I already posted this in the Nexus dev forum, and I hate to clutter the boards, but I know that there are a bunch of talented devs here that may not see it on those forums. If we can bust the 30fps cap for good, and get the GPU overclocked, then we could see some serious gaming performance out of our Snapdragon devices.
I'm sure we all heard about being able to overclock the GPU on some of the old MSM devices, but the Snapdragon handles the graphics chip in a different way. The goal of this thread is to try and overclock the GPU on Snapdragon devices as well.
There is nothing GPU related in acpuclock-scorpion.c (the Snapdragon cpu clock settings) at least for setting gpu clock speed as far as I can tell.
In board-mahimahi.c (Nexus board file) and board-supersonic.c (Evo board file) there is some kgsl init code, but so far as I can tell it isn't setting the clock there, instead it seems to be pointing to PWR_RAIL_GRP_CLK to set the clock in both devices. It defines the variable in each of those files but I'm not sure where that variable is set, since it doesn't seem to be in any of the other board files as far as I can tell. I could be completely off here too though.
In drivers/video/msm/gpu/kgsl/kgsl.c there is a method called kgsl_clk_enable that seems to be called whenever the GPU is enabled. It looks like this:
Code:
/* the hw and clk enable/disable funcs must be either called from softirq or
* with mutex held */
static void kgsl_clk_enable(void)
{
//clk_set_rate(kgsl_driver.ebi1_clk, 128000000);
clk_set_rate(kgsl_driver.ebi1_clk, 245000000); // Looks like it sets the GPU clock, right? Wrong.
clk_enable(kgsl_driver.imem_clk); // Enable the clock
clk_enable(kgsl_driver.grp_clk); // Enable another clock, but why?
}
The line that's been commented out is the original value, I replaced it with my value on the line below it in a failed attempt to overclock it. Probably a stupid effort on my part, I doubt it's that simple, but it was worth a shot. The comments at the end of the line are also my additions.
According to the clk.h files in the standard linux kernel, clk_set_rate is obviously a method to set the clock rate. The first variable is a struct that tells it which clock to set, and the second variable is a long value that is the rate you want it set at. Is it setting the right clock there for Snapdragon chips? Or is it only the clock for older chips?
I'm in way over my head with this source, I'm but a lowly Java dev, but I really wanna solve this. Can anyone with a little more experience than me throw in a little more info? Sorry if it doesn't make much sense, if it isn't clear just ask me & I'll try to explain a little more.
Regards,
Jesse C.
EDIT: Okay I did a little more digging and those kgsl settings should work for QSD8x50 chips. In the config file, under Drivers, Graphics Support, it allows you to enable 3D accelleration for QSD8x50 & MSM7x27 chips. The tag for that is CONFIG_MSM_KGSL_MMU. If you check in kgsl.c it checks to see if that is enabled in the config, and if it is then it compiles and uses kgsl.c & all of the kgsl code. That tells me that the clock is either not being set, or the wrong clock is being set. I'm adding some debug code right now so I can see in dmesg what code is actually being run.
Dude I have no idea how to do this but bump and best of luck
Sent from my PC36100 using XDA App
Isn't this Helping ? http://androidhtc.git.sourceforge.net/git/gitweb.cgi?p=androidhtc/kernel.git;a=blobdiff;f=arch/arm/mach-msm/clock-7x00.c;h=ef178abfcd46cf78dd47962d75298691f887ebf2;hp=d68fea09d1996daeff99365ced9785d65b1cb001;hb=23ff83048726252bc785699fc749a3e364a3bdb0;hpb=110e73c591db3fd23d8558659f8a6a0dfe5ba912
It looks like here he sets the clock writel(grp,MSM_CLK_CTL_BASE+0x84); 0x84
And here is the Orginal topic http://forum.xda-developers.com/showthread.php?t=697673..
fstluxe said:
Isn't this Helping ? http://androidhtc.git.sourceforge.net/git/gitweb.cgi?p=androidhtc/kernel.git;a=blobdiff;f=arch/arm/mach-msm/clock-7x00.c;h=ef178abfcd46cf78dd47962d75298691f887ebf2;hp=d68fea09d1996daeff99365ced9785d65b1cb001;hb=23ff83048726252bc785699fc749a3e364a3bdb0;hpb=110e73c591db3fd23d8558659f8a6a0dfe5ba912
It looks like here he sets the clock writel(grp,MSM_CLK_CTL_BASE+0x84); 0x84
And here is the Orginal topic http://forum.xda-developers.com/showthread.php?t=697673..
