Overclock.net › Forums › Components › Memory › An Independent Study: Does The Speed Of RAM Directly Affect FPS During High CPU Overhead Scenarios?
New Posts  All Forums:Forum Nav:

An Independent Study: Does The Speed Of RAM Directly Affect FPS During High CPU Overhead Scenarios? - Page 15

post #141 of 277
I would say that the FPS gains in Crysis 3 could be considered within the margin of error. I don't think margin of error is taken into account enough with FPS in video games. If you're only getting a 6 FPS boost I would say that's a normal variation and has little to do with any hardware upgrades.

As for Tomb Raider, that's a pretty significant boost though.
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
post #142 of 277
This is interesting. By all means update periodically with new games. Of course, since I have 2400 mhz CL 10 ram I'm curious about that particular benchmark, but still... awesome. Always nice to see some tests done that the reviewing sites don't bother with.
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K 4.4GHz ASUS MAXIMUS VII FORMULA ASUS Strix Geforce GTX 980 ASUS Strix Geforce GTX 980 
RAMHard DriveOptical DriveCooling
G.SKILL Trident X 32GB 2400 Mhz CL10 DDR3 Samsung 850 Pro SSD 256GB ASUS Blue Ray Swiftech H240-X CPU Liquid Cooling Kit 
OSKeyboardPowerCase
Windows 7 Professional Natural Ergonomic EVGA 750W G2 be quiet silent atx mid tower 
  hide details  
Reply
    
CPUMotherboardGraphicsGraphics
Intel Core i7-4790K 4.4GHz ASUS MAXIMUS VII FORMULA ASUS Strix Geforce GTX 980 ASUS Strix Geforce GTX 980 
RAMHard DriveOptical DriveCooling
G.SKILL Trident X 32GB 2400 Mhz CL10 DDR3 Samsung 850 Pro SSD 256GB ASUS Blue Ray Swiftech H240-X CPU Liquid Cooling Kit 
OSKeyboardPowerCase
Windows 7 Professional Natural Ergonomic EVGA 750W G2 be quiet silent atx mid tower 
  hide details  
Reply
post #143 of 277
Quote:
Originally Posted by Thready View Post

I would say that the FPS gains in Crysis 3 could be considered within the margin of error. I don't think margin of error is taken into account enough with FPS in video games. If you're only getting a 6 FPS boost I would say that's a normal variation and has little to do with any hardware upgrades.

As for Tomb Raider, that's a pretty significant boost though.

It's fairly easy to get way less than a 9% margin for error in testing. I don't even bother testing if i have margin for errors anywhere near that big!

My starcraft 2 tests when i did them in HOTS were not only accurate to 1%, but accurate to within 0.1%. I've even ran a 30 second benchmark three times and had two of the results come back nearly identical - literally the same amount of frames, like 3083 frames in both benchmark results giving the same min, avg and max FPS. Now other games like Crysis 3 are non deterministic but 9% is a LOT!

I'm fairly sure that BradleyW wouldn't have had to rely on tests with a 10% margin for error (that happened to randomly fall in his favor across testing more than 2 RAM speeds) in order to get results like this.
Edited by Cyro999 - 10/15/15 at 5:30pm
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
post #144 of 277
Quote:
Originally Posted by Cyro999 View Post

It's fairly easy to get way less than a 9% margin for error in testing. I don't even bother testing if i have margin for errors anywhere near that big!

My starcraft 2 tests when i did them in HOTS were not only accurate to 1%, but accurate to within 0.1%. I've even ran a 30 second benchmark three times and had two of the results come back nearly identical - literally the same amount of frames, like 3083 frames in both benchmark results giving the same min, avg and max FPS. Now other games like Crysis 3 are non deterministic but 9% is a LOT!

I'm fairly sure that BradleyW wouldn't have had to rely on tests with a 10% margin for error (that happened to randomly fall in his favor across testing more than 2 RAM speeds) in order to get results like this.

Yeah I think to eliminate the margin for error you should test at least 3 times. But because framerates can vary so widely from scene to scene, to say that more RAM gave you a 2 or 3 fps boost is kind of hard to say isn't it? You would have to test many times in order to be positively sure that those 2 or 3 extra frames were actually because of the RAM. If it were 10 or more frames, then it would be obvious from the beginning what was doing it, but if you're talking about a small amount, even if they're averages from 3 different tests each, you could very well see them moving closer and closer together with more testing to the point where you can't tell a difference. Like a bell curve effect.

edit: I might have misused the term margin of error, I haven't taken stats yet for my degree :/ but hopefully my overly long explanation can clarify what I mean.
Edited by Thready - 10/15/15 at 10:34pm
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
post #145 of 277
Quote:
Yeah I think to eliminate the margin for error you should test at least 3 times

