GMABooster. Give your a integrated graphics a tidbit more power

If you haven't heard already GMABooster will boost the performance of your netbook graphics. Go here to read about it http://www.engadget.com/2009/04/30/gmabooster-gives-your-netbooks-integrated-graphics-a-nudge/

Compatible with Windows, Linux and even OS X.

 

As always, Engadget is couple month behind of current news.
I tested this out a couple month ago. The performance of gma 950 is so bad that even with the overclock, games still runs at horrible FPS.

 

yea this wont help at all rele..

 

Hmm I'm going to try this out on my mini 9.
I don't play games on it but I watch a'lot of shows on hulu when I'm in bed with it and it often is laggy so maybe this will help

Thanks

 

It doesn't work like that. Overclocking your graphic card have absolutely 0 effect on flash based videos. You'd get much better result by overclocking your CPU since that's all of the rendering processes run. You can play 1080P high def video on your netbook with 1/4 of the current GPU processing power.

 

I'd like to see the 950 play 1080P high def video using only 1/4 of its power.

AFAIK, don't you need a X4500/ATI 3200 or the Nvidia ion (9200?) to play full HD video?

 

CoreCodec will do 1080p on a netbook.

 

Isnt this the same program that was proven bogus like 3 or 4 times and had several threads closed about it?

 

no its all cpu, you can have certian video card & codec combo's help the cpu by offloading some of the work to the hardware of the gpu. However with a good cpu you can have the most cruddy gpu.

the Atom does 720p without any issues as long as I am in XP, in Vista its a bit choppy on some video, but 1080p is too much for the atom.

 

With coreavc codec + Kmplayer I can do h.264 3-5Mbit/s BD-rip at 1080P. It's not too much for Atom. Though, playing straight from BD-rom at over 40Mbit/s will fail for sure though.

Since it's 100% cpu rendering, a 10 year old graphic card with 8MB of Video ram will handel 1080P resolution. As long as the graphic card is capable of displaying 1080P 2d images at 60FPS.. then you're all set. Graphic card from 10 years ago have less than 1/4 of the power of Intel IGP 950.

 

It works, however, its not FREE performance. It's not a FEATURE by getting your MOBILE part from 133MHz to 400MHz(the max using the "GMABooster"). It's OVERCLOCKING.

Take caution. You won't benefit much in most games because GMA 950 is not fillrate bound.

 

lol lol like i didnt know its overclocking and what overclocking is, im saying the mods have closed a thread about a program like this one several times and never once was anybody able to post proof that it actually worked.

I am pretty sure this is the same one again, we need proof it works not just claims.

If its not the same program, then so be it thats what I was asking.

 

It does work. It's not a software optimization but a overclocking of the core. But don't expect much more than 30-40% boost in real games when going from say 133MHz to 400MHz(which is a LOT more than the actual performance increase).

http://www.tomshardware.co.uk/forum/page-263635_15_0.html

 

I would agree with Kenpachi =)) And from the old threads I don't remember anyone getting any significant improvements

 

Well.. this one definetly works. I've even discussed with the developer about this program couple month ago when he was proposing his idea and rolling out the very 1st beta. This thing definetly works, but the amount of improvement may vary. For a few game, you get 50%-100% improvement and some games you get none. Nevertheless, this program still works.

 

50-100% improvement? I find that hard to believe. Is there are proof to these claims (benchmarks, etc)? Also, there have to be some drawbacks to using this, such as more heat, instability, power draw, etc.

 

Here's some benchmark or experiences from gmabooster users.
http://forum.eeeuser.com/viewtopic.php?id=60434&p=2

"Benchmark from Cystal Mark
1.6Ghz/GPU stock

Mark 25069
ALU 5349
FPU 4600
MEM 4519
HDD 5152
GDI 1840
D2D 2926
OGL 683

1.6Ghz/GPU 400Mhz

Mark 26068
ALU 5349
FPU 4600
MEM 4519
HDD 5152
GDI 1926
D2D 3842
OGL 680"




Basically this tool is to "un-underclock" the underclock IGP on the netbooks. There are tonnes of benchmark on regular GMA950 and you can compare them to the underclocked ones on netbooks. In some games, the FPS does in fact improve 50-100%, but that's rarely the case. Most games probably get about 1-5 more FPS. The underclocked GMA 950 is 166mhz and it can be safely overclocked to the standard 400mhz. That's almost 2.5x faster clock speed. Since overclocking performance isn't proportionate to the actual performance thus the speed improvements will be around 100-150% tops. In games, the difference will vary depending on the game engine and the edsign of the game.

Of course, there's going to be draw backs, mainly due to more heat and more power ussage. Since the IGP is downclocked and undervolted ( I think), the energy saved is exponential. Increasing the clock speed is just going to bring it back to standard clock speed thus it's unlikely to have stability issues.