Click to expand...
Click to collapse
Yeah I saw that, but that won't work for us, because it is for one of the old MSM7500 chips instead of the new QSD chips we have in Snapdragon devices.
I am working on a few things now and I recommend you check the Nexus forum since that thread is a lot more active.
Geniusdog254 said:
Yeah I saw that, but that won't work for us, because it is for one of the old MSM7500 chips instead of the new QSD chips we have in Snapdragon devices.
I am working on a few things now and I recommend you check the Nexus forum since that thread is a lot more active.
Click to expand...
Click to collapse
okey, But as far as I know that code is not standard... Becouse the Vugue msm7500 device is originally WM device.. so the Smart dzo managed this custom code.
I wil follow the nexus form .. This is Damn interesting !
i wish you guys the best of luck..wish i could help..i love xda..lol
I'm going to bump this, would be nice if more people were working on this.
Man... I absolutely love my phone. It really sux that this even needs to be asked. This device is so much more capable and I cannot understand why HTC would cripple it so badly. Look at the Samsung galaxy s or the Droid x. Or even a more similar phone like the nexus or incredible. They all stomp this phone when it comes to anything graphics wise even if it is just swiping home screens. It makes me sad. Lol
I love how it makes you sad, but you still lol. Lol (correctly placed)
This is big but first let's get a true cap fix!
Sent from my Evo 4G
How is there such little interest in this subject? This intergrated gpu will probably overclock like a beast... and maybe we can get actual fps.
jigglywiggly said:
How is there such little interest in this subject? This intergrated gpu will probably overclock like a beast... and maybe we can get actual fps.
Click to expand...
Click to collapse
I'm thinking this is also one of those things awaiting source. Even the overclocked kernals make my phone reset so I'm waiting for a more stable release myself. Since you're still posting here does that mean you decided to keep your evo?
psych2l said:
I'm thinking this is also one of those things awaiting source. Even the overclocked kernals make my phone reset so I'm waiting for a more stable release myself. Since you're still posting here does that mean you decided to keep your evo?
Click to expand...
Click to collapse
I don't know, yet I am in deep havoc lol. I activated my phone on the 20th. I have till 13th of July to pay my bill, so don't know.
I mean the locked bootloader of the Droid X is kind of lame. Also on the Evo, I have a 10$ discount because of the IO conference... unlimited texts too, and I bought two batteries for it with a charger(for 10$ lol, and they actually last a long time 1500ma, I put the 2 extras in my wallet)
Also I think I can upgrade my phone right away already if I read things right, so if Sprint gets an awesome phone I can upgrade to that whenever.
Verizon is more expensive, but I like how they let me do a 1 year contract. The Droid X is also nicer, I mean it has a better camera, and the mic quality when taking videos isn't piss, hopefully someone fixes that. It also has a real graphics card.
So yeah, it's hard to decide. A better phone, vs saving me some money. Though I plan to attend the IO conference once again. I'm still porting my distrubuted prime client onto Android, hopefully it will be high performance. So it's not like I just go there for a phone ^^. Hell I didn't even know about the phones till they said "Where do you want the Droid delivered" I lol'd. Though my time is limited, summer classes at UC Berkeley are tough : /
jigglywiggly said:
How is there such little interest in this subject? This intergrated gpu will probably overclock like a beast... and maybe we can get actual fps.
Click to expand...
Click to collapse
Yep, looks like til we get the kernel in our hands they're tied.
Also what's the point of overclocking if their is still a limit on FPS..
Eventually this will be the next thing on our todo list
some of the source for the OpenGL ES GPU core was announced as being released today
Qualcomm releases open-source 3D Snapdragon driver
http://www.androidcentral.com/qualcomm-releases-open-source-3d-snapdragon-driver
havent looked at it yet, but im sure in the very least this should at least help shed some more light on overclocking the GPU and add another piece to the puzzle!
git location -
https://www.codeaurora.org/gitweb/q...b819424af4be;hb=refs/heads/android-msm-2.6.32
topdnbass said:
Yep, looks like til we get the kernel in our hands they're tied.
