Overclock.net › Forums › Intel › Intel Motherboards › Gaming and mouse response BIOS optimization guide for modern PC hardware
New Posts  All Forums:Forum Nav:

Gaming and mouse response BIOS optimization guide for modern PC hardware - Page 287

post #2861 of 3535
Quote:
Originally Posted by AndiWandi View Post



Reminder you don't have to turn off all power-saving options in BIOS to get decent latency.
This is with C-States, HPET, Turbo Boost all enabled.

You realize 0.0 page fault errors just means you sat idle at desktop with nothing open? Matter of fact, I wouldn't be surprised if you photoshopped it, as LatencyMon itself generates pagefaults. Even then, this is a useless screenshot, as it doesn't show whether your system can handle latency well. The point of the tool is troubleshoot during audio/video playback to see if your system can handle real-time playback without any major faults. It's rather misleading to people who don't know any better.

"Hey guys, check out my super low latency while I'm idle at desktop with everything closed! I won't even be using the system, but man it feels so responsive looking at the desktop with the 0.05ms improvement!"

This thread always give me a good laugh.
Edited by gene-z - 6/5/16 at 7:57pm
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
post #2862 of 3535
Quote:
Originally Posted by gene-z View Post

You realize 0.0 page fault errors just means you sat idle at desktop with nothing open? Matter of fact, I wouldn't be surprised if you photoshopped it, as LatencyMon itself generates pagefaults. Even then, this is a useless screenshot, as it doesn't show whether your system can handle latency well. The point of the tool is troubleshoot during audio/video playback to see if your system can handle real-time playback without any major faults. It's rather misleading to people who don't know any better.

"Hey guys, check out my super low latency while I'm idle at desktop with everything closed! I won't even be using the system, but man it feels so responsive looking at the desktop with the 0.05ms improvement!"

This thread always give me a good laugh.

Nothing "open" (running*) would be when the PC is off. Page faults are unrelated to latency. We use the tool to measure latency.

But nice troll attempt. thumb.gif
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
post #2863 of 3535
Quote:
Originally Posted by gene-z View Post

You realize 0.0 page fault errors just means you sat idle at desktop with nothing open? Matter of fact, I wouldn't be surprised if you photoshopped it, as LatencyMon itself generates pagefaults. Even then, this is a useless screenshot, as it doesn't show whether your system can handle latency well. The point of the tool is troubleshoot during audio/video playback to see if your system can handle real-time playback without any major faults. It's rather misleading to people who don't know any better.

"Hey guys, check out my super low latency while I'm idle at desktop with everything closed! I won't even be using the system, but man it feels so responsive looking at the desktop with the 0.05ms improvement!"

This thread always give me a good laugh.

Seeing as you cared enough to respond to my post, I'd like to address some of your points:

Page fault errors being zero after five minutes of idling is not a safe bet. In fact, I've had problems with it spiking into the high red regions after a minute even on idle, giving cause to the red LatencyMon notification. One of the measures I have done fixed it. It's that simple. Also, why would I want to photoshop a screenshot for an internet forum? It's weird you should assume I take this stuff that seriously.

If you go over my posts here, you will see that I'm far from advocating every placebo tweak here on this thread, and the only thing I mentioned in this post was that turning off all power-saving options might not be necessary. As far as testing under load goes, you are correct. In no way would my screenshot reflect load latencies.
post #2864 of 3535
Quote:
Originally Posted by Trull View Post

Nothing "open" (running*) would be when the PC is off. Page faults are unrelated to latency. We use the tool to measure latency.

But nice troll attempt. thumb.gif

You realize the application is measuring the time it takes to resolve the page fault? If it was unrelated to latency, it wouldn't even be in the application, lol. Why do you think it's measured in µs cycle time like the rest of the counters? Take some time and go read the LatencyMon website.
Quote:
Originally Posted by AndiWandi View Post

Seeing as you cared enough to respond to my post, I'd like to address some of your points:

Page fault errors being zero after five minutes of idling is not a safe bet. In fact, I've had problems with it spiking into the high red regions after a minute even on idle, giving cause to the red LatencyMon notification. One of the measures I have done fixed it. It's that simple. Also, why would I want to photoshop a screenshot for an internet forum? It's weird you should assume I take this stuff that seriously.

If you go over my posts here, you will see that I'm far from advocating every placebo tweak here on this thread, and the only thing I mentioned in this post was that turning off all power-saving options might not be necessary. As far as testing under load goes, you are correct. In no way would my screenshot reflect load latencies.

Exactly my point. You're measuring an idle system. You're not getting an accurate representation of what you're after and what the tool is intended to do, unless of course your goal is sitting and starting at a desktop wallpaper. You're just showing off a cherry picked screenshot to brag and it's misleading to people that don't know any better.

This is equivalent of me buying a brand new GTX 1080 and telling you I got 900 fps benchmarking it.........at the main menu of the game. What's the point? Want to show something relevant? Show your results during a gaming session, or during heavy video playback, browsing the web, listening to music, etc - when the drivers are really put to use:
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
post #2865 of 3535
Quote:
Originally Posted by gene-z View Post

You realize the application is measuring the time it takes to resolve the page fault? If it was unrelated to latency, it wouldn't even be in the application, lol. Why do you think it's measured in µs cycle time like the rest of the counters? Take some time and go read the LatencyMon website.

