Overclock.net › Forums › Intel › Intel CPUs › Haswell Overclocking Guide [With Statistics]
New Posts  All Forums:Forum Nav:

Haswell Overclocking Guide [With Statistics] - Page 1064

post #10631 of 19539
Quote:
Originally Posted by angelotti View Post

Warning: Spoiler! (Click to show)
The x264 benchmark/stability test needs major rework, not just updated executables.
The video file included in the test is major crap in terms of quality. When one executes the test's batch file, a presentation is displayed claiming it's going to encode a full HD 20Mb/s video file.
The video in question is only 1920x656 plus black bars at 6.8Mb/s (not 20), it is also sharp and quite pixelated in places. In order to stress the CPU as much as possible, you need a very complex source, with high quality noise (not sharp like disney toons), lots of details and motion.
Here is a test i've done. Downloaded a 20Mb/s trailer and encoded it through megui along with the included video. Both encodes resulted in smaller size output.
Important MeGUI settings CQ:18, preset - medium, profile - high, B-frames - 3, ref frames - 5, adaptive B-frames optimal, ME range - 24, ME algorithm - multi hex, subpixel - 10 qp-rd, macroblocks - all.
These are typical settings for a quality output, but i could go allot further with things like: preset - placebo, me algorithm - exhaustive or subpixel - full RD, etc.
The test:
Code:
Video                              trailer                    included 1080p
size (MB)                 353 (321 output)        101 (80 output)
source quality             20Mb/s                      6.8Mb/s             - the included video is 194% lower in quality
length                             2:27                              2:03                 - the included video is ~ 20% shorter
pixels                         1757184                       1259520             - the included video has ~39% less pixels (i removed the black bars)
frames                          3525                              2953                 - the included video has 19% less frames
fps                                     6                                  12.8                - the trailer's frame rate was 213% lower
time                                582                                230                  - the trailer took 253% more time to encode

So from the 20/39/19% that the included video is less in length/pixels/frames you could expect it to encode about 25% faster, but from those 213/253% in FPS and time, you can see that the trailer didn't take 25% but approx 233% longer. And this is all down to the quality of the source and the analysis of it and not so much the output.
If the x264 is to be an efficient real world stress test, then it needs some updating.

I hope i didn't look like a troll with this post.

Agreed and that's why 90% of the results ran with this are not really worth-while. The other 10% did use the latest encoder which made it better, I guess you've expanded on it. So is there a good/better source file that people can then use? That could then be put into the package on 1st page.

BTW: I also tried the package from 1st page for giggles just then but it just scrolls through all the tests (15 odd), nothing happens.
Edited by error-id10t - 3/2/14 at 2:18pm
Just a 'puter
(18 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 6700K ASUS Maximus VIII Hero iGPU thanks to Nvidia 970 3.5GB Corsair Vengeance LPX CMK8GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Samsung 950 PRO 512GB Crucial M4 256GB RX360 V3 Koolance CPU-380I 
CoolingCoolingOSMonitor
Koolance RP-452X2 Reservoir Koolance PMP-450 12V Variable Speed Pump Windows 10 Pro x64 benq XL2411Z 
KeyboardPowerCaseMouse
logitech G510 Seasonic X-760 Corsair 800D Mionix Castor 
AudioAudio
Creative Sound Blaster X-Fi Titanium HD DT 990 Premium 250Ohm 
  hide details  
Reply
Just a 'puter
(18 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7 6700K ASUS Maximus VIII Hero iGPU thanks to Nvidia 970 3.5GB Corsair Vengeance LPX CMK8GX4M2B3200C16 
Hard DriveHard DriveCoolingCooling
Samsung 950 PRO 512GB Crucial M4 256GB RX360 V3 Koolance CPU-380I 
CoolingCoolingOSMonitor
Koolance RP-452X2 Reservoir Koolance PMP-450 12V Variable Speed Pump Windows 10 Pro x64 benq XL2411Z 
KeyboardPowerCaseMouse
logitech G510 Seasonic X-760 Corsair 800D Mionix Castor 
AudioAudio
Creative Sound Blaster X-Fi Titanium HD DT 990 Premium 250Ohm 
  hide details  
Reply
post #10632 of 19539
Quote:
Originally Posted by Cyro999 View Post

