G73/53SW Owners - CPU Throttling to 798mhz and how to fix it

Hi, I initially noticed this when I first got my G73SW-XA1 back in mid march. but didn't think anything of it until a poster had brought this up as some HP owners noticed that their Envy 17 SB is doing this to them. EDIT:This affects all G series models with a sandybridge cpu(G73SW/G74SX/G53SW/G53SX), so you will have to run the temporary fix until there is an official fix.

EDIT: You can also try my Unofficial bios HERE. It's sad we had to fix it ourselves, but its a more permanent fix than the throttlestop method. Simply flash, and forget all about any games throttling/possibly throttling.

THE FIX

ThrottleStop 4.00
Downloads | Tech|Inferno

Once you start it up, TS(throttlestop) changes whatever stops the throttling until you reboot.You don't need to do anything else, as monitoring mode only seems to still cause throttling to stop. So what I usually do, since I still prefer to start it manually, is add this to the throttlestop.ini file, ExitTime=5.This tells it to close 5 seconds after being open(so you can edit the number to however many seconds you like, lowest number you can go is 1). Alternatively, if you'd like to have it start/close automatically, after putting the exit time in the ini file. Just create a shortcut, and drop it in the "Start>>All Programs>>Startup" folder. This will allow it to start up and close without you ever having to click anything. :>

Also, there is no known downsides to using throttlestop, your CPU will still downclock properly when not in use, and it does not run hotter, nor have I found any conflicts with battery draining faster.

You do NOT have to worry about any overheating, even with throttling disabled I see no higher than a peak 81c on my CPU/GPU in a ambient room temp of 84-86f. So you really don't need to worry about that. However, the PSU does get hotter when throttling is disabled when there is significant load on CPU + GPU. So just watch for that, though it should be ok for the most part.



Confirmed Games that are affected by Throttling(click on the game to see throttling log of that game)

• Grand Theft Auto 4
• GRID
• StarCraft 2
• World of Warcraft
• Left 4 Dead 2
• Team Fortress 2
• Just Cause 2
• Skyrim (in 3D mode)
• Sonic Generations

Video of throttling in GTA IV

If you don't see a game that you've proven to have throttling through logs on the list, please post it so I can add it. Thanks!

<embed src="http://www.youtube.com/v/0JdRUHHn0UM?version=3&hl=en_US" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" height="349" width='425'>


The Problem is that the CPU drops to 798mhz intermittently during games that put significant load on GPU and CPU at the same time, causing stutter and low frame rates. Sometimes I've even noticed in logs that it's dropping to 1300mhz before hitting 798mhz and then it jumps back up to normal speeds of around 2594mhz. This can be tested/shown easily also by running Prime95 and Furmark at the same time, using CPUz, or my preferred clock checking/cpu load program is now RealTemp 3.67.

Start prime95 in 8 threads, and you will see temps rise, as well as the clocks will be around 2300mhz-2594mhz. Then fire up Furmark and you will see instantly of the CPU dropping to 798mz and stay there. You can grab the programs below to test if you like, but it's not really necessary to do so. However, I do recommend using RealTemp 3.67 with logs enabled , to log the cpu during a game that your not sure if throttling is happening in or not, this will confirm any suspicion and be 100% proof of it happening.


Prime95
64 bit: Prime95 (64bit) - 25.11 Download - EXTREME Overclocking

Furmark 1.9.0
(GPU Tool) FurMark 1.9.0: The Return of the Furry Donut - 3D Tech News, Pixel Hacking, Data Visualization and 3D Programming - Geeks3D.com

CPU-Z
32 and 64 bit: CPUID - System & hardware benchmark, monitoring, reporting

RealTemp 3.67(for logging and reading correct cpu clocks/multiplier)
|MG| Real Temp 3.67 Beta Download

(click to enlarge)

(this pic demonstrates the throttling when furmark and prime95 is being run at the same time.)

If you would like to read the HP thread about it, link is below.
http://forum.notebookreview.com/hp-envy-hdx/575817-hp-envy-17-sb-throttling-test.html





NOTES I've gathered so far:

• This is not a temperature problem, temps are perfectly well in range during test. However, intel could be throttling for laptops that possibly would have high temperatures with both running.
• This is not strictly from caused by nvidia as the HP Envy uses an ATI GPU, so this is either motherboard or strictly the CPU.
• Using a higher watt PSU has no affect on throttling.
• Using the program Throttlestop 3.00 stops the throttling, only need to run it once on boot, it doesnt even need to stay running for it to work.
• HP released a BIOS update fixing the throttling problem for Envy 16 gen2 owners on May 29th, 2011.
• Using realtemp 3.67 to log cpu during games has proven that throttling is happening outside of artificial test.
• GTA4 is the most affected by throttling, frames are around 10-25, but instantly jump to 30+, once throttlestop is opened.
• All G53/73SW models are affected by this throttling problem until Asus releases an update to fix it, most likely it will be a new bios.
• Ken Lee from GenTechPC has been great help, no one at Asus was really responding to me. I asked if Ken could talk to them, and now he is working with Asus to produce a fix. :>
• If you'd like to show support, use this link http://vip.asus.com/eservice/techserv.aspx to complain, let Asus know you do not appreciate the throttling!

(just as a side note, this thread was for testing the problem, but I've come to learn a lot since I first made it, so I cleaned up and deleted most of what I said at first to make it easier to read/understand, as well as talk about the fix a bit more, so the below posts for the first 3pages are in response what was previously said.)

 

Probably thermal throttling... what are the CPU temps? But anyway, CPU-Z is no use for monitoring the CPU speed anymore... use Throttlestop if you want to know what your CPU is doing exactly.

Higher PSU won't work in most cases since the power draw is not limited by the PSU but by the mobo. Do you know what your system consumes under load?

 

It's certainly nothing to do with temps. When running prime95 it shoots up to 74-78c, then backs down and stays steady around 70-72c because of the fans.

When starting furmark, temps instantly start dropping down into the 50s and high 40s which seems concur with the 798mhz speeds, since CPU still is stressed at 100% in task manager.

Thanks for the throttlestop link though, I had tried it, but I guess I had an older version. So. I will definitely check it this one.

And, no I can't say I know, maybe one day I'll get an electric meter as I have been wanting one. :>

I'll add the pic I have of it what it does.

 

Check throttlestop, or better make a log. See whether it's throttling or whether the multi goes down...

Temps are definitely fine...

 

I was thinking that maybe they should try the clock fix for overclocking.

bcdedit /set useplatformclock true

 

Same problem reported in reviews.

