*HP HDX DRAGON Owners Lounge, Part 1*

Yeah? Check this out...





That's nice, right?

 

And with powermizer set to off, they each switch frequency in real-time...

It's easy to make. It's just a toolbar added that's a directory named "nvMODE" with shortcuts to the profiles in it. The nVidia icon is from a nvidia DLL.

 

Added an Underclock profile: (lowest one can safely go before video distortion)

 

Perfect test to demonstrate the need to cut powermizer off.

Go to hulu.com and pick any TV show you like. Watch it in full screen and choose hi-res. You'll see the choice for hi-res after the brief commercial if you don't see it right away.

If it's jerky, stuttery, or something Land 'O Lakes wouldn't call butter, then you could benefit from turning powermizer off.

With powermizer off and fixing my GPU to 200/100/400 (core/mem/shdr) it's as smooth as... let's just say it does Land 'O Lakes proud.

It's the same for OTA HDTV feeds. Kills the jitter that's caused by other than poor reception.

 

I'm so glad you've finally understood me!
(And you've got it now finally understood correctly)
But a small step is still lacking.

Powermizer "on" -200/100/400 = 43C idle (because "Standard2D mode" ca: GPU voltage 0.8 V)
Powermizer "off" -200/100/400 = 48C idle (because "MaxPerformance3DMode" ca: GPU voltage 1.3 V)

This is just the temperature! (Voltage is not measured)
Powermizer "off" + idle:
1. 200/100/400 + 5C = 48C Fan Stage 1
2. 215/110/465 + 6C = 49C Fan Level 1 (Standard RickiBerlin after long testing phases, explanation follows *)
3. 500/799/1250 + ? C = 49C Fan Stage 2! (+? = Ca.20C if fan would remain at Level 1)

No3. 500/799/1250 of the corresponds with Powermizer "on", because the GPU is also yes in the "MaxPerformanc3DMode" works.

RickiBerlin do not like fan noise.

Fan Level 1 to 53C then level 2 is active. (cooling down to 47 then Level 1)

* (See above) from my standard clock (215/110/465 PM "off"), I 4C to 53C can be achieved. (49C +4 C = 53C)
This goes for light load, such as the Internet, Office, TV, BlueRay to Fan1 to stay. (also depending on outside temperature)

Have current GPU 52C at 23C room with sun on GPU page and write you. (Fan is on 1)

200/100/400 PM "off" vs.215/110/465 PM off:

QuickPlay (TV mode) crashed at 200Mhz sometimes distracting. Otherwise I had no problems with 107/106/220. At the same time, my standard clock a healthy mix of everything.

Bravo 2.0, have made it like me (profiles). I had 20 profiles to test, then 4 profiles (as you now) and now only 3 profiles.

Next Treat warnings ...

Ps: I edit my Treats up to 30 min after the send.

 

A real underclocking with 8800M GTS is not possible!

1. Powermizer "on" profile, for example 107/106/220:

"Powermizer automatic" when GPU load up to 107/106/220, down to 383/301/766 under 30s, under 15s down to 275/301/550, under 15s down to 200/100/400.
Max = 107/106/220
Idle = 200/100/400

That makes no sense!


2. Powermizer "off" profile, for example 107/106/220:

GPU works always in the "MaxPerformance3DMode" = more energy (CA1, 3V)

107/106/220 idle = 45C GPU Temp


Reminder: When Powermizer "on" 200/100/400 GPU is at 43C idle, because the 8800M GTS "Standard2D" works. (ca 0,8V)


Denote all levels:

200/100/400 = Standard2D Mode
275/301/550 = 3DLowPower Mode
383/301/766 = 3D Performance Mode
500/799/1250= 3D Max Performance Mode <-------When overclocking, or "downclocking" are always MHz only for the "3D Max Performance Mode" changed. (Never the voltage) It follows also 107/106/220 would be in the "3D Max Performance Mode" at 1.3 V.

further warnings follow ...

 

Anyone mentioned that the very new Nvidia 185.85 just came out a couple days ago?

 

Yep, 2 pages back. Lost in the shuffle...