Also what's the point of overclocking if their is still a limit on FPS..
Eventually this will be the next thing on our todo list
Click to expand...
Click to collapse
There isn't really a fps cap with the novatek kernel hack. It made the display much smoother for me and I am using it on EvolutionX. So beast.
Good news about the drivers, maybe we can actually see some overclocking now.
I want this to happen!
Nova runs pretty well on the Evo. Gameloft updated it recently.
I would just like for iphone ports to run well.
I hear the Palm Pixi has the same GPU. It is clocked at 190mhz. What about the Evo?
patelkedar91 said:
I hear the Palm Pixi has the same GPU. It is clocked at 190mhz. What about the Evo?
Click to expand...
Click to collapse
The Pixi doesn't have the same GPU, it has one of the low/mid-range MSM7227 (I think thats the model?) chips. It's similar as it's one of Qualcomms latest gen chips, but it is NOT a Snapdragon so it's different.
Also, the Snapdragon GPU's on HTC devices are clocked as follows:
GRP_CLK: The actual graphics clock, 256mhz
IMEM_CLK: The gpu memory clock, 256mhz
EBI1_CLK: As best I can tell, the bus that the GPU is on is EBI1, this is the clock for it, 128mhz
If you want more info, look here: http://forum.xda-developers.com/showthread.php?t=710850 since that thread is a lot more active, but please don't clutter it, there's actual devving going on there
Thanks for the info. I just wonder how they're going to get 3d games to work on the Pixi then... weird.
It was announced that the Pixi would get 3d gaming with the 1.4.5 update.
Much better (Y)
http://forum.xda-developers.com/and...ck-overclocking-tuning-snapdragon-s1-t2883708
Related
hey guys,
I saw winmo devs OCing this same Magic's CPU up to 800Mhz with great performance boost.
any one could reach same with android OS so far?
URL: http://wmpoweruser.com/?p=7461
thnx!
first off, the touch pro 2 isn't a magic. at all.
second of all, this is completely wrong forum. please read rules.
third, it probably never will happen. someone did it a while back and there was a huge debate about whether it worked or not and it was finally confirmed that it didn't work.
I can't find the word magic anywhere in there!!! Damnitt!!
so it leads to this question,
How did you come to the assumption that you could OC the magic when nothing about the magic is there?
The cpu is the same on the magic...
Deviser said:
The cpu is the same on the magic...
Click to expand...
Click to collapse
yup thats why I've asked...thanks!
Code:
http://forum.xda-developers.com/showthread.php?t=630450
Rugga said:
Code:
http://forum.xda-developers.com/showthread.php?t=630450
Click to expand...
Click to collapse
It's a shame that project died. I really believe that it's possible if the right person would try cause it already worked on a lot of phones. I used to clock my Omnia at 800Mhz which gave it a great boost and there was no real battery drain. They've also managed to clock the N900 cpu at 1Ghz:
http://talk.maemo.org/showthread.php?t=39753
renegadEEE said:
It's a shame that project died. I really believe that it's possible if the right person would try cause it already worked on a lot of phones. I used to clock my Omnia at 800Mhz which gave it a great boost and there was no real battery drain. They've also managed to clock the N900 cpu at 1Ghz:
http://talk.maemo.org/showthread.php?t=39753
Click to expand...
Click to collapse
indeed sad :/ could give us nexus cpu power on magic if some one could work this out....
rommark said:
indeed sad :/ could give us nexus cpu power on magic if some one could work this out....
Click to expand...