Asus G73SW-91182V Notebook im Test Spielen mit Tiefgang auf notebookjournal.de

Also with the new Medion X6813.
http://www.notebookcheck.com/Test-Medion-Erazer-X6813-Notebook.51973.0.html

You should read the comments there too.
I think it's a Bios problem or a protection system on new sandy's.

 

Does the 1FC toggle make any difference? See http://forum.notebookreview.com/asu...asus-g51j-x-cpu-throttling-investigation.html

 

Thanks for posting that tilla, that's kind of info I'm looking for to see just how widespread this is.

I'd still like some G73SW owners here to post their results in this thread to get some good documentation going for future references. However, I have been doing some further testing and have some interesting results.(see below)

I read through the thread and remember seeing something like that prior to making this thread. But I dunno if ThrottleStop 3001 enables it automatically or what. Because after running it just once(i did not touch anything in throttlestop, just using it to monitor the multi and cpu speed) and I do NOT get anymore throttling from running prime95 +furmark.

I can confirm this even after a good 30mins the CPU stays at normal speeds and multi stays between 22-26. So, I do a reboot, and just start up furmark, prime95 and HWINFO32 to monitor the temps.

It stays at a dead 797mhz and multi is 8x. The second I start up throttlestop, it jumps to 2594mhz(26x multi) and only drops to 2200mhz(22x multi) at its lowest. I can even close out throttlestop and restart the test multiple times. Yet no matter what, it stays at full speed like it should, as well as still speedstepping down when there is no cpu load.

I understand throttlestop has helped with the downclocking, but I did not know it would do it just by running the program, since I still haven't touched any settings within throttlestop.

Either way, with the CPU able to run at full speed with furmark running and it doesnt downclock for the whole time ive run it, it shows there is definitely no power issues and should make it possible for a driver to fix this problem if throttlestop can somehow cause it to stay at full speed just by running once.

I will continue to read more on throttlestop as I've really yet to understand the program at all. Now to just figure out what exactly throttlestop does so I can replicate it without it needing to run at least once on boot.

 

0x1fc is an undocumented MSR, it's a feature added to throttlestop after that thread brought it to light via an anonymous contribution on the g51j 0x124 bsod wiki. From what you posted I'm not clear on if this is the problem with your machine or not, however that's it in a nutshell.

I wouldn't run on battery with that - there were some reports that it was unstable (possibly ate some batteries) unless you are on ac power.

 

Thanks for explaining, I'm still unsure if this is the problem as well. But I just had a HP Envy 17 owner PM me and just running throttlestop as "monitor only", stopped his throttling as well. Though, his clock stayed at 1895mhz while mine goes anywhere from 2200mhz to 2400mhz while running Prime95 and Furmark at the same time. I didn't get a single drop lower than that in a 2hour run. However, my PSU did get hot and start to smell of burning a little bit. So I definitely don't want to be pushing it like that often. I think I will still invest in a 180w psu sometime or another.

His low clocks could have been due to high temps, he did mention his were at 87-88c, while mine stay at 75-80c

I have run it on battery since and it seems to be exactly the same. So thanks for the warning, I will still keep an eye out just in case.

The HP thread keeps growing and growing, but not many have truly tried the throttlestop idea yet.

 

It sounds like high power consumption is triggering the Performance Control register to be set to the minimum value which I think is 8 for your CPU. When ThrottleStop first starts up, it's probably resetting this register to the default maximum value.

MSR Tool

If you want to have a look at the PERF_CTL register then you can download my MSR Tool and in the MSR Number box at the bottom type in 0x199. After that click on the Read MSR button before you start running Prime95 and Furmark. After you get those programs running, click on the Read MSR button again and see if this register has dropped. Finally start up ThrottleStop and have another look at that register by clicking on ReadMSR one more time.

If this only needs to be taken care of once by ThrottleStop then you can add this to the ThrottleStop.ini configuration file.

ExitTime=10

This will allow ThrottleStop to exit about 10 seconds after it starts if you don't need it anymore. Minimum time period is 1.

 

Thanks for posting, here is my results, I know I PMed you too, but just to have it in this thread I guess.

Before anything is running, no TS yet.


Running Furmark + Prime95(still no TS)


And with TS, Furmark + Prime95

 

0x1D hex is 29 decimal which is the maximum turbo multiplier.

Were you monitoring your multiplier with CPU-Z or RealTemp during your second test, Running Furmark + Prime95(still no TS)?

 

Was using both during that test, though, even with cpuz and realtemp closed, and just using HWMonitor for temps I can see it has dropped because temps become significantly lower.

 

In the HP forum a few were pretty insistent that it did affect real world performance, I havent really seen or felt any personally, yet.

But I will definitely give that encoding vid a try, so thanks for the links. I'll try to post some results later today or tomorrow.

 

This throttling seems to be triggered by power consumption so it's possible that any app that works the CPU and GPU hard at the same time might cause it to drop to 800 MHz. It would be interesting if someone had a Kill-a-Watt meter to keep an eye on that to see if there is a direct connection between power consumption and throttling. In many Dell laptops this relationship was very obvious.

 

I was thinking about buying this to check, but wasnt sure how good it would be for this type of situation.

Amazon.com: P3 International P4460 Kill A Watt EZ Electricity Usage Monitor: Home Improvement

Any other not so expensive recommendations would be cool too. In the notebookjournal link @tilla posted, the reviewer said laptop wasnt over drawing power when it throttled on them. But apparently it could be from the PSU getting too hot(if i understood them correctly??)

Here is the page the quotes are from, translated. Google Translate


I saw that someone in the HP forums tried a higher watt PSU, yet throttling still persisted. So I'm going to guess me buying a 180psu(im on stock 150w now) wouldnt solve it either, though if anyone has a G73SW and a 180psu, if you could run the test with it, that would be great.

 

I used a kill-a-watt (normal version) to test the g51j throttling. It's probably the cheapest tool for the job. I deleted my youtube vids after we learned of the 1fc fix for that, but it was pretty obviously tied to power consumption. I might still have the vids somewhere if you need an example.

 

Ahah, guess maybe I will go ahead and get that one then. I've been wanting to monitor some other things anywho(just for the fun of it :>).

And thanks to reading a post by TailGunner in the G53SW owner lounge thread, GTA4 is the perfect test game for throttle, which i decided to log in RealTemp 3.67 and of course, I easily noticed the frames dropping.

Before running throttlestop.