 

Hi CJ,

I have 32bit Win7 RC and installing the same drivers I had on Vista worked without any problems.

 

@ 2.0

I said in my posts about 4x before "HW Monitor" warned.

Please check it out my last warning, understand the warning and then translate it into your language!

This is in Powermizer "off" extremely important.

Caution statement follows with an example (your oc value):

1. Powermizer "off"
2. HW monitor off ( and not in Autostart)
3. Windows system restarts
4. HW-monitor does not start!
5. Nvidia Profile "610/975/1525" box.
6. In idle, the temperature rises above 53C now (do not measure!)
7. At 53C, the fan in Level 2 and remains there. Please just follow the noise! Even after 20 min of the fans would not be in Stage 1, as the high performance GPU. (check not required!)
9. Now run "HW Monitor"!
10.Was happens in this moment with the fan?

So he goes with me (despite over Temp 53C) immediately in level 1
And no matter what I do now, the fan stays on level 1
Only a restart Windows will help.

How is it with you?
For me, it is with every "HW Monitor" version equal.
However, I have Vista32bit.
If you do not experience this problem arises, ok.
It could, however, in other cases.

This means, for example:

Pinballman, 500/799/1250, Powermizer "off", PlayGame, running in the background "Hardware Monitor". The temperature rose to about 85C, depending on the game-GPU requirement. (da fan remains in Level 1)

 

Running test...

Don't have that problem running Hwmonitor v1.10 and higher on vista x64. Using HwMonTray 1.5 to launch at start normally. Disbaled startup launch for test. Ran COD4 to get GPU temps up on overclock. Exited game. Fan on MAX. Ran HWmonitor. Fan dropped from high to med when temp from 55-> 53. Stays on med thereafter. To get it out of medium and into lowest fan setting, I have to switch to standard mode to drop temps below 49 or wait until gpu cools enough @ overclock doing nothing down to 49C.

Running any game or app that raises gpu temp >51C next fan level kicks in. Drops back to lowest fan level once 49C or below achieved.

Max temp w/ COD4 was 71C using overclock. Fan @ max as it was when powermizer is on. After game exit, clicked on standard profile(200/100/400) about 1min - 1.5 min, temp @ 48 fan drops to lowest setting.

Ambient room temp, 24C at time of tests.

Idle at 200/100/400 = 44/45C.

 

I think you have no HDX!

If something is time, you can perhaps try someone check Vista32bit. (for example Pinballman, or other?)
-I can not explain better in English.
If not, well. (yes I know the problem for me)

Question: Why did you Powermizer now "off"?

After your information (your HDX has no Microruckler), it actually makes no sense. On the contrary, the temperature is always higher than with the PM "on".


Info @ Hovercraftdriver: Thank you for Smiley and remember ...

Powermizer "on" or "off" switch just by rebooting OS!
(because registration needs to be reloaded)

 

LOL, secret is out!

Likely it's has to do with 32bit Vista. As both hDX's I have using Vista x64 respond the same way. The only differences between the two is the GDDR3 VRAM is Hynx and Samsung and the USB2.0 controller on one is faster than on the other.

Only way to understand what you were talking about was to experiment. Found that the problem occured with certain flash videos (FLV) like movie trailers in HD and on HULU Hi-res for certain shows.

Also, it benefits the community as others might need to do it and may need help along the way.

My next endeavor is to figure out a way to force Performance 2D on my lowest 2 modes.

 

Quote 2.0: "My next endeavor is to figure out a way to force Performance 2D on my lowest 2 modes."

If I understand this correctly, like you have a Nvidia profile in 2D mode. So at about 0.8 V?

Have fun!
This I have been trying for almost a year.
Nvidia does not answer me and I'm really angry.
If you have it, there are a beer in Berlin on my bill.
(Flight goes on your account)

Or did I misunderstand you?

 

That's correct.

LOL. Well, let's say that the flight is free but the beer in Berlin is worth $2000 eur.

If it can be figured out I'll be sure to post it. Might take another year...

 

The fear I have.
Have tried everything as it comes with what I einfiel.

1. Nvidia bios - no way!
2. Registration database - yes, yes, but where and how?
and, and

I'm with my skills at the end and live with my standard clock + PM "off".