You may be running old encoder which is slower speed and worse for checking stability. I'm using the download that was posted here that has a few things together, with a manually dropped in encoder version. 300mhz won't make my chip 22.5% faster than yours

Grabbed some download links out of my firefox log:

http://dc581.2shared.com/download/yNZzmwrI/x264_Stability_Test.7z?tsid=20140129-102227-7849b374

^modified x264 bench 5.0.1 from this thread, can use the x264 stability test file to run it after extracting it to a folder



and then encoder version that i used:
http://download.videolan.org/pub/videolan/x264/binaries/win64/x264-r2389-956c8d8.exe

^drop in the folder next to x264-64 (it's maybe in a subfolder?), then rename the original and make this the new x264-64

This mod doesn't seem to work for me. I replaced with the r2389 file that you linked and when I run the test, it skips pass 2 of every run.
post #10633 of 19539
Quote:
Originally Posted by angelotti View Post

The x264 benchmark/stability test needs major rework, not just updated executables.
The video file included in the test is major crap in terms of quality. When one executes the test's batch file, a presentation is displayed claiming it's going to encode a full HD 20Mb/s video file.
The video in question is only 1920x656 plus black bars at 6.8Mb/s (not 20), it is also sharp and quite pixelated in places. In order to stress the CPU as much as possible, you need a very complex source, with high quality noise (not sharp like disney toons), lots of details and motion.
Here is a test i've done. Downloaded a 20Mb/s trailer and encoded it through megui along with the included video. Both encodes resulted in smaller size output.
Important MeGUI settings CQ:18, preset - medium, profile - high, B-frames - 3, ref frames - 5, adaptive B-frames optimal, ME range - 24, ME algorithm - multi hex, subpixel - 10 qp-rd, macroblocks - all.
These are typical settings for a quality output, but i could go allot further with things like: preset - placebo, me algorithm - exhaustive or subpixel - full RD, etc.
The test:
Code:
Video                              trailer                    included 1080p
size (MB)                 353 (321 output)        101 (80 output)
source quality             20Mb/s                      6.8Mb/s             - the included video is 194% lower in quality
length                             2:27                              2:03                 - the included video is ~ 20% shorter
pixels                         1757184                       1259520             - the included video has ~39% less pixels (i removed the black bars)
frames                          3525                              2953                 - the included video has 19% less frames
fps                                     6                                  12.8                - the trailer's frame rate was 213% lower
time                                582                                230                  - the trailer took 253% more time to encode

So from the 20/39/19% that the included video is less in length/pixels/frames you could expect it to encode about 25% faster, but from those 213/253% in FPS and time, you can see that the trailer didn't take 25% but approx 233% longer. And this is all down to the quality of the source and the analysis of it and not so much the output.
If the x264 is to be an efficient real world stress test, then it needs some updating.

I hope i didn't look like a troll with this post.

If you've got a source file, I'd be happy to put it and the new encoders in the stability test download.
post #10634 of 19539
Quote:
Originally Posted by Zahix View Post

This mod doesn't seem to work for me. I replaced with the r2389 file that you linked and when I run the test, it skips pass 2 of every run.

Glad to see I'm not the only one haha. Curious what's going on really.
Main
(20 items)
 
  
CPUMotherboardGraphicsRAM
i5 4670K @ 4.7Ghz [1.284v] Z87X-UD4H [F7] MSI GTX 1070 Gaming X 2x4GB Samsung MV-3V4G3; 10-10-10-28 @ 2133Mhz [... 
Hard DriveHard DriveHard DriveHard Drive
1x Samsung 850 Pro 256GB 1x Crucial M500 960GB 1x WD4003FZEX 1x WD30EFRX 
CoolingCoolingOSMonitor
NH-D14 3x A15s @ 600RPM 2x Phanteks F140SP BBK (front), SFF21E (bottom) Win 10 Pro x64 Catleap 2B @ 119hz +1 
MonitorPowerCaseMouse
U3014 eVGA 750G2 Fractal R5 - Blackout Edition MS WMO 1.1a 
Mouse PadAudio
fUnc 1030 Creative Sound Blaster Z 
  hide details  
Reply
Main
(20 items)
 