I kinda already assume that most tests do that. There's no need for the games that give very consistent results (2 tests is fine if they match up) but for others, 3-5 test runs help a lot to increase certainty of results
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
post #146 of 277
Quote:
Originally Posted by Cyro999 View Post

I kinda already assume that most tests do that. There's no need for the games that give very consistent results (2 tests is fine if they match up) but for others, 3-5 test runs help a lot to increase certainty of results

I think I was referring to within normal variance. Not within margin of error. Oh I'm not going to pass graduate stats am I? lol

And I was just saying 3 tests but I'm not sure how many tests are usually done. I'm just saying in general, it's a good practice to use 3 or more because in psychology we like to make things look as smooth and as much like a bell curve as possible, and to get that you have to test and retest and retest. etc, etc. I'm not actually sure how many tests we use with things like FPS in games.

My thinking is that if it's within a few frames of the other RAM, how do you know the RAM is the reason why, and how do you know it's the size of RAM and not other things like some timings or something that doesn't have to do with the speed of the RAM. Because with 2 or 3 frames, it could be virtually anything.

I'm not that knowledgeable about how benchmarks are calculated and I'm wanting to get into it a bit more so I'm just trying to think outside the box here and learn a thing or two.
Edited by Thready - 10/16/15 at 9:37am
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
post #147 of 277
Thread Starter 
Quote:
Originally Posted by Thready View Post

I would say that the FPS gains in Crysis 3 could be considered within the margin of error. I don't think margin of error is taken into account enough with FPS in video games. If you're only getting a 6 FPS boost I would say that's a normal variation and has little to do with any hardware upgrades.

As for Tomb Raider, that's a pretty significant boost though.

Margin of error would be 1 FPS.

All tests have been repeated several times. Results were always consistent. A 6 FPS gain was hardware based. 6 FPS is an excellent boost in min frames.
Quote:
Originally Posted by Thready View Post

I think I was referring to within normal variance. Not within margin of error. Oh I'm not going to pass graduate stats am I? lol

And I was just saying 3 tests but I'm not sure how many tests are usually done. I'm just saying in general, it's a good practice to use 3 or more because in psychology we like to make things look as smooth and as much like a bell curve as possible, and to get that you have to test and retest and retest. etc, etc. I'm not actually sure how many tests we use with things like FPS in games.

My thinking is that if it's within a few frames of the other RAM, how do you know the RAM is the reason why, and how do you know it's the size of RAM and not other things like some timings or something that doesn't have to do with the speed of the RAM. Because with 2 or 3 frames, it could be virtually anything.

I'm not that knowledgeable about how benchmarks are calculated and I'm wanting to get into it a bit more so I'm just trying to think outside the box here and learn a thing or two.

Capacity does not increase FPS in any of these particular tests. This variable can be eliminated. If the applications could use more than 16GB, then such variable would come into play.
Edited by BradleyW - 10/17/15 at 7:11pm
X79-GCN
(22 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3930K 4.5GHz HT GIGABYTE GA-X79-UP4 AMD R9-290X GEil Evo Potenza DDR3 2400MHz CL10 (4x4GB) 
Hard DriveCoolingCoolingCooling
Samsung 840 Pro 120GB EK Supremacy (CPU) NF F12's P/P (360 Rad)  NF A14's (420 Rad)  
CoolingCoolingCoolingCooling
XSPC Chrome Compression Fittings EK RES X3 150 Primochill PremoFlex Advanced LRT Clear 1/2 ID EK-FC (R9 290X) 
CoolingCoolingCoolingOS
EK D5 Vario Top-X  Phobya G-Changer V2 360mm Phobya G-Changer V2 420mm Win 10 x64 Pro 
MonitorKeyboardPowerCase
BenQ XR3501 35" Curved Corsair Vengeance K90 Seasonic X-1250 Gold (v2) Corsair 900D 
MouseAudio
Logitech G400s Senn HD 598 
  hide details  
Reply
X79-GCN
(22 items)
 
  
CPUMotherboardGraphicsRAM
Intel 3930K 4.5GHz HT GIGABYTE GA-X79-UP4 AMD R9-290X GEil Evo Potenza DDR3 2400MHz CL10 (4x4GB) 
Hard DriveCoolingCoolingCooling
Samsung 840 Pro 120GB EK Supremacy (CPU) NF F12's P/P (360 Rad)  NF A14's (420 Rad)  
CoolingCoolingCoolingCooling
XSPC Chrome Compression Fittings EK RES X3 150 Primochill PremoFlex Advanced LRT Clear 1/2 ID EK-FC (R9 290X) 
CoolingCoolingCoolingOS
EK D5 Vario Top-X  Phobya G-Changer V2 360mm Phobya G-Changer V2 420mm Win 10 x64 Pro 
MonitorKeyboardPowerCase
BenQ XR3501 35" Curved Corsair Vengeance K90 Seasonic X-1250 Gold (v2) Corsair 900D 
MouseAudio
Logitech G400s Senn HD 598 
  hide details  
Reply
post #148 of 277
Quote:
Originally Posted by BradleyW View Post

Margin of error would be 1 FPS.