Too stupid to a Powermizer shows it is not a hardware problem.
There are no documented Nvidia settings.

It could also, I buy the beer and send me a plane. Then I visit you in New York.

 

185.85 Blu-Ray ok

 

imo it's a big failure.
that's why I am going to prepare a lite version of vista with Vlite app to have less than 50 processes running.
Maybe I install this one on the second drive to use it only for games.

 

Sorry, I do not understand what this is consistent with my quote has to do.
There is but only the GPU temperature. (not for CPU)

 

it means that if you set PM off your gpu will always work on full voltage not o.8v as with PM on even if you set clocks manualy for very low level.

 

Yes, this is a statement from me.

But what does this have with your 50-Vista processes to be done?

 

Thank you to
2.0 & RickiBerlin

Between both of your post there has been a great deal of useful info that I used to customize my GPU settings. I now have PowerMizer off and created custom profiles that I can change in real time with a custom toolbar.

Here is the toolbar thanx to 2.0:



I am not sure why but I am now able to OC farther than 605,905,1450. Ever since I installed 185.85 I have been able to. It may be coincidence. I am able to use 610.975.1525 in most games and to finish 3Dmark with a score of 9735, 605,905,1450 is still the most stable with no crashing an any games, and 630,1007,1575 will not finish 3dMark with out crashing and only a few games are stable. All games test were with 30Min of playing so long term stability has yet to be seen. (Except 605,905,1450 with many hours nonstop)

Also here is some additional info from my experience for those who go with custom profiles.

When I edited my .nsu files here are what was in mine:

[BoardConfiguration]
Version=200
CPU=10676
MCP=65535
SPP=65534
GPU=NVIDIA GeForce 8800M GTS

[ClockSettings]
FSBMHZ=0
AGPMHZ=-1
PCIe2MHZ=-1
CpuCore0Mul=14.0
CpuCore1Mul=14.0
CpuCore2Mul=-1.0
CpuCore3Mul=-1.0
CpuCore4Mul=-1.0
CpuCore5Mul=-1.0
CpuCore6Mul=-1.0
CpuCore7Mul=-1.0
TRAS=0
TRCD=0
TRP=0
GPUCOREMHZ0=200
GPUMEMMHZ0=100
GPUSHADERMHZ0=400

[VoltageSettings]
CPUVOLTAGE=-1
NFORCEVOLTAGE=-1
MEMVOLTAGE=-1
AGPVOLTAGE=-1
AUXVOLTAGE=-1
CPUFAN=-1
AUXFAN1=-1
HTCPUSPPVOL=-1
HTSPPMCPVOL=-1
NFORCEFAN=-1
AUX1FAN=-1
AUX2FAN=-1
GPUFAN3D0=0

I did not have a [GPUSettings] heading or a Min Max setting. Just one setting for each under [ClockSettings]. The settings in bold are the only ones I changed. It Might be a different version of the tools.

Also it looks like Voltage and Fan setting can be changed here. I have not played with them yet. But there may be a way to set the voltage lower to emulate "2D" performance and lower the core temp.

Also something to consider is editing the osbootpf.nsu profile this is what your computer will default to after reboot. I set mine to 200.100.400. If you use a setting that is within the available range then before you close NVIDA Control Panel it will ask you if you want it set as default.

Now all we need is a way to do this for the CPU! :-D

 

so if you have your os optimized there shouldn't be any problems with PM and you will have the lowest temps

 

maybe clockgen

 

Looking good CompT. Looking good.

I noticed that to. Played around with the voltage one by setting it to zero and 1. Wanted to see if that un-greyed out the voltage setting in tools. No effect noticed.

 

Two things I observed about this:

I have 2 HDX's. One has Hynix RAM on the MXM board, the other Samsung. The one with Hynix doesn't have the same issues that one with Samsung has. The one with Hynix has slightly more processes, 55 -57, and doesn't experience and stutter except on some Hulu and CBS videos in hi res. It also clocks to 383/301 and stays there the entire time while watching video with powermizer enabled..

