX 980M 4 GB 16 GB DDR3L-1600 
Hard DriveOS
SanDisk 128 GB + HGST 1 TB Windows 10 Home 
  hide details  
Reply
Haswell-E
(11 items)
 
   
CPUMotherboardGraphicsRAM
Intel Core i7-5820K @ 4.4 GHz Asus X99-A/USB 3.1 (3402) MSI GTX 980 Ti Gaming 6 GB  Corsair Vengeance LPX 16 GB (4x4) DDR4-3000 
Hard DriveHard DriveCoolingOS
Samsung 850 Pro 1 TB Samsung 950 Pro NVMe 512 GB Noctua NH-D15 Windows 10 
MonitorPowerCase
Asus VG236H 23" 120 ZM1000-HP CoolerMaster HAF 922 
CPUMotherboardGraphicsRAM
Intel Core i7 4720HQ @ 2.60 GHz Asus G751Y Asus GTX 980M 4 GB 16 GB DDR3L-1600 
Hard DriveOS
SanDisk 128 GB + HGST 1 TB Windows 10 Home 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [VC] 290Xvs780 In Non Quiet Mode At 2160P