  
CPUMotherboardGraphicsRAM
i5 4670K @ 4.7Ghz [1.284v] Z87X-UD4H [F7] MSI GTX 1070 Gaming X 2x4GB Samsung MV-3V4G3; 10-10-10-28 @ 2133Mhz [... 
Hard DriveHard DriveHard DriveHard Drive
1x Samsung 850 Pro 256GB 1x Crucial M500 960GB 1x WD4003FZEX 1x WD30EFRX 
CoolingCoolingOSMonitor
NH-D14 3x A15s @ 600RPM 2x Phanteks F140SP BBK (front), SFF21E (bottom) Win 10 Pro x64 Catleap 2B @ 119hz +1 
MonitorPowerCaseMouse
U3014 eVGA 750G2 Fractal R5 - Blackout Edition MS WMO 1.1a 
Mouse PadAudio
fUnc 1030 Creative Sound Blaster Z 
  hide details  
Reply
post #10635 of 19539
Quote:
Originally Posted by angelotti View Post

The x264 benchmark/stability test needs major rework, not just updated executables.
The video file included in the test is major crap in terms of quality. When one executes the test's batch file, a presentation is displayed claiming it's going to encode a full HD 20Mb/s video file.
The video in question is only 1920x656 plus black bars at 6.8Mb/s (not 20), it is also sharp and quite pixelated in places. In order to stress the CPU as much as possible, you need a very complex source, with high quality noise (not sharp like disney toons), lots of details and motion.
Here is a test i've done. Downloaded a 20Mb/s trailer and encoded it through megui along with the included video. Both encodes resulted in smaller size output.
Important MeGUI settings CQ:18, preset - medium, profile - high, B-frames - 3, ref frames - 5, adaptive B-frames optimal, ME range - 24, ME algorithm - multi hex, subpixel - 10 qp-rd, macroblocks - all.
These are typical settings for a quality output, but i could go allot further with things like: preset - placebo, me algorithm - exhaustive or subpixel - full RD, etc.
The test:
Code:
Video                              trailer                    included 1080p
size (MB)                 353 (321 output)        101 (80 output)
source quality             20Mb/s                      6.8Mb/s             - the included video is 194% lower in quality
length                             2:27                              2:03                 - the included video is ~ 20% shorter
pixels                         1757184                       1259520             - the included video has ~39% less pixels (i removed the black bars)
frames                          3525                              2953                 - the included video has 19% less frames
fps                                     6                                  12.8                - the trailer's frame rate was 213% lower
time                                582                                230                  - the trailer took 253% more time to encode

So from the 20/39/19% that the included video is less in length/pixels/frames you could expect it to encode about 25% faster, but from those 213/253% in FPS and time, you can see that the trailer didn't take 25% but approx 233% longer. And this is all down to the quality of the source and the analysis of it and not so much the output.
If the x264 is to be an efficient real world stress test, then it needs some updating.

I hope i didn't look like a troll with this post.

Of course not
Quote:
Originally Posted by Forceman View Post

If you've got a source file, I'd be happy to put it and the new encoders in the stability test download.

+1 for this. Better source is good but we need to be able to compare - this would be good for wider distribution (new source file, 2389 encoder thrown in and used, etc)
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 #10636 of 19539
Quote:
Originally Posted by Forceman View Post

If you've got a source file, I'd be happy to put it and the new encoders in the stability test download.