The HDX with the least processes has Samsung RAM and stutters with unbuffered FLV (flash videos) of all kinds with powermizer enabled. Also jumps the clocks all around then settles @ 200/100/400. With powermizer disabled, it's stutter free except on CBS FLV videos because the problem is on their end. Their stream speed fluctuates.

The two HDX's side by side, Hynix with PM on, Samsung with PM off, the Samsung vRAM based one does a better job of video playback on the same video source and reveals that the Hynix one has a slight stutter on video playback.

So it's one of those things where it depends on the individual setup. Much like overclocking or undervolting. YMMV.

The other thing about temps... a higher idle is actually marginally better than a lower one. This may seem unconventional at first and counter-intuitive. Allow me to explain. Heat a metal and cool it often enough it will deteriorate no matter what. The greater the degrees between hot and cold, the sooner it will break down. Just as running too hot (>90degC) will break it down on a molecular level. As with the Nvidia solder bump issue, the solder (a metal alloy with a low relative melting point that joins metals with higher melting points) breaks down not just because of high heat but because of the hot/cold cycle (DegC between hot and cool) is great.

So if you're idling in the 40's and your top is in the 70's, that's good for the life of the GPU. If you idle too high(>50DegC), you make the fans work more and generate more radiant heat that will affect the nearby circuitry over time.

This all assumes you use your computer for 6 to 8+ hours a day for non-gaming or video watching purposes.

In any event, the HDX doesn't ever get hot enough to worry about temps. Just keep your top end at or below 75DegC (80DgeC max) and your GPU will last longer than your screen probably would. The talk about idle temps when cutting off powermizer is all about fan noise. The threshold for the medium and audible fan setting kicks in >50DegC. You don't want to be browsing, typing or whatever and listening to that noise all day.

Oh, yeah and just don't run overclocked all the time. That's another issue besides just heat.

 

how to check if my gpu has hynix or samsung?

 

It's not possible.

 

Dissasembly of the HDX is the only way since the GPU doesn't report it to software like Sandra. Ram is labeled on the board.

 

Has anyone tried this little program? The thread is littered with favorable posts and atta-boy's, but I haven't played with it yet. I'm not entirely sure what all it does, but enables PM on/off with ease among other things.

http://forum.notebookreview.com/showpost.php?p=4698205&postcount=1

 

you can try this app
http://www.iobit.com/gamebooster.html

It closes processes when you want to play a game.
In my case it closed 6 processes.

 

IDT driver with SRS:
ftp://ftp.hp.com/pub/softpaq/sp41501-42000/sp41693.exe

Enhancements:
Improves audio playback performance.
Provides an improved sound experience.

 

Do you know if it works with HDX 9000? Please give the community a proper heads up.

 

Looks familiar, Smells familiar, Sounds familiar...could be:

Really
Unreliable
Moronically
Bad
Advice

Oh, and it's not for HDX, I checked.

 

How do you use this? Do you just click on the cmd file? Import the adm file in regedit? I only clicked the cmd file..but now I'm thinking I didn't do the right thing...:/

 

This is a good find. Nice and polished.

 

That's hysterically witty!



Someone should get a pool going...

 

Ok it wasn't just me thinking that.

 

Thanks...I kinda just advanced the thought you gave me several days ago.

 

It doesn't even use the technically correct codec: 92HD71 B7, instead of Stac9271.

The funny thing is, there are HP drivers that do use 9271, as I'm sure that you are aware. I have tried modding WRT.ini's and other affected files to get one to even function, so we can have a little more useful UI. I have spent about 20-25 hours since Pieter started his over on LV2go, on about 6 different driver sets.

The only thing is you have to change so much, and then find correlating entries in the Inf and other files. Plus, can't alter the dll's (or don't know how), and they are different for each driver. So no way of knowing if keeping them or swapping in 5607 files is helping.

It's a nightmare. Nothing like playing with nVidia files. Pieter said he spent months and almost got divorced to get things right on his Dell, but even the mod(s) he posted isn't meshing with other peoples' machine very nicely.

I'm not giving up, but there is a limit...I prefer to keep my sanity and my wife.