Click to collapse
No, it wouldn't. You're buying into the megahertz myth. The magic has a ARM11 (ARMv6) chip, whereas the N1 has a Cortex (ARMv7). It's more like a P3 against a P4 - the clockspeeds may match, but the performance is still going to be off. Not to mentium ARMv7 has a bunch of new instructions (NEON), that are not present on the v6. And the GPU still wouldn't support OpenGL ES 2.0. It may speed things up sure, but it wouldn't really matter.
sadly it's not the cpu preventing me from overclocking it! it's the software in the phone, the pll's are defined in the spl and it would take i SIGNIFICANT skill of hexing and backporting to do something about them! and by the way it would almost certainly brick the phone! i'm sorry guys the only thing i ever was able to play a litlle with was some parameters giving a little better gpu performance, but it's unstable
Right so I posted this in both the kernel thread and the overclock thread but nobody saw it so I'll just post it again:
While looking at the vogue forum, I found a thread about GPU overclocking : http://forum.xda-developers.com/showthread.php?t=697673 the changes made to the kernel to enable this are the following: http://git.linuxtogo.org/?p=groups/...ff;h=23ff83048726252bc785699fc749a3e364a3bdb0
Unfortunately I am no programmer and I do not see how to incorporate this code in to kernel sources for MSM7xxxA processors, though it should probably work because the gpus of A and non-A processors are pretty much the same (same basic specs such as number of triangles etc.), so if a kernel dev. could try to integrate this patch, that would be awesome.
this really looks like some awesome stuff! i hope someone can figure this out for the g1. thanks for supporting the g1 thread!
GPU overclocking
Holy crap, a dev should take a look at that source NOW
agreeeeed^
just saw this thread for the first time.
gona take a look
Even though its a good idea in theory..
Once you start over clocking the GOS you really gona kill not only battery life, but your phone will be able to cook eggs lol
Thanks kingklick, hopefully you'll figure it out so we can run newer 3D games better, as for overheating, dzo (who created this modification) says that he hasn't noticed anything differences in temperature after overclocking the GPU on the Vogue
It will be cool if someone can overclock the GPU , these days everybody is talking about overclocking the cpu , but what allot of people do when they are traveling is playing games, some games we cant play because they are a little bit to heavy for the GPU. Kingklick thanks for looking at this topic,
personally I have really much confidence in you!
For this to work.
Someone has to implement this for MSM72XX, which is completely different than what's in the TI phones.
I am not sure if that code is even in the kernel (GPU clock). I don't recall seeing it but I haven't really looked at the kernel all that much.
It's all in the board files if it's anywhere.
/arch/arm/mach-msm/
Are you sure it's safe to OC the GPU?
It's not like we're using a adreno or Power VR in these phones. Also doesn't android use the CPU for things like scrolling,animations,swiping,etc...
For this to work.
Someone has to implement this for MSM72XX, which is completely different than what's in the TI phones.
I am not sure if that code is even in the kernel (GPU clock). I don't recall seeing it but I haven't really looked at the kernel all that much.
It's all in the board files if it's anywhere.
/arch/arm/mach-msm/
Click to expand...
Click to collapse
Um, the HTC Vogue (or Touch CDMA) AFAIK uses the Qualcomm MSM7500 which pretty much has the same GPU, just clocked slower than in the MSM7XXXA series, so I think it's just a case of recoding the modifications made by dzo (android kernel developer for the Vogue) for other phones.
The GPU code for the MSM chips is in /drivers/video/msm/gpu/kgsl/
I haven't looked at any of the code much, but kgsl.h defines some structs for what I think are clock values and in kgsl.c it contains:
Code:
/* the hw and clk enable/disable funcs must be either called from softirq or
* with mutex held */
static void kgsl_clk_enable(void)
{
clk_set_rate(kgsl_driver.ebi1_clk, 128000000);
clk_enable(kgsl_driver.imem_clk);
clk_enable(kgsl_driver.grp_clk);
}
Which to me looks a lot like its setting the GPU clock at 128Mhz.
Just my $.02
Geniusdog254 said:
The GPU code for the MSM chips is in /drivers/video/msm/gpu/kgsl/
I haven't looked at any of the code much, but kgsl.h defines some structs for what I think are clock values and in kgsl.c it contains:
Code:
/* the hw and clk enable/disable funcs must be either called from softirq or
* with mutex held */
static void kgsl_clk_enable(void)
{
clk_set_rate(kgsl_driver.ebi1_clk, 128000000);
clk_enable(kgsl_driver.imem_clk);
clk_enable(kgsl_driver.grp_clk);
}
Which to me looks a lot like its setting the GPU clock at 128Mhz.