I will search for a proper source (better than the one i tested), but more is needed.
The 2 pass encode procedure came to life long ago from the need to give the CPU some “room to breathe”. It splits the analysis from decision making (the actual compress) in two separate processes. It is also a ‘target bitrate’ method, wich further lowers the cpu stress. What we need is the opposite, a ‘constant quality’ encode, wich has to make the analysis and the decisions at the same time. Also, being a quality oriented encode (as apposed to fixed or average bitrate) it has to decide wich frames need less ‘bits’ along with the decision of how to transit from one pixel to the next and from one frame to the next. CQ is more taxing to the CPU, therefor better suited.
post #10637 of 19539
Personally i don't 2-pass encode at all (aside from this bench), if you're not targeting a very specific file size i prefer to use crf; any live encoding for streaming etc will be done with bitrate-capped crf or CBR

Most of my encodes on that front are veryslow preset (offline), maybe medium depending on the content, length and what i want to do with it etc, whereas live encoding is more targeting ~medium-veryfast. I don't know enough about the encoder to tweak individual settings to great effect, aside from some of the more basic stuff, would be nice to know more
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 #10638 of 19539
x264 or prime95 to test the OC?
post #10639 of 19539
Quote:
Originally Posted by Caos View Post

x264 or prime95 to test the OC?

Did you read the OP? tongue.gif

Prime27.9 custom fft 1344-1344 with >90% of max ram after checking error checking and round off checking and setting it to high priority if you want harsh is a decent option. x264 is good and pretty aggressive too but it's not quite a catch-all i think, you have to or at least used to have to use it a bit smartly
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 #10640 of 19539
Thread Starter 
Quote:
Originally Posted by Minusorange View Post


I understand where you're coming from dude, and I'm not arguing for or against any one guide as they all have their merit and I believe anyone who truly wants to learn how OC properly will do what I did and use a combination of guides to get as much information as possible AND ask questions if they're stuck on something smile.gif As for real world vs synthetic, I honestly think if you can pass a synthetic then you should be fine with anything the real world throws at you!

I'll see what I can do.

 

I do my best to take the best info I see from other guides and incorporate it into my guide... and whatever info I need that can't be found, I test myself.

Undelwalt (2017)
(29 items)
 
 
Celapaleis (2013)
(16 items)
 
CPUMotherboardGraphicsRAM
7600k @ 5.197/1.392 1.392v z170 Asus Hero EVGA 1080ti SC Black (2100/6250) x2 8gb GSkill Trident Z 3600 @ 3804 15-15-15-32-2T 
Hard DriveHard DriveHard DriveCooling
Samsung 950 Pro 512gb (Undelwalt) WD Red 2tb (Pack Yak II) Seagate Expansion Drive 5tb (Phoenix Down II) 3x560mm HardwareLabs Nemesis GTX 
CoolingCoolingCoolingCooling
XPSC Raystorm Pro Watercool Heatkiller 1080ti Full Cover Waterblock EK XTOP Revo Dual D5 (Serial, PWM, v4) EK ZMT (1/2 - 3/4 ID OD) 
CoolingCoolingCoolingCooling
14xEK ACF (Compression Fittings) EK x4 250 (v2) Distilled Water + PT Nuke (Copper Sulphate) 24x140mm Silent Wings 3 
OSMonitorKeyboardPower
Windows 10 Pro 64bit Catleap 27 Inch 2560x1440 IPS 60hz Display Coolermaster Storm Trigger (Brown Switches) EVGA P2 1000w 
CaseMouseMouse PadAudio
Corsair Air 540 Logitech G Pro Tek Syndicate "Raze the World" Desk Mat O2 + Odac by Mayflower Electronics (Rev A) 
AudioAudioAudioAudio
Sennheiser HD 800 2xKrk Rokit 6 G2 Blue Yeti (Grey) Rode PSA-1 Arm 
Other
Aquaero 6 LT + Aluminum Heatsink 
CPUMotherboardGraphicsRAM
i5 6600k Delid @ 4.848/4.848 Asus z170 Hero MSI GTX 980ti @ 1499/4082 1.25v/134% Power 2 x 8gb Gskills Ripjaws V 3131 16-16-16-32 
Hard DriveHard DriveHard DriveHard Drive
Samsung Pro Series 840 256gb (Celapaleis) 1 TB Samsung Drive (Pack Yak) Samsung Pro Series 850 256gb (Celapaleis Reprise) Phoenix Down (External Backup) 
CoolingCoolingOSMonitor
Noctua D14 Kraken x61 Push-Pull for GPU Windows 10 Pro Catleap 2560x1440 @ 60hz 
KeyboardPowerCaseMouse
CM Storm Trigger Brown Switches EVGA Supernova 1000w P2 Corsair 540 Air Razer Abyssus 
Mouse PadAudioAudioAudio
Steelseries QcK Mini Blue Yeti Microphone Objective Dac + Amp Sescom AB Switch 
AudioAudioAudioAudio
Krk Rokit 6 G2 x2 Radioshack Ground Loop Isolator (For Rokits) Wooaudio Aluminum Headphone Stand Sennheiser HD 800 
AudioOther
Logitech c920 Webcam NZXT Sentry Mix 2 Fan Controller 
  hide details  
Reply
Undelwalt (2017)
(29 items)
 