Code:

 
  DATE     TIME     MHz   C0 C1 C2 C3 LOAD%
05/12/11 22:47:45 1446.77 48 50 43 43 16.5
05/12/11 22:47:50 2584.22 57 57 51 49 39.7
05/12/11 22:47:55 848.10 58 55 47 49 44.3
05/12/11 22:48:00 997.77 61 60 54 54 41.4
05/12/11 22:48:05 798.22 55 55 47 47 40.8
05/12/11 22:48:10 798.22 55 54 47 47 38.5
05/12/11 22:48:15 798.22 55 54 47 47 40.4
05/12/11 22:48:20 1546.54 62 62 56 56 59.6
05/12/11 22:48:25 897.99 55 55 47 47 50.3
05/12/11 22:48:30 798.22 55 54 46 47 37.1
05/12/11 22:48:35 798.22 54 55 46 46 38.2
05/12/11 22:48:40 2235.00 62 65 58 57 64.8
05/12/11 22:48:45 1446.77 55 55 45 47 58.7
05/12/11 22:48:50 798.22 55 54 46 46 36.7
05/12/11 22:48:55 2594.20 65 67 60 60 81.3
05/12/11 22:49:00 1546.54 60 58 51 54 66.7
05/12/11 22:49:05 2594.20 67 69 63 62 81.4
05/12/11 22:49:10 2594.20 69 70 64 63 82.5
05/12/11 22:49:15 2584.22 72 70 63 64 73.4
05/12/11 22:49:20 798.22 62 62 55 56 38.5

After running throttlestop.

Code:

  DATE     TIME     MHz   C0 C1 C2 C3 LOAD%
05/12/11 22:50:15 2624.12 62 58 51 52 15.3
05/12/11 22:50:20 2584.23 66 65 57 58 44.8
05/12/11 22:50:25 2594.20 70 69 64 61 70.1
05/12/11 22:50:30 2594.20 72 73 67 65 80.8
05/12/11 22:50:35 2594.20 74 74 68 67 68.0
05/12/11 22:50:40 2594.20 74 75 70 70 76.8
05/12/11 22:50:45 2594.20 75 76 70 70 68.9
05/12/11 22:50:50 2594.20 77 75 70 69 76.3
05/12/11 22:50:55 2594.20 77 76 71 70 80.9
05/12/11 22:51:00 2594.20 77 77 71 71 76.2
05/12/11 22:51:05 2594.20 78 76 71 70 78.0
05/12/11 22:51:10 2594.20 78 77 71 71 81.7
05/12/11 22:51:15 2594.20 77 76 71 69 72.0
05/12/11 22:51:20 2594.20 77 76 73 71 78.2
05/12/11 22:51:25 2594.20 78 75 72 70 75.8
05/12/11 22:51:30 2594.20 76 76 71 71 76.0
05/12/11 22:51:35 2584.22 73 70 65 63 50.2

FPS/Game was way smoother as well. So I guess this throttling thing does suck and happens outside of synthetic test.

 

When the throttling problems first started being discovered in the Dell laptops, the Dell reps made excuses that these were synthetic tests and they didn't represent real world use.

Your log files make it very obvious that this is a problem for real world apps too. Any application that works both the CPU and GPU has the potential to trigger this throttling. You might want to try logging while running a 3D rendering benchmark like Cinebench.

CINEBENCH 11.5

Intel CPUs are rated to run reliably right up to 100C. In your second log file it peaks at 78C which is well within the Intel spec so there is no need to throttle the CPU to protect the CPU.

Some manufacturers turn to throttling to protect the power delivery circuit if it has been under designed and they think it is not up to the job. Asus needs to be contacted and asked to provide a solution for this problem. I used to talk with Jean Jose Guerrero at Asus. He was able to provide me with special bios versions to correct the throttling problems in some of their P6T desktop boards. These problems only happened when doing some extreme overclocking but they worked hard to come up with a solution for enthusiasts. The problem you are having is just during normal use so hopefully they can get this sorted out.

Compared to much more expensive test equipment, in many situations, the Kill-a-Watt is not nearly as accurate. I have one and I find that even if it is not 100% accurate, it tends to be fairly consistent and is good enough to help understand power based throttling problems.

 

I contacted Brian@Asus from the ROGForums about it, and he was the one that said for me to do a full post in the G-section about it, also that he would forward it to the staff. I noticed you found the thread, thanks for posting in it, and if it's cool with you I may go ahead and edit in some of the things you mentioned in your post above since that is a very good explanation of what this problem most likely is.

I'm downloading Cinebench now, but I have been doing some testing in 3dmark06,vantage and 11. I logged the cpu clocks and it seems to be the same before running throttlestop and after, scores pretty much match as well. So I guess we can rule out any throttling in those apps.Though, once I get my logs all organized better and I've run more test while logging, I'll put them up for DL in case you or someone wants to look at them/compare.

I also still need to download that video and encode it, i had started it a while back, but i my inet has been wonky the past few days, so I havent had the chance to finish. Still lots of test for me to do, and I know I really should log not just CPU clock/temps/multi, but i should log from GPUz as well to find some more accurate data on just how much load on the GPU before CPU decides to throttle.

Thats good to know about the Kill-A-Watt meter, I figured it wouldnt be super accurate, but I guess considering I don't have to meter anything professionally and I dont have much money to spend for a good one. I know it can be good enough for situations like this. I've also gone ahead and ordered a 180w psu, Figured since it's at a good price, I might as well.
Amazon.com: Lite-On AC Power Adapter PA-1181-08, APA05US. 19V/9.5A, 180W

EDIT: K, I ran Cinebench 11.5 while running Furmark, before and after TS.

Before TS, CB + Furmark

Spoiler

Code:

  DATE     TIME     MHz   C0 C1 C2 C3 LOAD%
05/14/11 01:44:40 2683.96 61 60 47 46 29.6
05/14/11 01:44:41 2733.85 61 60 46 46 25.8
05/14/11 01:44:42 2733.85 61 60 48 46 25.6
05/14/11 01:44:43 2763.78 62 61 47 47 25.5
05/14/11 01:44:44 2713.89 61 61 48 47 25.5
05/14/11 01:44:45 2683.96 61 61 54 49 28.8
05/14/11 01:44:46 2604.14 62 62 50 56 37.6
05/14/11 01:44:47 1546.52 54 55 46 47 62.7
05/14/11 01:44:48 798.20 51 54 46 47 86.2
05/14/11 01:44:50 798.20 54 54 46 47 74.3
05/14/11 01:44:51 1047.64 56 57 48 48 92.3
05/14/11 01:44:52 798.20 53 54 46 47 100.0
05/14/11 01:44:53 798.20 53 54 45 47 100.0
05/14/11 01:44:54 798.20 54 53 46 46 100.0
05/14/11 01:44:55 848.09 52 55 46 46 100.0
05/14/11 01:44:56 798.20 53 54 46 46 100.0
05/14/11 01:44:57 798.20 52 54 45 46 100.0
05/14/11 01:44:58 798.20 53 54 45 45 100.0
05/14/11 01:44:59 1047.64 55 54 46 47 99.9
05/14/11 01:45:00 798.20 51 53 45 46 100.0
05/14/11 01:45:01 798.20 51 54 45 45 100.0
05/14/11 01:45:02 798.20 51 53 45 45 100.0
05/14/11 01:45:03 947.87 51 52 45 45 100.0
05/14/11 01:45:04 798.20 51 53 44 45 100.0
05/14/11 01:45:05 798.20 50 52 44 45 100.0
05/14/11 01:45:06 798.20 50 51 44 45 100.0
05/14/11 01:45:07 947.87 51 51 44 45 100.0
05/14/11 01:45:08 798.20 50 51 43 44 100.0
05/14/11 01:45:09 798.20 51 52 43 45 100.0
05/14/11 01:45:10 798.20 50 51 44 44 100.0
05/14/11 01:45:11 1297.08 52 55 46 46 99.9
05/14/11 01:45:12 798.20 51 51 44 45 100.0
05/14/11 01:45:13 798.20 50 51 43 43 100.0
05/14/11 01:45:14 798.20 49 51 44 45 100.0
05/14/11 01:45:15 947.87 50 52 44 45 100.0
05/14/11 01:45:16 798.20 50 51 44 45 100.0
05/14/11 01:45:17 798.20 50 50 44 43 100.0
05/14/11 01:45:18 798.20 50 51 43 43 100.0
05/14/11 01:45:19 947.87 50 51 42 45 100.0
05/14/11 01:45:20 798.20 51 52 43 43 100.0
05/14/11 01:45:21 798.20 50 51 43 43 100.0
05/14/11 01:45:22 798.20 50 51 43 43 100.0
05/14/11 01:45:23 897.98 50 51 43 43 100.0
05/14/11 01:45:24 798.20 50 51 44 44 100.0
05/14/11 01:45:25 798.20 50 51 43 44 100.0
05/14/11 01:45:26 798.20 50 50 42 44 100.0
05/14/11 01:45:27 1097.53 50 51 43 43 100.0
05/14/11 01:45:28 798.20 49 51 42 43 100.0
05/14/11 01:45:29 798.20 49 51 42 44 100.0
05/14/11 01:45:30 798.20 49 50 42 44 100.0
05/14/11 01:45:31 947.87 50 50 42 43 100.0
05/14/11 01:45:32 798.20 49 50 43 43 100.0
05/14/11 01:45:33 798.20 49 51 42 44 100.0
05/14/11 01:45:34 848.09 56 57 44 45 100.0
05/14/11 01:45:35 798.20 50 51 43 44 100.0
05/14/11 01:45:36 798.20 50 50 43 44 100.0
05/14/11 01:45:37 798.20 50 50 44 43 100.0
05/14/11 01:45:38 897.98 51 52 46 45 100.0
05/14/11 01:45:39 798.20 48 50 43 44 100.0
05/14/11 01:45:40 798.20 50 50 44 44 100.0
05/14/11 01:45:41 798.20 50 51 43 43 100.0
05/14/11 01:45:42 1147.42 60 60 53 53 100.0
05/14/11 01:45:43 1047.64 50 52 45 44 100.0
05/14/11 01:45:44 798.20 50 50 44 43 100.0
05/14/11 01:45:45 798.20 50 50 44 44 100.0
05/14/11 01:45:46 848.09 50 51 42 44 100.0
05/14/11 01:45:47 798.20 49 50 43 44 100.0
05/14/11 01:45:48 798.20 48 51 44 44 100.0
05/14/11 01:45:50 798.20 49 51 43 43 100.0
05/14/11 01:45:51 1047.64 50 53 44 45 100.0
05/14/11 01:45:52 798.20 49 50 44 44 100.0
05/14/11 01:45:53 798.20 49 50 43 43 100.0
05/14/11 01:45:54 798.20 50 51 44 43 100.0
05/14/11 01:45:55 1346.97 54 55 47 46 100.0
05/14/11 01:45:56 798.20 50 52 44 44 100.0
05/14/11 01:45:57 798.20 48 51 43 44 100.0
05/14/11 01:45:58 798.20 49 51 44 45 100.0
05/14/11 01:45:59 947.87 49 51 44 44 100.0
05/14/11 01:46:00 798.20 49 50 44 44 100.0
05/14/11 01:46:01 798.20 48 50 44 44 100.0
05/14/11 01:46:02 798.20 49 52 44 43 100.0
05/14/11 01:46:03 1047.64 50 52 45 44 100.0
05/14/11 01:46:04 798.20 50 51 44 44 100.0
05/14/11 01:46:05 798.20 50 51 44 44 100.0
05/14/11 01:46:06 798.20 48 51 44 44 100.0
05/14/11 01:46:07 848.09 50 51 43 45 100.0
05/14/11 01:46:08 798.20 50 51 42 44 100.0
05/14/11 01:46:09 798.20 50 51 43 44 100.0
05/14/11 01:46:10 798.20 50 50 43 43 100.0
05/14/11 01:46:11 2055.38 62 63 56 56 100.0
05/14/11 01:46:12 1047.64 51 54 45 46 100.0
05/14/11 01:46:13 798.20 51 52 45 46 100.0
05/14/11 01:46:14 798.20 51 52 44 45 100.0
05/14/11 01:46:15 897.98 50 52 44 45 100.0
05/14/11 01:46:16 798.20 50 51 44 44 100.0
05/14/11 01:46:17 798.20 50 51 44 45 100.0
05/14/11 01:46:18 798.20 51 51 44 45 100.0
05/14/11 01:46:19 1097.53 50 52 44 45 99.9
05/14/11 01:46:20 798.20 50 52 43 44 100.0
05/14/11 01:46:21 798.20 50 52 43 45 100.0
05/14/11 01:46:22 798.20 50 52 44 48 100.0
05/14/11 01:46:23 897.98 51 51 44 45 100.0
05/14/11 01:46:24 798.20 50 51 45 44 100.0
05/14/11 01:46:25 798.20 51 51 44 45 100.0
05/14/11 01:46:26 897.98 57 57 50 52 100.0
05/14/11 01:46:27 798.20 51 51 43 44 100.0
05/14/11 01:46:28 798.20 50 51 42 44 100.0
05/14/11 01:46:29 798.20 50 51 43 44 100.0
05/14/11 01:46:30 897.98 50 53 45 44 100.0
05/14/11 01:46:31 798.20 50 50 43 45 100.0
05/14/11 01:46:32 798.20 49 52 43 44 100.0
05/14/11 01:46:33 798.20 49 51 43 44 100.0
05/14/11 01:46:34 897.98 50 52 44 45 100.0
05/14/11 01:46:35 798.20 50 51 43 44 100.0
05/14/11 01:46:36 798.20 49 52 42 44 100.0
05/14/11 01:46:37 798.20 50 51 43 45 100.0
05/14/11 01:46:38 1147.42 54 55 45 46 100.0
05/14/11 01:46:39 798.20 49 51 43 45 100.0
05/14/11 01:46:40 798.20 50 51 43 45 100.0
05/14/11 01:46:41 798.20 50 51 44 44 100.0
05/14/11 01:46:42 897.98 50 52 44 44 100.0
05/14/11 01:46:43 798.20 50 51 42 44 100.0
05/14/11 01:46:44 798.20 49 51 43 43 100.0
05/14/11 01:46:45 798.20 50 52 43 44 100.0
05/14/11 01:46:46 947.87 50 52 44 44 100.0
05/14/11 01:46:47 798.20 50 50 43 44 100.0
05/14/11 01:46:48 798.20 50 50 43 44 100.0
05/14/11 01:46:49 798.20 50 51 43 44 100.0
05/14/11 01:46:50 947.87 50 51 43 44 100.0
05/14/11 01:46:51 798.20 49 51 43 44 100.0
05/14/11 01:46:52 798.20 48 51 43 43 100.0
05/14/11 01:46:53 798.20 49 50 42 44 100.0
05/14/11 01:46:54 947.87 50 51 44 43 100.0
05/14/11 01:46:55 798.20 50 50 44 44 100.0
05/14/11 01:46:56 798.20 50 51 43 44 100.0