 

guys does anyone if the sound card in the HDX is any good for external 5.1 surround speakers (namely G51)

or will i be wasting money buying them?

 

I'm sure it's great. You can even do 7.1 on this baby. I'm a bit of an audiophile myself and have some very good headphones and headphone amps, and I have to say that the line output on the HDX is very clean and crisp.....much better than your typical laptop sound.

 

Believe me, my system is completely optimized. (+ No error no Warning in Event Viewer, reliability surveillance Index 10.00)
The video plays with 200/100/400 PM "on". Temperature everything ok. This is not the problem!
The disks are very fast. (I had a lot to do with 7200 and employs SSD and those plans quickly discarded)

7200 - u / min = loud (am very sensitive)
SSD - 1.Single Level Cell-too expensive for large capacity 2.Multi Level Cell - Error on pre-programmed time

Moreover, this problem is not with the hard drives, nothing to do with "Vista-Prozessen" and nothing with video buffering to be done.

I can unfortunately not all over again to explain. It is too complex.

I could find some in Germany, where this video the same triggers. It is the only video I know of. All other videos are running PM "on" smoothly. As I wrote before long, the clock after clock down to 200/100/400 never changed. (if the video in IE remains) This behavior of Powermizer wished me well in other situations. For example, in GTA. Unfortunately, however, after the trit Down Clock on 200/100/400 a thunderstorm on latency. The trigger for this, I suppose, in a non-adjusted voltage of Standard2DMode.Beziehungsweise the exact Powermizer switching threshold of the next cycle with a higher voltage. This may, at any HDX (GPU) vary.
The problem here that nobody has, I think not!


You see, only now has 2.0 understood what I wrote months ago. In the following time are still many lessons to follow, which in my letter on page 542 + -> already included.

For example:

1. The attempt of Extration and decoding of the Nvidia Bios. It is also possible, but it can no longer be played back.
(The problem here is not the decoding, but the encryption on the installation)

2. (and thus CompTrakkie started already ) The attempt, on the NSU files other parameters to change. This is all not possible because the bios is locked. (secure)

3. Then this is followed by further attempts Reg Nvidia's changing. ( "PowerMizerDefault", "PowerMizerEnable", "PowerMizerLevel", "PowermizerLevelAC", "PowerSaverHsync" and more)
All this leads unfortunately to no results for fixed "Standart2DMode"!

Thanks for trying me to help!

 

Not to thank!

Profile 0 makes no sense, as the temperature falls very little. (compared to 200/100/400) And it comes in software (for example QuickPlay) for errors.
But as I started at that time also
Had 20 Profile and have an infinite amount of testing.
Now I have only 3 Profile ( "minimal 215/110/465", "500/799/1250 normal" and "max oc)

NSU-files - no change from values other than clock possible (Nvidia Bios Lock)

osbootpf.nsu = normal profile without any special characteristic
Previous names: nvsutil.nsu

Related software settings profile and baseline.npe not really usable.

Why do you use PM "off"?
To test such as 2.0, right?

 

What are the benefits of OCing the gpu. Could you show your results from fraps (for example cod4)

 

Great question.

COD4, Warpig Level , street fight. 1680x1050, max everything except anisotropic filtering set to ~25%. Fraps 2.9.7 build 7771

Standard clock of 500/799/1250;

2009-05-11 07:58:38 - iw3sp
Frames: 7280 - Time: 167617ms - Avg: 43.432 - Min: 27 - Max: 84

Overclock of 610/975/1525;

2009-05-11 08:03:14 - iw3sp
Frames: 11523 - Time: 205236ms - Avg: 56.145 - Min: 30 - Max: 113


A 30% improvement in frame rates. Booyah!

Edit: Forgot to add, driver 185.85.

 

VERY NICE

PS
what temps after oc?

 

65C as a high but the ambient is 22.5C with HDX running only 30mins . I've seen temps go to a high of 71c with ambient @ 26C testing various games for two hours.

 

So the best deal then still is to use the SRS drivers?

 

Looks like it. The odds of someone getting both the bass and the SRS extensions to work in a driver for all HDXs are very slim. It may be possible for the limited number of HDX's that are able to use the 5929 driver.