Just my $.02
Click to expand...
Click to collapse
I modified this on the Eris's kernel to 160000000 and ended up with slower benchmark scores in both Neocore and Quadrant. So... possibly not that.
ivanmmj said:
I modified this on the Eris's kernel to 160000000 and ended up with slower benchmark scores in both Neocore and Quadrant. So... possibly not that.
Click to expand...
Click to collapse
Can you try others like:
352x
480x
176x
245x
Ace42 said:
Can you try others like:
352x
480x
176x
245x
Click to expand...
Click to collapse
384 gave me close to normal speeds but still slower than 128... 160 was slower.
It doesn't make sense.
If overclocking the GPU is anything similar to the CPU, then shouldn't we be using multiples of 19.2? instead of 128, try 147.2 or 176.4 etc....
mejorguille said:
If overclocking the GPU is anything similar to the CPU, then shouldn't we be using multiples of 19.2? instead of 128, try 147.2 or 176.4 etc....
Click to expand...
Click to collapse
But the stock is set to 128, which is not a multiple of 19.2.
If messing with it caused the phone to go slower, then it must be possible to mess with it to make it go faster.
dzo, linked me to his commit and it doesn't correspond to anything in the eris source.
http://androidhtc.git.sourceforge.n...;hpb=110e73c591db3fd23d8558659f8a6a0dfe5ba912
Time to poke around...
Owning both a Kaiser and a G1, the Kaisers 3D is actually faster. A lot of apps render quicker and smoother. Games play better too.
Would be nice if someone could pick this up, and get some over clocking action going for the GPU. Even without the overclock, my Kaisers GPU seems so much faster, but yet it's a weaker GPU then what my G1 has.
I am posting here as I am not allowed to do so in development subforum.
Anyway, I am the developer of the OpenFIMG project (formerly GLES6410), which is aiming to provide proper OpenGL support on devices with FIMG 3DSE 3D engine, found in S3C6410, S5PC100 and probably also in S5P6442. The project is in a pretty advanced state as it is already capable of running Android 2.3 with hardware acceleration. Still many OpenGL extensions and some core features (like lighting) are still missing. More info can be found here: https://github.com/tom3q/openfimg/wiki.
It is very likely (and almost confirmed) that the SoC used in Galaxy 3 (S5P6442) contains the same GPU as the one in S3C6410, which is the chip inside Galaxy Spica and similar phones, at least basing on what Quadrant and GLbenchmark show and on GL libraries supposed to be dumped from Galaxy 3.
What I am trying to say is that my project may also be useful on Galaxy 3, but I am the only developer working on it and I am doing it in my free time, so it does not progress as fast as one may expect. In other words, I am looking for some other developers interested in this project.
If you are interested, then do not hesitate to drop me a PM.
Mod EDIT : moving this to development
EDIT: Attached some documents about FIMG 3DSE (based on S3C6410 documentation and my reverse engineering)
EDIT: The project has been successfully used on G3. Builds of ICS for G3 use OpenFIMG as primary graphics driver currently and there are update packages for CM7.
Very interesting .. I send you PM.
Galaxy 3 has no GPU.
mpbm23 said:
Galaxy 3 has no GPU.
Click to expand...
Click to collapse
Based on what?
All my sources state that it has exactly the same FIMG 3DSE rev. 1.5 as in S3C6410.
Based on the fact that no site says that the I5800 has a GPU and that graphics on games are really laggy.
Unless you are not talking about a discrete GPU.
No mobile phone contains a discrete GPU. All of them are embedded inside some SoC chip, some are better (SGX, Adreno), some are worse (FIMG 3DSE).
It is exactly the same as with Spica and similar phones. Games are laggy because the hardware is not a speed daemon and the performance is even more impacted by really _broken_ drivers.
Then why nobody write that it has GPU like all the other phones like Galaxy S ,Nexus S etc?
mpbm23 said:
Then why nobody write that it has GPU like all the other phones like Galaxy S ,Nexus S etc?
Click to expand...