We're measuring DPC latency. That screenshot you quoted was pointing out DPC latency, not page faults (the page faults were 0, so why are we even talking about this? It's a mystery to me). Do you not understand that? Everything in your computer causes latency. DPC latency is a good measurement for overall system/kernel latency, and it can be measured via software like LatencyMon and DPC Latency Checker in Windows.
Edited by Trull - 6/6/16 at 10:33am
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
post #2866 of 3535
Quote:
Originally Posted by gene-z View Post

This is equivalent of me buying a brand new GTX 1080 and telling you I got 900 fps benchmarking it.........at the main menu of the game. What's the point? Want to show something relevant? Show your results during a gaming session, or during heavy video playback, browsing the web, listening to music, etc - when the drivers are really put to use:

So if another card gets a consistent 500 fps in a menu while the 1080 gets 900, assuming 100% load on both cards and same API, etc., there's no point in that data?

500 fps = 2 ms frame latency
900 fps = 1.1 ms frame latency -> Proves the 1080 is almost twice as fast as the other card, and that it isn't by 20%, in terms of the amount of frames it can display per second. We also know that there is an improvement in latency of 45%.

That is absolutely relevant data. Otherwise what you're looking at is average frame rates during random sections of gameplay and the likes (not saying you can't get consistent results that are just as relevant, if not more). It's good for consumers to look at graphs and see if they can get above whatever refresh rate they play at and that's about it. If the card accomplishes consistent full load under the menu then you're not taxing it any less than actual gameplay would, lol. It would have to be a pretty awesome menu, though (something like Battlefleet Gothic Armada), or some pretty awesome scalability.
Edited by Trull - 6/6/16 at 12:01pm
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
post #2867 of 3535
Quote:
Originally Posted by Trull View Post

We're measuring DPC latency. That screenshot you quoted was pointing out DPC latency, not page faults (the page faults were 0, so why are we even talking about this? It's a mystery to me). Do you not understand that? Everything in your computer causes latency. DPC latency is a good measurement for overall system/kernel latency, and it can be measured via software like LatencyMon and DPC Latency Checker in Windows.

Oh, so now you want to ignore that the application measures the time it takes to resolve hard page faults when we were just talking about it? You literally just said "Page faults are unrelated to latency. We use the tool to measure latency.", now you're telling me everything in your computer causes latency. So which it is? I suggest you go read the LatencyMon page and get a better grasp of what the tool does before you start calling people trolls.
Quote:
Originally Posted by Trull View Post

So if another card gets a consistent 500 fps in a menu while the 1080 gets 900, assuming 100% load on both cards and same API, etc., there's no point in that data?

500 fps = 2 ms frame latency
900 fps = 1.1 ms frame latency -> Proves the 1080 is almost twice as fast as the other card, and that it isn't by 20%, in terms of the amount of frames it can display per second. We also know that there is an improvement in latency of 45%.

That is absolutely relevant data. Otherwise what you're looking at is average frame rates during random sections of gameplay and the likes (not saying you can't get consistent results that are just as relevant, if not more). It's good for consumers to look at graphs and see if they can get above whatever refresh rate they play at and that's about it. If the card accomplishes consistent full load under the menu then you're not taxing it any less than actual gameplay would, lol. It would have to be a pretty awesome menu, though (something like Battlefleet Gothic Armada), or some pretty awesome scalability.

The fact that you even tried to breakdown a metaphor is laughable. The figure of speech was to show how absurd the way he is using the tool is.

Majority of people in this thread are measuring and showing off latency measured at an idle desktop, as if it's relevant. I get it if you want to provide a baseline, but there needs to also be shown the system under load. Posting one screenshot of your system idling with super low latency, as is said, is misleading to those that don't know better. It's completely unrealistic, unless, again, you sit and watch your wallpaper. You're not going to get those results he showed when you're using the system. IT SHOULD BE USED AS A BASELINE RESULT.

Example, I can also kill every possible background process and cherry pick a screenshot:



but in reality, when testing the system properly, these are real-world results and they get worse with more demanding games:

    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
post #2868 of 3535
Quote:
Originally Posted by gene-z View Post

Oh, so now you want to ignore that the application measures the time it takes to resolve hard page faults when we were just talking about it? You literally just said "Page faults are unrelated to latency. We use the tool to measure latency.", now you're telling me everything in your computer causes latency. So which it is? I suggest you go read the LatencyMon page and get a better grasp of what the tool does before you start calling people trolls.