This program definetly works, but it's not effective. Today's game relies too much on the CPU. Even if you have 1000% actual IGP performance increase, you'll still get faster speed by getting one of low end current generation card.

 

Not to mention the fact that since the GMA 950 doesn't have hardware T&L nore vertex shaders, all of that work is done on the CPU.
Overclock it all you like, but it's not going to make that big of a difference.

 

Please don't say that. A mobile part binned and set to run at 166Mhz is indeed being overclocked when its forced to run at 400MHz. It is an overclock. Overclock doesn't always require increased voltages either.

The only GMA950 capable of running at 400MHz default clock are the desktop versions. The 945G on the desktop has a TDP of 22W. The mobile 945GMS used in low power EEEPC/notebooks only use mere 4-5W.

That's like claiming ULV CPUs are desktop CPUs that are underclocked. They are not. They are specifically chosen as "ULV" and clocked that way.

Will it consume more power and run hotter? That doesn't even need to be asked. It's being clocked higher. It's the LEAST you'll get from clocking it higher. It's an overclock, not by little, but up to 200% increase. This is not your regular "software booster" where it might really work without any issues. Don't come back and complain if the chip dies later.

 

The chipset used in netbook is actually a different chipset, mobile i945GSE. As far as I know, many people, including myself, who helped test out the program since from its early beta stage didn't get any instability when the IGP is overclocked to 400mhz. I haven't heard a confirmed stability issue from using this software yet.

The example you're referring to is probably inaccurate in my opionion. You can't compare mobile cpu with desktop CPU.

A ULV cpu is actually a highier bin grade that standard cpu. In theory, if given the same voltage under the same condition, ULV cpu should be able to run at the speed of the standard cpu if not faster. So with this analogy, it's possible that the mobile chipset is a highier bin quality thus is able to run at 400mhz standard clock speed with undervolting.

Overclocking without increasing the voltage will rarely damage the IGP or CPU. The killer factor is the voltage used. Too much voltage will fry your CPU/IGP. In this case, I assume the IGP is brought up to standard clock speed with lower voltage thus it's not overclocking and it in theory won't damage the IGP.

 

You are promoting a controversial software based on the fact that its totally safe. See, that's not totally true. I've seen the forum, and one of the reasons stated for 400MHz not bringing too much advantage over 250MHz is because the GPU is throttling!

Even if you are totally right that it does not damage the chip, the form factor designs are usually based off the fact that its running on a 133MHz or a 166MHz variant of the IGP NOT a 400MHz that's only used in desktops.

And how much are you gaining for all that potential side issues in real games?? 10-20% gain for a 200% or a 3x increase in clock? Don't be mistaken, the GPU is not at all fillrate bound. In terms of memory bandwidth usage and fillrate its so good that the "next" generation architecture(X3100) falls behind by 30-40% in some games.

You are helping zero by not stating the caution you have to take. Yea, I'm pretty sure that's why the original "GMABooster" threads all got locked. Because it really brings negligible benefits for a potential much worse side effect.

 

I am saying is that it does overclock the GPU and the program works. The actual performance improvement you're going to get varies quite a bit with the games. Personally, I don't use to program since i don't usually play games on my netbook.If the program had a option to underclock the already underclocked GMA 950, then that would be a different story. All I need is to play 1080P videos on my netbook and GMA 950 is too powerful and too power hungry for the purpose.

 

GMABooster works for me! I use it since February. No overheating, no stability problems. Here are the scores in 'Fillrate Benchmark 2004' on my Atom based netbook:

 

Here are PCMark graphics tests (166MHz and 400MHz):

 

I also just installed this program hope it does work,im going to test it on a few games..........

 

Please stop posting synthetic benchmarks and start posting actual fps in games.
I've tried GMAbooster and even checked the settings it changes in the registers with wpcredit and compared it to Intel's datasheets.
400MHz at 1.05V is not possible. On an Atom equipped netbook GMAbooster changes clock speed from 166MHz to 200MHz by altering the F0 register of the graphics adapter, that's all.

Synthetic benchmarks are worth nothing. GMA950 is just crap and sits idle while the CPU does the work.

 

Is a similar tool available for Intel's 4500 mhd. I'd like to see for myself if it works or not.

 

Holy necro posting batman!

GMAbooster .... nobody even has really proven it works yet let alone mentioned a similar program for the 4500.

Try this

As I can see, there is no way to do it and if there was it would be little to no point (just lik the GMAbooster)

 

Is it really not proven to work? From what I can tell, the benchmarks show synthetic gains.

There's a difference between "does nothing" and having negligible real gains. I believe that GMAbooster falls under the second category.