All tests have been repeated several times. Results were always consistent. A 6 FPS gain was hardware based. 6 FPS is an excellent boost in min frames.
Capacity does not increase FPS in any of these particular tests. This variable can be eliminated. If the applications could use more than 16GB, then such variable would come into play.

thanks for the explanation
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
Zen
(17 items)
 
  
CPUMotherboardGraphicsRAM
i5-6600k Gigabyte Z170XP-SLI RX 480 4GB 24 GB DDR4 2133MHz 
Hard DriveHard DriveHard DriveHard Drive
Samsung 850 evo OCZ Vertex 4 Crucial MX 500 256GB WD Black 3 TB 
Hard DriveCoolingOSMonitor
Seagate 1 TB H60 Windows 10 Asus mx27a 
MonitorKeyboardPowerMouse
hp 2009 m Corsair MR Brown Antec Earthwatts 650 Razer Naga 
Audio
Soundblaster Omni 
  hide details  
Reply
post #149 of 277
Nice to see some tests on single vs double sided dims
Seems that it improves bandwidth but games mostly care about overall latency
Thanks for your work testing

Here is some ram tests that are in progress with a 4930k and on the third page of that thread there is some links to some older tests i ran on ram with a 2600k in arma\thief at different cpu clock speeds and ram latency
http://hardforum.com/showthread.php?t=1872293
post #150 of 277
I just got a 5930k and some 3200mhz Trident Z. If I can I'll run some benchmarks over the weekend on it. Subbed too. I never really gave any good thought to the effects of ram speeds. I'm ignorant.
2012-2015
(14 items)
 
2015-present
(13 items)
 
 
CPUMotherboardGraphicsRAM
i7-3930k @ 3.2 MSI X79A-GD45 Plus EVGA GTX 980 SC ACX 2.0 16GB Adata 
Hard DriveHard DriveCoolingOS
1 TB Toshiba 7200RPM Intel 730 Series 256GB SSD CLLC on CPU Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Chroma 850 Watt In Win Mana-136 Razer Deathadder Chroma 
Mouse PadAudio
Generic On-Board 
CPUMotherboardGraphicsRAM
i7-5930k @ 4.4ghz 1.276v MSI X99 Godlike Gaming MSI GTX 980 Ti Lightning @ 1460 core 3850 mem Trident Z 3200Mhz 16GB 
Hard DriveOptical DriveCoolingOS
Intel 730 series 240GB None Corsair H110i Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Chroma EVGA 1200 P2 EVGA DG-87 Razer Deathadder Chroma 
Mouse Pad
Generic RIT Tiger Mousepad 
  hide details  
Reply
2012-2015
(14 items)
 
2015-present
(13 items)
 
 
CPUMotherboardGraphicsRAM
i7-3930k @ 3.2 MSI X79A-GD45 Plus EVGA GTX 980 SC ACX 2.0 16GB Adata 
Hard DriveHard DriveCoolingOS
1 TB Toshiba 7200RPM Intel 730 Series 256GB SSD CLLC on CPU Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Chroma 850 Watt In Win Mana-136 Razer Deathadder Chroma 
Mouse PadAudio
Generic On-Board 
CPUMotherboardGraphicsRAM
i7-5930k @ 4.4ghz 1.276v MSI X99 Godlike Gaming MSI GTX 980 Ti Lightning @ 1460 core 3850 mem Trident Z 3200Mhz 16GB 
Hard DriveOptical DriveCoolingOS
Intel 730 series 240GB None Corsair H110i Windows 10 
KeyboardPowerCaseMouse
Razer Blackwidow Chroma EVGA 1200 P2 EVGA DG-87 Razer Deathadder Chroma 
Mouse Pad
Generic RIT Tiger Mousepad 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Memory
  • An Independent Study: Does The Speed Of RAM Directly Affect FPS During High CPU Overhead Scenarios?
Overclock.net › Forums › Components › Memory › An Independent Study: Does The Speed Of RAM Directly Affect FPS During High CPU Overhead Scenarios?