05/14/11 01:46:57 798.20 49 50 43 43 100.0
05/14/11 01:46:58 947.87 50 50 43 43 100.0
05/14/11 01:46:59 798.20 50 50 44 44 100.0
05/14/11 01:47:00 798.20 49 50 43 43 100.0
05/14/11 01:47:01 798.20 50 51 43 44 100.0
05/14/11 01:47:02 897.98 49 51 43 44 100.0
05/14/11 01:47:03 798.20 50 51 43 44 100.0
05/14/11 01:47:05 798.20 50 50 42 44 100.0
05/14/11 01:47:06 798.20 49 51 42 43 100.0
05/14/11 01:47:07 1147.42 51 50 43 44 100.0
05/14/11 01:47:08 798.20 50 50 43 44 100.0
05/14/11 01:47:09 798.20 49 51 43 44 100.0
05/14/11 01:47:10 798.20 50 51 43 44 100.0
05/14/11 01:47:10 897.98 49 51 43 44 100.0
05/14/11 01:47:12 798.20 50 51 43 43 100.0
05/14/11 01:47:13 798.20 48 51 42 44 100.0
05/14/11 01:47:14 798.20 48 51 43 43 100.0
05/14/11 01:47:15 798.20 49 50 43 44 100.0
05/14/11 01:47:16 798.20 50 50 42 44 100.0
05/14/11 01:47:17 798.20 49 50 42 44 100.0
05/14/11 01:47:18 947.87 58 59 50 50 100.0
05/14/11 01:47:19 947.87 50 51 43 44 100.0
05/14/11 01:47:20 798.20 50 51 43 44 100.0
05/14/11 01:47:21 798.20 50 51 43 44 100.0
05/14/11 01:47:22 897.98 50 53 44 44 100.0
05/14/11 01:47:23 798.20 50 51 43 44 100.0
05/14/11 01:47:24 798.20 50 51 43 44 100.0
05/14/11 01:47:25 798.20 49 51 43 45 100.0
05/14/11 01:47:26 897.98 50 52 43 46 100.0
05/14/11 01:47:27 798.20 50 51 43 44 100.0
05/14/11 01:47:28 798.20 50 50 43 44 100.0
05/14/11 01:47:29 798.20 50 50 43 43 100.0
05/14/11 01:47:30 1097.53 52 53 45 45 100.0
05/14/11 01:47:31 798.20 51 50 42 44 100.0
05/14/11 01:47:32 798.20 49 51 43 44 100.0
05/14/11 01:47:33 798.20 49 50 43 43 100.0
05/14/11 01:47:34 947.87 50 51 43 45 100.0
05/14/11 01:47:35 798.20 49 50 42 45 100.0
05/14/11 01:47:36 798.20 49 50 42 43 100.0
05/14/11 01:47:37 798.20 49 50 42 44 100.0
05/14/11 01:47:38 897.98 49 50 42 44 100.0
05/14/11 01:47:39 798.20 50 51 43 43 100.0
05/14/11 01:47:40 798.20 49 50 43 44 100.0
05/14/11 01:47:41 798.20 49 51 43 44 100.0
05/14/11 01:47:42 1097.53 50 51 43 44 100.0
05/14/11 01:47:43 798.20 49 50 43 44 100.0
05/14/11 01:47:44 798.20 49 51 43 44 100.0
05/14/11 01:47:45 798.20 50 51 42 44 100.0
05/14/11 01:47:46 947.87 50 52 43 44 100.0
05/14/11 01:47:47 798.20 49 51 42 43 100.0
05/14/11 01:47:48 798.20 50 51 42 44 100.0
05/14/11 01:47:49 798.20 50 51 43 44 100.0
05/14/11 01:47:50 848.09 50 51 42 44 100.0
05/14/11 01:47:51 798.20 50 51 42 44 100.0
05/14/11 01:47:52 798.20 49 51 43 43 100.0
05/14/11 01:47:53 798.20 49 51 43 44 100.0
05/14/11 01:47:54 947.87 49 51 43 44 100.0
05/14/11 01:47:55 798.20 49 51 42 43 100.0
05/14/11 01:47:56 798.20 49 50 42 43 100.0
05/14/11 01:47:57 798.20 49 51 42 44 100.0
05/14/11 01:47:58 997.75 48 51 43 44 100.0
05/14/11 01:47:59 798.20 49 51 42 45 100.0
05/14/11 01:48:00 798.20 50 51 42 44 100.0
05/14/11 01:48:01 798.20 49 51 43 44 100.0
05/14/11 01:48:02 798.20 50 51 42 44 100.0
05/14/11 01:48:03 798.20 50 51 42 43 100.0
05/14/11 01:48:04 798.20 50 51 43 44 100.0
05/14/11 01:48:05 798.20 50 51 42 44 100.0
05/14/11 01:48:06 798.20 50 50 43 43 100.0
05/14/11 01:48:07 798.20 50 52 43 43 100.0
05/14/11 01:48:09 798.20 50 51 43 43 100.0
05/14/11 01:48:09 897.98 58 58 51 50 100.0
05/14/11 01:48:10 947.87 50 51 43 45 100.0
05/14/11 01:48:11 798.20 50 50 42 43 100.0
05/14/11 01:48:13 798.20 50 51 42 43 100.0
05/14/11 01:48:13 897.98 50 53 43 44 100.0
05/14/11 01:48:15 798.20 49 51 42 44 100.0
05/14/11 01:48:16 798.20 50 51 42 44 100.0
05/14/11 01:48:17 798.20 49 50 43 44 100.0
05/14/11 01:48:18 897.98 50 51 44 43 100.0
05/14/11 01:48:19 798.20 50 51 42 43 100.0
05/14/11 01:48:20 798.20 49 51 42 43 100.0
05/14/11 01:48:21 798.20 48 50 42 43 100.0
05/14/11 01:48:22 897.98 49 50 43 44 100.0
05/14/11 01:48:23 798.20 50 51 43 43 100.0
05/14/11 01:48:24 798.20 48 51 42 43 100.0
05/14/11 01:48:25 798.20 49 50 42 43 100.0
05/14/11 01:48:26 1097.53 50 52 44 45 100.0
05/14/11 01:48:27 798.20 48 51 43 43 100.0
05/14/11 01:48:28 798.20 48 51 43 43 100.0
05/14/11 01:48:29 798.20 49 51 43 43 100.0
05/14/11 01:48:30 897.98 50 52 44 44 100.0
05/14/11 01:48:31 798.20 48 51 42 43 100.0
05/14/11 01:48:32 798.20 49 51 43 43 100.0
05/14/11 01:48:33 798.20 48 52 43 43 100.0
05/14/11 01:48:34 848.09 48 52 43 43 100.0
05/14/11 01:48:35 798.20 48 51 43 43 100.0
05/14/11 01:48:36 798.20 49 50 42 43 100.0
05/14/11 01:48:37 798.20 48 50 43 44 100.0
05/14/11 01:48:38 1297.08 50 53 44 44 100.0
05/14/11 01:48:39 798.20 49 52 43 44 100.0
05/14/11 01:48:40 798.20 50 51 43 44 100.0
05/14/11 01:48:41 798.20 50 51 43 44 100.0
05/14/11 01:48:42 1047.64 50 51 44 43 99.9
05/14/11 01:48:43 798.20 50 50 43 43 100.0
05/14/11 01:48:44 798.20 48 51 43 43 100.0
05/14/11 01:48:45 798.20 49 51 44 44 100.0
05/14/11 01:48:46 897.98 50 52 44 44 100.0
05/14/11 01:48:47 798.20 50 51 43 43 100.0
05/14/11 01:48:48 798.20 50 50 43 43 100.0
05/14/11 01:48:49 798.20 50 50 42 43 100.0
05/14/11 01:48:50 1047.64 50 51 43 44 100.0
05/14/11 01:48:51 798.20 49 51 43 44 100.0
05/14/11 01:48:52 798.20 50 50 44 43 100.0
05/14/11 01:48:53 798.20 50 51 43 44 100.0
05/14/11 01:48:54 947.87 50 50 44 44 100.0
05/14/11 01:48:55 798.20 50 50 42 43 100.0
05/14/11 01:48:56 798.20 49 51 43 43 100.0
05/14/11 01:48:57 798.20 50 51 43 43 100.0
05/14/11 01:48:58 897.98 49 50 41 42 46.9
05/14/11 01:48:59 798.20 48 50 41 43 26.4
05/14/11 01:49:00 798.20 48 50 41 42 26.3
05/14/11 01:49:01 897.98 55 55 41 43 26.1
05/14/11 01:49:02 2753.80 59 59 45 44 25.6
05/14/11 01:49:03 2773.76 60 59 45 45 25.4
05/14/11 01:49:04 2773.76 60 59 45 45 25.3
05/14/11 01:49:05 2773.76 60 60 46 44 25.4
05/14/11 01:49:06 2773.76 60 60 46 45 25.3
05/14/11 01:49:07 2773.75 60 60 47 45 25.4
05/14/11 01:49:08 2763.78 61 60 46 45 25.4
05/14/11 01:49:09 2763.78 61 60 46 46 25.4
05/14/11 01:49:10 2773.76 60 60 47 46 25.3
05/14/11 01:49:11 2723.87 62 60 47 46 25.5
05/14/11 01:49:12 2773.76 62 61 48 46 25.4
05/14/11 01:49:13 2773.76 62 61 47 46 25.4
05/14/11 01:49:14 2763.78 62 62 48 47 25.3
05/14/11 01:49:15 2743.82 62 61 47 47 25.6

