I've read your post on overclock.net and i have to say, some motherboards such as GIGABITE had real problems regarding 3D performances dunno why. I noticed it when FarCry 2 came out a few years ago, i made test with multiple motherboards CPU and other hardware, the problem was coming from the GIGABITE motherboards. Dunno why.

Other than that you prove nothing with your thread and it looks realy just complet paranoid, while back in the days when i detected that problem with GIGABITE motherboards (was the P35 and or P45 series iirc) i could see it while ingame with fps decreasing drasticaly and very specific software benchmark for 3D.

DPC Latency checker will show flawless green line with or without all the craps you mentionned ON or OFF, just saying.

Give some evidences first...

edit : you do realise the latency in DPC is in 0.001 of a second? All you can feel is placebo. Also RAM latency numbers are in NANO seconds all you can feel there is placebo too.

Yes he realizes it and even says this:
"DPC Latency is mostly a measure of how ready, willing, and able a non-real time OS is to deal with requests from the user. It's not a measure of input lag, but will give you a good idea of possible problematic devices."

His screenshot of DPC is total bullshit btw, you gotta let it run a few minutes to get a result he took a screenshot within the first 20s. Showing everyone his 2 microseconds of latency and absolut maxiumum of 19.. come the fuck on :>

It doesnt change anything if you let it run more, my average is from 4-8, with some rare spikes to 30-40, and got one very rare spike 170, but I heard hdd sound with it also, much be antivirus checking things.
You just can clearly see that hpet or eist/c1e is turned off with it. With only hpet on its ~30, with all bs on its 50-100.

As an obsessive tweaker/optimizer myself, I must say you make a lot of ridiculous claims. I'm sure you mean well, and you've got the right idea, but you absolutely exaggerate some things.

Let's see for example, I'll quote: "all the dynamic power and speed adjustment modes they have added are latency monsters, easily noticeable by a B-grade gamer".
They are not noticeable by anyone, let alone a B-grade gamer as you put it. In fact they don't really do anything , other than what they're supposed to - save power comfortably when you're idling/browsing etc. (I do use Speedstep/C1 myself.)
HPET on or off also doesn't make any remotely noticeable difference either.

It can be important/useful to configure those settings properly, but most of them really don't do anything for you in practice.
So, you should stick to the basic things that, you know, ACTUALLY matter. A good example would be OC-ing a mouse to 500/1000 hz from 125. The difference there is very easy to both quantify and perceive, in game or even on desktop.

"HPET on or off also doesn't make any remotely noticeable difference either"

Most people that don't even play video games can notice a large difference with this on or off, like the one guy that posted in the thread that says it feels like you're running a wireless, bluetooth mouse with it turned on.

Honestly, something is wrong with your system if you can't notice a difference with this on or off. It just means you have a much larger latency source than what altering HPET affects.

Another guy posted in the thread saying it feels worse with HPET off, probably because the mouse feels a little quicker and he didn't actually want the mouse to feel quicker than it was with his current settings.

It's really not even up for debate if this setting makes a difference or not because there's already been hundreds of internet forum posts about this setting before and the majority of people notice a large difference. If there is any one setting that's completely impossible to be "placebo", this is the one.

You already said you use things like speedstep and C-states, so changing other settings without altering those is the equivalent of turning on a ceiling fan to put out a fire in your house.

HPET makes huge difference, made huge difference on my old pc and on new one too. Some people on other forums notice latency from it too. But its a little bit less noticeable on CRT for me.

You are the closest to real mouse optimisation, than anyone else. But still, not close enough. :)

I thinking to write topic about it, because it is long overdue and the technique is available for a few years now. But it will take a lot of time to do properly, because the staff is really of seniour programmer level, and I'm not seniour windows applications developer. :)

I reached a stage in windows tweaking where i realized 99% of it is either snake oil or gives .0001% better performance. I bet it's the same with all this bios stuff.

Calling it placebo is great, but when I upgraded from old system with less power saving to new one I noticed significant latency, was much harder to hit things on desktop, in-game aiming was floaty.
There was no thread by r0ach so I went and disabled EIST and C1E state that I seen in cpu overclocking thread, and huge input lag was gone.

You think top 10 duel players for exemple gives a fuck about that? Obvious answer is NO.


:>

Point is he is paranoid.

another "but pro's" argument
Its about comfort, I can see the difference clearly and I didnt like latency increasing things even on my radioactive CRT(latency goes to human bearable level, so its less annoyng). No one pays me to adapt to whatever bs pc's have, if I can with a simple tweak make it better.