Click to collapse
Because it is a low end phone, software support for this GPU is very bad and the GPU itself is not a speed daemon.
tom3q said:
Because it is a low end phone, software support for this GPU is very bad and the GPU itself is not a speed daemon.
Click to expand...
Click to collapse
Okay..so your project involves developing real good drivers so that even the g3 low end gpu can perform better, right?
Sent from the 3rd Galaxy !
Yes. Of course, it will not work on par with Adreno 200 or similar GPUs, but should at least work a bit better. The main target is to run Android 2.3 (and future versions, which will not work will original drivers, because of missing extensions) with full hardware acceleration and without bugs found in Samsung drivers.
Hmmm.. Nice .. Good luck with your project.. One of our devs marcellusbe is working on porting CM7 to G3.. This would surely help him ..
Sent from the 3rd Galaxy !
cool... if so this is gonna make new benchmark .. pun intended ... best of luck man.... I had thought of throwing this phone a long time back but you guys always give hopes... and ofcourse result..
Wow interesting!
This should be moved to dev section I have sent a PM to haree
Cool. So what can others help with?
Too interesting
I'ld like to help with whatever i can
I'll back you up in spirit, sorry,i know nothing about programming
VERY intresting!
and yes the galaxy 3 does have a gpu and yes it has been confirmed (or atleast i heard) that it has this gpu..cant wait to see the outcome!
Actually, I will need some people who would compile, test and eventually fix the code for Galaxy 3, because the only phone with this GPU I have is a Galaxy Spica (i5700) and there may be some subtle differences between them. (Especially in the kernel part, where kernel modules may require changing of some addresses or some other fragments of code.)
Preferably, I would like someone to help me with the project, but I understand this is not an easy task, so not everyone can apply.
I don't mind being a tester
Edit: also will we be able to play gameloft games and nfs shift?
Sent from my GT-I5800 using XDA App
dilzo said:
I don't mind being a tester
Click to expand...
Click to collapse
Nice, thanks.
However, you have to understand that it is not a simple drop-in replacement of standard GL libs. This project replaces the whole graphics subsystem of the phone, including some kernel modules and this makes testing a bit more complicated as it needs pretty big modification of the phone software.
If it is not a problem for you then ok.
dilzo said:
Edit: also will we be able to play gameloft games and nfs shift?
Click to expand...
Click to collapse
It all depends if all the extensions required by these games will get implemented. Also, there may be some problems with screen resolution of Galaxy 3, which is a not standard one. I cannot give any claims regarding the performance as the real performance of this hardware is unknown.
We're in the dawn of mobile revolution; I appreciate the effort of Devs for fixing, tweaking, modding just for us to enjoy this phone... however, i have some queries and points about this dual core trend lately?
first is Android itself; i read somewhere that 2.3.4 does have true multi-core support..
quoted from one of the Devs over S2's board
In Androids before 2.3.4 there is no "Real" multicore support. just a sort of SMP "hack". They simply restrict all threads of a process to one core. This is, in some situations, faster in singlecore applications than the real thing. But it requires more power and multicore applications can't profit. In Android 2.3.4 and later, there is a part from Honeycomb in it. True Multicore support.
Click to expand...
Click to collapse
Did a little more reading, and although Gingerbread 2.3.x is not truly dual core optimized like Honeycomb or the new 2.3.4 , multithreaded programs will still make use of the 2nd core for functions such as garbage collection, background syncing, preloading pages, etc...
Can Devs here shed us some light about this matter?
Based from infos that i gathered here and in S2's board; i think HTC's approach is more efficient (where second core kicks in when the first core is loaded) than that of Samsung (Full time Dual Core)... Why would devs is forcing the second core making it always alive? I mean, Android does run butter smooth with 1ghz single core processor (Xperia Arc and Incridble S for example), so making both cores always on will only have greater battery drain without significant boost in performance?
Instead of turning both cores on, why not just improve HTC's efficient approach?
making second core breath when 75%load of the first core is reach for example or optimize the governor of the second core when assisting the first core?
Please enlighten me...
I agree with you on that this phone is very fast with just the way it is I don't even overclock no need to or nor will I change my stock kernal . I'm good with just letting it do its thing and enjoy my insane battery life and speed on the roms these devs are designing some great work on this board
Q&A section.......
nccu_eagles said:
Q&A section.......