After TS, CB + Furmark

Spoiler

Code:

  DATE     TIME     MHz   C0 C1 C2 C3 LOAD%
05/14/11 01:04:10 2644.09 63 58 55 54 26.8
05/14/11 01:04:15 2624.14 63 60 55 52 26.7
05/14/11 01:04:20 2594.20 69 69 63 62 97.2
05/14/11 01:04:25 2594.20 71 72 66 65 100.0
05/14/11 01:04:30 2594.20 73 74 67 68 100.0
05/14/11 01:04:35 2594.20 75 75 70 70 100.0
05/14/11 01:04:40 2594.20 76 76 70 70 100.0
05/14/11 01:04:45 2594.20 76 76 71 71 100.0
05/14/11 01:04:50 2594.20 76 77 71 71 100.0
05/14/11 01:04:55 2594.20 77 76 72 70 100.0
05/14/11 01:05:00 2594.20 77 77 72 72 100.0
05/14/11 01:05:05 2594.20 77 77 71 72 100.0
05/14/11 01:05:10 2594.20 79 77 71 71 100.0
05/14/11 01:05:15 2594.20 78 78 70 72 100.0
05/14/11 01:05:20 2594.20 78 78 72 72 100.0
05/14/11 01:05:25 2494.43 77 77 71 71 100.0
05/14/11 01:05:30 2484.44 75 75 70 71 100.0
05/14/11 01:05:35 2474.47 75 76 69 71 99.9
05/14/11 01:05:40 2474.47 75 76 70 70 99.9
05/14/11 01:05:45 2474.47 75 76 70 70 99.9
05/14/11 01:05:50 2474.47 75 75 69 70 99.9
05/14/11 01:05:55 2634.11 70 65 63 60 26.8
05/14/11 01:06:00 2624.13 62 63 61 58 14.7

I had ran the After TS with 5sec intervals, but then realized I could prob see the drops better if I put it at 1second intervals, so after rebooting, I changed it to that for the Before TS test, and I think I will keep it like that for the rest of the time I do logs. Scores I get are 1.61 before, and 4.93 after just in case anyone is wondering. When there is no GPU load, whether TS is ran or not, I get 4.92-95, so scores really don't differ until GPU load is introduced.