N'est-ce pas.

Also I quickly checked his post, didn't really see anything clear. I'm not keen about that stuff, but if there is no way to say how the input lag really is, if it's a real deal, if it really matters, and what kind of user is going to care... I mean, great that you found these things, but there is no "conclusion" (xD just checked the bottom of the post)... Personally I cba because it's not clear, and I bet it's the same for many people :/.

Execute Disable Bit I disable it just because the odds of any negative effects for gaming, latency, or other issues outweigh the security benefit.

How can a simple bit in page table entries have any effect on performance?

It can't. This guy seems obsessed with 'lag' beyond reason.
I mean, I thought this was a guide, not a survey, right?
But his argumentation is like: "Well! Lot of ppl said this and this in my thread so it must be true!"
Even if it were true, I suspect the reason for that might be older/problematic systems, where some settings might cause issues. That's the only explanation I can think of.
However, personally I'm leaning towards the notion that ppl don't know what the fuck they talkin about. That's my rule of thumb and it usually works well. :)

Itīs called projection.

quote myself from another thread here:

"this guy is really mental, and worse part is, he made me mental too yesterday. :D

i got a new system since one week, also intel cpu and testet this http://www.humanbenchmark.com/tests/reactiontime/index.php
after a long time again just to notice that i couldn't go below 200ms
no matter how hard i tried, when this wasn't a real problem in the past.

so i tried everything he describes in his thread, including switching to my old graficcard(gtx275) just to be able to use an "inputlag less" driver that he recommends.

long story short:
tried the reactiontest in IE10 at last instead of firefox, and reached
the results i was expecting, also with all the evil settings enabled."

HAHAHA xD

I had consistent ~150ms results on old pc with ?60hz ps/2 ball mouse and ~ 85hz CRT + win xp

Now I cant test good with lcd as colors are shimmering and only possible do normal score on 1st try while my eyes dont hurt, especially in this test with red color.

what result you were expecting?
do you get same results with nvidia scaling on?

i was expecting the results i was getting when i did this test one year ago. :P

i get the same results with nvidia scaling on aswell, yes(still using native res ofc, which he says increases inputlag aswell).

wow a year ago, and how many ms you get now? I highly doubt this is good way to measure input lag, than aiming 180 degrees back and forth and clearly seeing if its sluggish or responsive.

I just tried again hpet+nvidia scaling. Randomly running with LG vs 6 nightmare bots(few bones) aiming was much slower and sluggish, got amazing 42-45% accuracy, with just nvidia scaling its 50-55%, with all disabled its 60-65%, all on late 2006 LCD, old blurry CRT gives +15-20%, but it was few months ago that I tested with it.
I even have better lg with scaling on some mice with bad gliding feet.
"Placebo" works for me, and a little lag works for shit mice, but it still feels sluggish.

having it like this(j4f): http://i.imgur.com/TFTjph6.jpg
and turned HPET on again...

everything the same as before, also in my warmup setup on dm6.

And for the 10th time, the vast majority of people can easily notice a change in mouse movement by turning HPET off or on. If you can't, then it's something wrong with either you or you computer.

rofl.

You're delusional.

You underestimate the power of placebo. Many people thought blackviper's service tweaks speeds up windows. But tests show otherwise. Many people stroke there dicks 1hr a day in hopes for it to become larger. But tests prove otherwise.

Yes... I use that reaction test. Just wanna confirm, I had the same experience.
They've recently changed something about it, not the functionality just the design. There's some added lag when using Firefox, not sure why but that wasn't the case before. However in IE it still works like it used to.

So, Roach, the question of all questions...

How much % did your LG aim go up after all those optimizations?

Can't wait for the reply :3

he doesn't play fps

Turned off HPET. Made no difference whatsoever.

That's normal :)

depends on exact OS and PC configuration, generally it shouldn't really make any difference since OS should be using invariant TSC instead of HPET(instead of LAPIC if you got hpet off) for almost everything anyway(unless you are using vista/unpatched win7)

r0ach try the newest nvidia drivers. http://www.geforce.com/drivers/results/70184 They were released earlier today I have tested the dpc latency and for me it is actually good. For me it stays around 6 to 20 . I would like to see what you think of them. The drivers before this made mine jump to 300+. There does seem to be some spikes but I don't know if that is my anti virus or what.