I'm not the one pointing out page fault latency (technically it's the highest resolution time for a specific page fault but I'll give it a pass as potentially having an impact on overall latency under load situations) on a SS that literally shows 0 page faults, lol. A SS that was not trying to deceive anyone btw, like you seem to be suggesting. You're talking about page faults for no reason, most likely so that you can feel better about yourself:
Quote:
Originally Posted by gene-z View Post

This thread always give me a good laugh.
It's like several people pointing out how much gas their car consumes when running completely motionless, and then you come completely out of the blue and tell them: "yeah, that's great, but it's actually going 0 miles per hour, so gas consumption while the car isn't moving isn't relevant. What's relevant is the maximum time it takes for the gasoline to get pumped into the engine when the car is running at top speeds... because reading about it somewhere gave me an excuse to mention it for no reason". Whatever makes you happy, but don't expect anyone to jump on your nonsense ego-train.
Quote:
Originally Posted by gene-z View Post

The fact that you even tried to breakdown a metaphor is laughable. The figure of speech was to show how absurd the way he is using the tool is.

Majority of people in this thread are measuring and showing off latency measured at an idle desktop, as if it's relevant. I get it if you want to provide a baseline, but there needs to also be shown the system under load. Posting one screenshot of your system idling with super low latency, as is said, is misleading to those that don't know better. It's completely unrealistic, unless, again, you sit and watch your wallpaper. You're not going to get those results he showed when you're using the system. IT SHOULD BE USED AS A BASELINE RESULT.

Example, I can also kill every possible background process and cherry pick a screenshot:

but in reality, when testing the system properly, these are real-world results and they get worse with more demanding games:

So now fps on a game menu with a 1080, a card that's just been released, is a "metaphor". Weird. I thought it was a hypothetical case bordering on realistic, but what do I know. You know better. Obviously. It's so obvious that you had to come here to show us all how baseline DPC latency doesn't mean anything (despite the fact that it's the best software-based way of measuring system latency and despite the fact that measuring DPC latency at any workload relies on baseline values for accurate/consistent data), and how highest resolution times for specific page faults are what it's all about. Damn, son. Where ya find that brain?
Edited by Trull - 6/6/16 at 5:19pm
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
Gud
(15 items)
 
  
CPUMotherboardGraphicsRAM
Xeon E3-1240 v3 ASRock H87 Pro4 Sapphire HD 7970 GHz Vapor-X (comes w/ free noise) KHX1866C9D3 
Hard DriveCoolingOSMonitor
Samsung 840 Pro Noctua NH-D14 Windows 7 Home Basic Asus VN279QLB 
KeyboardPowerCaseMouse
Ozone Strike Battle (Browns) Corsair CS750M Antec P280 Logitech G100s 
Mouse PadAudio
Razer Goliathus Speed (old version) Creative SB X-Fi Fatal1ty 
  hide details  
Reply
post #2869 of 3535
Quote:
Originally Posted by Trull View Post

I'm not the one pointing out page fault latency (technically it's the highest resolution time for a specific page fault but I'll give it a pass as potentially having an impact on overall latency under load situations) on a SS that literally shows 0 page faults, lol. A SS that was not trying to deceive anyone btw, like you seem to be suggesting. You're talking about page faults for no reason, most likely so that you can feel better about yourself:
It's like several people pointing out how much gas their car consumes when running completely motionless, and then you come completely out of the blue and tell them: "yeah, that's great, but it's actually going 0 miles per hour, so gas consumption while the car isn't moving isn't relevant. What's relevant is the maximum time it takes for the gasoline to get pumped into the engine when the car is running at top speeds... because reading about it somewhere gave me an excuse to mention it for no reason". Whatever makes you happy, but don't expect anyone to jump on your nonsense ego-train.
So now fps on a game menu with a 1080, a card that's just been released, is a "metaphor". Weird. I thought it was a hypothetical case bordering on realistic, but what do I know. You know better. Obviously. It's so obvious that you had to come here to show us all how baseline DPC latency doesn't mean anything (despite the fact that it's the best software-based way of measuring system latency and despite the fact that measuring DPC latency relies on baseline values for accurate/consistent data), and how highest resolution times for specific page faults are what it's all about. Damn, son. Where ya find that brain?

This right here sums up the majority of people in this thread. Talking about gas and comparing it to latency now. This thread never fails to deliver a comical experience.

You're here just to argue and provide nothing relevant to the discussion besides calling people with opposing opinion trolls. My initial reply was to say 0 page faults and such low latency is misleading, because it's not an accurate representation of how the system handles latency, again, unless you're staring at a desktop doing nothing. It's a good baseline result, that's it. Thus the metaphor saying, it's like benchmarking a game at the main menu, you're not getting the full picture of how it performs. And you're now on my ignore list.
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5 2500k, 4.2ghz @ stock voltage asus gene-z evga 1060 sc gskill 2x4gb ddr3 1866mhz 
Hard DriveOSMonitorKeyboard
samsung ssd 830 256gb w10 pro x64 dell s2417dg g-sync @ 165hz pok3r mx silent 
MouseMouse PadAudioOther
zowie fk1 @ 400 dpi, 1000hz zowie ps-r, zowie camade fulla 2 dac, akg k7xx sony a5000, 20.1mp 
  hide details  
Reply
post #2870 of 3535
Arguing over a tangent.

I lost some brain cells reading that Trull. Really, I did.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel Motherboards
Overclock.net › Forums › Intel › Intel Motherboards › Gaming and mouse response BIOS optimization guide for modern PC hardware