 

dkillone: Good work with your testing. This should make it easy for Asus to duplicate this problem in the lab and hopefully come up with a fix for it.

The way Sandy Bridge CPUs are designed, after a preset amount of time when running at 100.0% load, usually between 28 to 56 seconds, the turbo TDP limit will decrease and this will decrease the amount of turbo boost so the multiplier and MHz will decrease. If you keep an eye on the power consumption number in ThrottleStop on the left hand side, you should see this number and the MHz drop at about the same time. You might have to run a steady load like Prime95 Small FFTs to see this happen. It looks like this might have happened near the end of your second test. This slight amount of throttling is normal. Throttling down to 798 MHz is definitely not normal.

You can copy and paste anything I've said anywhere you want if it helps you get this solved.

 

I confirm this, and I don't even need to run furmark for it to be an issue (though that did help in provoking/testing it). Thanks for the Throttlestop fix.

 

Ah, yeah indeed I have been finding more instances where throttling is occurring, I should edit my main post with what games I've seen/heard throttling happening. So far the list goes,

• Grand Theft Auto 4
• StarCraft 2
• GRID
• World of Warcraft

If you have anymore games or programs that you'd like to add, I'll put that in the main post as well.

K, I cleaned up the main post and changed the thread title to a more fitting name. Asus techs are supposed to be currently testing this problem and hopefully we will see a fix. But just gotta be patient and I wouldn't hold your breath. Though an awesome man(Ken Lee) will keep pushing them to get a fix out, so that is the good news of all this.

 

Having the throttling problem here too. Using Throttlestop even without turning it on (odd ) fixes it though!

Though it seems after a little bit of time it drops from 2.6Ghz down to ~2.2Ghz as well as the Wattage down to ~42W, so this is probably the normal 28 to 56 second TDP decrease that unclewebb mentioned above.

Thanks for the guide +rep

 

So to fix this problem I just need to run throttlestop in monitoring mode? I don't even need to turn it on?

Also, thanks for your work on this.

 

Yup, just in monitoring mode, though, you don't even need to leave it open at all. Once you start it up, it changes whatever stops the throttling until you reboot. So what I usually do, since I still prefer to start it manually, is run it, then close it right after it opens. I'm a low process freak. So while it isn't a very recourse intensive app, I still prefer to not have it on 24/7 since it just needs to be run once for it to work.:>

 

dkillone: If you don't need to keep ThrottleStop running after it sets up your CPU, just add this to the ThrottleStop.ini configuration file.

ExitTime=5

This option tells ThrottleStop to automatically exit 5 seconds after it starts.

 

Ahah, thanks, I knew you said it before, but kept forgetting to find the post to look it up. I will go ahead and add that to the main post for others in case they'd like to do the same.

Throttlestop is a great program, but since there isn't a way to stop it from stopping the throttling unless I reboot. I try to stick to RealTemp to log before and after so the logs are set up the same.

Which, have you made any progress on finding out what TS is doing to stop it? If I had some spare money, I'd seriously buy you a sandybridge laptop for you to play with. You certainly deserve it for all your hard work. :>

Thanks

 

any setting to run it automatically in minimize, and after 5secs, exit?
and auto turn on without i needed to press the option.

 

In the options menu there is a "start minimized" box you can check. As for the turn on button, you don't have to press it to stop throttling, just having TS in "Monitor Mode Only" stops any and all cpu throttling.

 

i didnt notice that, my bad. thankss

 

Has anyone heard if Asus is going to release a BIOS fix for this? There is no new BIOS downloads at Asus's site and you'd think they'd fix this since it can impede performance.

 

I'm working with Ken from GenTechPC, who is in turn working with Asus about it. Just did a video for them as further proof.


<embed src="http://www.youtube.com/v/0JdRUHHn0UM?version=3&hl=en_US" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width='425' height="349">

 

Dkillone - if you and Ken are good buddies with Asus, could you put in a word for an AES-NI fix as well?

 

I replied in a Pm, sorry it took me so long. Asus sure is being stubborn about this throttling problem. So who knows if we will even see an update at all. Hopefully we can have good luck with getting both solved in one clean update(prob too much to ask for though =P).

EDIT: @everyone else,

K, I found a nice convenient place to complain to Asus when reading through Skywise AES thread. If we ALL complain about BOTH problems, maybe we will see them update the bios and start taking us seriously about the throttling and adding AES.

http://vip.asus.com/eservice/techserv.aspx



You're welcome to copy and paste this if you don't feel like writing, should be simple, but yet effective enough/not individual sounding.

 

The reason why just running throttlestop as "monitor only", stop throttling is because, it unchecks the BD PROCHOT as soon as the programs is turned on which causes it to disable throttling. You dont even need to restart your computer to disable it. Just go in the option check off "Unlock bidirectional PROCHOT" and checkoff BD PROCHOT in the main screen. By doing this it will re-enable the throttle like before.

The BD PROCHOT actually controls the throttling
BD PROCHOT = uncheck| No Throttling
BD PROCHOT = checkoffed | Throttling

 

Wow, awesome, thanks. This indeed does bring the throttling back. However it is fluctuates, from 798.22 to 845-870mhz. Unlike without, it stays a strict 798.22mhz, no matter what. However, with that said, still cool that it indeed allows throttling since that is a step further into understanding what is going on. I'm sure unclewebb could explain in much more detail and could give me a few more things to test out. Maybe the suggestion of 0x1fc in the MSR thing ALLurGroceries suggested didn't do anything since the BD PROCHOT wasn't checked off by default, and I never thought to mess with it.

Thanks again, I will do some MSR Tool testing to see if I see changes when inputing 0x199 as well as 0x1fc. In the past I couldn't find any changes before and after running TS.

EDIT: K, the 0x199 still led to no changes. However, with 0x1fc, I get 0x0004005 F, and after starting up TS, I get 0x0004005 E. I never noticed it in the past, but I can confirm by toggling on/off the BD PROCHOT option. So indeed that must be the MSR throttlestop is altering by default that stops the throttling, even when in Monitor Mode. +rep to you good sir, Illz! :>

 

Intel does not document MSR 0x1FC bit[0] but that's the bit that ThrottleStop toggles on and off to solve this problem.

BD PROCHOT stands for Bi-Directional Processor Hot. This allows other items to signal the CPU to tell it that the CPU is hot and that it should throttle even when it isn't. It might be the GPU or the power regulator when it gets too hot or starts putting out too much power that signals the CPU to throttle and conserve power and cool down the overall system temperature.