 
Celapaleis (2013)
(16 items)
 
CPUMotherboardGraphicsRAM
7600k @ 5.197/1.392 1.392v z170 Asus Hero EVGA 1080ti SC Black (2100/6250) x2 8gb GSkill Trident Z 3600 @ 3804 15-15-15-32-2T 
Hard DriveHard DriveHard DriveCooling
Samsung 950 Pro 512gb (Undelwalt) WD Red 2tb (Pack Yak II) Seagate Expansion Drive 5tb (Phoenix Down II) 3x560mm HardwareLabs Nemesis GTX 
CoolingCoolingCoolingCooling
XPSC Raystorm Pro Watercool Heatkiller 1080ti Full Cover Waterblock EK XTOP Revo Dual D5 (Serial, PWM, v4) EK ZMT (1/2 - 3/4 ID OD) 
CoolingCoolingCoolingCooling
14xEK ACF (Compression Fittings) EK x4 250 (v2) Distilled Water + PT Nuke (Copper Sulphate) 24x140mm Silent Wings 3 
OSMonitorKeyboardPower
Windows 10 Pro 64bit Catleap 27 Inch 2560x1440 IPS 60hz Display Coolermaster Storm Trigger (Brown Switches) EVGA P2 1000w 
CaseMouseMouse PadAudio
Corsair Air 540 Logitech G Pro Tek Syndicate "Raze the World" Desk Mat O2 + Odac by Mayflower Electronics (Rev A) 
AudioAudioAudioAudio
Sennheiser HD 800 2xKrk Rokit 6 G2 Blue Yeti (Grey) Rode PSA-1 Arm 
Other
Aquaero 6 LT + Aluminum Heatsink 
CPUMotherboardGraphicsRAM
i5 6600k Delid @ 4.848/4.848 Asus z170 Hero MSI GTX 980ti @ 1499/4082 1.25v/134% Power 2 x 8gb Gskills Ripjaws V 3131 16-16-16-32 
Hard DriveHard DriveHard DriveHard Drive
Samsung Pro Series 840 256gb (Celapaleis) 1 TB Samsung Drive (Pack Yak) Samsung Pro Series 850 256gb (Celapaleis Reprise) Phoenix Down (External Backup) 
CoolingCoolingOSMonitor
Noctua D14 Kraken x61 Push-Pull for GPU Windows 10 Pro Catleap 2560x1440 @ 60hz 
KeyboardPowerCaseMouse
CM Storm Trigger Brown Switches EVGA Supernova 1000w P2 Corsair 540 Air Razer Abyssus 
Mouse PadAudioAudioAudio
Steelseries QcK Mini Blue Yeti Microphone Objective Dac + Amp Sescom AB Switch 
AudioAudioAudioAudio
Krk Rokit 6 G2 x2 Radioshack Ground Loop Isolator (For Rokits) Wooaudio Aluminum Headphone Stand Sennheiser HD 800 
AudioOther
Logitech c920 Webcam NZXT Sentry Mix 2 Fan Controller 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Haswell Overclocking Guide [With Statistics]