A better idea would be for manufacturers to start designing adequate cooling solutions with a power supply beefy enough to power the CPU and GPU at their rated speed. This would let a user buy a gaming laptop and actually be able to play a modern off the shelf game like GTA4 without it trying to throttle to a crawl.

 

Good stuff DKillone! I will fire off an email of protest about this obvious BIOS related issue to Asus as soon as I am done here.

First, it's interesting that this only occurs in games that test the CPU and GPU simultaneously. I use my laptop for Folding@Home on both GPU and CPU at the SAME TIME, and have no ill effects: I am getting 2500Mhz on all Cores (8) for the CPU doing this deed, and the GPU is GPU CPU stressed to 99% constantly, forever literally unless I pull the two prograrms, FAH and FAHGPU for nVidia video cards.

So I don't understand why I am not getting throttling in the CPU, but I'm not! I am using Throttlestop (turned ON) to remedy the CPU problem, while I monitor the GPU with OC Scanner, an EVGA utility that members of their Forums can get, a most xlnt utility! I am also using MSI Afterburner to OC the GPU not EVGA's Precision utility. I find Afterburner to work better in both the desktop medium and the laptop than Precision, after much testing and years of doing FAHome on both types of computers, 24/7 365 days a year. I am one of the Top 25 contributors to the EVGA Folding Team, and have been ever since I started Folding for that team in Nov. 2010...My average points totals are today in the low 100K's, while when I have all PC's and Macs doing Folding my average is about 162K PPD (points per day), a commendable number of points. I use CPU-Z and GPU-Z also to monitor both VGA card and the CPU, which covers every possible monitoring situation so I know my data is good for the 2500Mhz consistenly on ALL Cores for the CPU, and for the dual CPU's in my three types of video cards, all EVGA: GTX-285's in SLI; GTX-560 Ti SC's in SLI; and one GTX-470 single GPU that folds all by itself these days...I will soon enough have another if a current RMA goes through fine on my last GTX-285 video card from EVGA, which has a Lifetime Warranty through the company. EVGA RMA Dept is replacing all RMA'd 200-series cards with 400-series and 500-series cards at the RMA Dept's discretion, a nice thing!

That's all I wanted to post about this phenomenon, and I DO know about GTA:IV being a victim of the throttling in my notebook, see it every time I fire up that game. I've just begun playing WOW on my laptop too, and haven't seen it there yet, but I will most likely. I don't play the others, presently anyway, but will in the future since it's quite a powerful laptop, given the chance to perform with Throttlestop engaged. I have been enjoying the 3D aspect of the laptop immensely, and perhaps my 120Hz LG display for 3D is a reason that I am not seeing as much throttling as the rest of you are, just a possibility I suppose--but could be real. I also feel that the 3D model is pretty much getting extra assembly and testing care and feeding at the China, factory where they are made, as I have not had a single issue with the components, overclocking both CPU and GPU significantly during game playing, while I leave the default clocks on the CPU and GPU both for Folding@Home work, and it IS work of a kind that is difficult to categorize, as it uses both available RAM, GPU and CPU processors 100%, and GPU memory RAM to some extent also, so go figure!?

If anyone has a need to know more about the issues I am having vs the rest of you, feel free to PM me here, as that can sometimes be a better medium for expressing things 100% true, blue too!

Thanks for reading this, all of you...

rexrzer727

 

is this the same issue ? but with screen blinking

width='425' height="349"><param name="movie" value="http://www.youtube.com/v/b7TXN6hIG_g?version=3&amp;hl=en_GB"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="http://www.youtube.com/v/b7TXN6hIG_g?version=3&amp;hl=en_GB" type="application/x-shockwave-flash" width='425' height="349" allowscriptaccess="always" allowfullscreen="true"></embed></object>

 

Drawing too much power too quickly out of a battery can damage it or shorten its life. That's why many laptop manufacturers use extreme CPU and GPU throttling when a laptop is running on battery power. There are ways to bypass this throttling but your battery run time will be short and your battery's life time might be quite short too. To prevent too many RMA claims for defective batteries, it is easier/cheaper for manufacturers to use throttling. Most gaming laptops are crippled like this when you pull the AC plug but few if any manufacturers are willing to admit this up front.

The G73/G53 series can also be crippled by throttling when plugged in. The first post in this thread shows a list of demanding games that can cause this. When you buy a gaming laptop with great specs but can't play games at a reasonable frame rate when plugged in due to throttling, that's a problem.

 

how would i go about bypassing this ? just so i can test it?

 

Hmm, I'm not seeing drops like in your vids, however my CPU does stay a permanent 798mhz when on battery? Have you used RealTemp/GPUz to watch your clocks while on battery? Maybe since my CPU isn't fluctuating, even though I am getting lower performance, it's still smooth.

EDIT BoxKid, have you tried uninstalling power4gear/running at different power profiles?

 

@dkillone I've tried all the power modes in P4G and i have just tried uninstalling P4G and no change at all... id rather fix it than RMA as i wont have a computer for a while

can you play games like dirt 3 crysis 2 on battery fine? what fps do you get because im quite sure the first day i had it i didnt have this problem as i think i recall playing dirt 3 on my battery fine...

 

do you get screen blinking like i do ?
i even get it while watching movies on battery

do you think i should RMA it to fix the problem

 

@Unclewebb: Could this issue be caused by the system drawing too much from the battery, and throttling to protect itself? Or could it be an issue with NVIDIA driver on battery mode? (aka PowerMizer)

 

If BoxKiD ran a RealTemp and GPU-Z log file while gaming, we would probably see PowerMizer lowering the GPU clocks and memory speed and RealTemp would show the CPU throttling down to 798 MHz. It's likely a combination of both the CPU and GPU throttling that is causing the screen blinking. CPU throttling is usually triggered by the bios when switching to battery power to protect the battery. I think you can adjust PowerMizer to disable the GPU downclocking when switching to battery power but the battery might not like that long term.

RealTemp 3.67
http://www.mediafire.com/?jibd769x6z7n0c8

 

Sorry for the long delay in replying, but I don't have those games, so can't exactly test. However, nope, I don't get screen blinking at all when running on battery. Doesn't matter what I'm doing, I've never seen the screen blink like in your youtube video.

Hopefully you don't have to RMA to fix though, would be annoying :\.

 

I have an ENVY 17 3D and have been searching for why my games stutter. I had one of the first Quad core DV7s that were recalled and never had this problem. I have been thinking it was a graphics switching problem with the ATI and knew that if I called HP it would be a dead end.

I just ran the test as described in the OP and it behaves exactly as described; 798. When I run ThrottleStop I see the clocks back up where they should be.

So is there any definitive answer to what the problem is yet? Do you know if HP is working on a fix as well? Is this an OS, Intel, or GPU problem?