Click to expand...
Click to collapse
+1.........seriously its like they do it on purpose :-\
Sent from my HTC Sensation Z710e using xda premium
Sorry guys... i tried to put some post important points before in Q/A but it seems it was ignored... Mods can transfer this anyways...
Now, most of Devs are diving to this Dual Core thingy...
I salute you guys for making all of this for us but its better if you can share some thoughts about my queries here... itll be a benifit to us users if we know what we are doin and what we get with our device...
Dont get me wrong guys, im no expert thats why im asking all of this...
Madness? This.. is... dual.. core!
On a serious note, i don't think there's anything wrong with people experimenting with trying to get both cpu's to work together. However for a lot of people i get the feeling that it's all about benchmarks. That's all i ever seem to read about these days. Who cares about benchmarks really? They're not important, only real world performance is king imo.
LOL... ala 300?....
Thats my point... its all good if both SW and HW runs hand-onhand...
The phone does seem alot snappier with the second core enabled.
jaytana said:
Instead of turning both cores on, why not just improve HTC's efficient approach?
making second core breath when 75%load of the first core is reach for example or optimize the governor of the second core when assisting the first core?
Please enlighten me...
Click to expand...
Click to collapse
That part only came about because they're trying to get both cores overclocked and get the governors to stick. The devs are trying to figure out aSMP architecture, and it's going to take a bit.
Sent from my HTC Sensation 4G using XDA App
jaytana said:
Instead of turning both cores on, why not just improve HTC's efficient approach?
making second core breath when 75%load of the first core is reach for example or optimize the governor of the second core when assisting the first core?
Please enlighten me...
Click to expand...
Click to collapse
chef of insertcoin baadnewz, i think is already on it. as someone of the forum suggest that is it much better to enhance HTC stock kernel rather than both cpu are active, i tested it and the heat is much faster. also, i think there is a reason to this kernel issue that is why they haven't release a OTA 2.3.4 or 2.3.5 for asia yet IMO.
good to hear that... There's nothing wrong of turning both CPU on as long as its well optimize and efficient (S2 for example with its SMP; im on 30hrs mark with my moderate to heavy usage)... i understand we still have less progress on 2.3.4 build because its still new... MORE POWER to the DEVS..
So I got my i9500 and already did some foolery with it.
Fine device, but I hate the raised lip around the screen edge. Something I definitely did not miss on the S3 and something very annoying.
Other than that small design critique:
THE ****ING PHONE ISN'T RUNNING FINAL FIRMWARE!
Basically the CPU is running on the cluster migration driver, meaning it switches all four cores from the LITTLE to the big cluster, as opposed to the core migration driver who does this in an individual core-pair manner.
You can pretty much throw all battery comparisons out of the window: it's completely unfinished and unoptimal.
I already compiled the kernel and flashed it without the cluster migration tidbit, but the phone won't boot. So yea. Current sources also useless.
Cleverly enough: you can't really distinguish between the two drivers apart from one manner: if /sys/devices/system/cpu/cpufreq/iks-cpufreq/max_eagle_count is present, you're running an IKS driver. If it's not, then you're running the sub-optimal IKCS driver.
So yea. We'll see what Samsung does about this, currently the advantages of big.LITTLE are pretty much unused.
Another nail in the coffin on how rushed and unprepared this phone has been.
Wow, this is seriously turning out to be a fiasco.
ChronoReverse said:
Wow, this is seriously turning out to be a fiasco.
Click to expand...
Click to collapse
This is EXACTLY why at the end I don't care for technical details about socs but was rather waiting for real world usage first. As much I wanted to agree with Andrei Lux on how intelligent BigLittle is, I sort of felt that it wont be same at the end.
Question is now: Is this possible to fix in the near future?? So that maybe buying the Exynos will be beneficial if the devs take over. I wont bet on Samsung introducing mind-blowing improvements in that department in upcoming firmwares
Xdenwarrior said:
Question is now: Is this possible to fix in the near future?? So that maybe buying the Exynos will be beneficial if the devs take over. I wont bet on Samsung introducing mind-blowing improvements in that department in upcoming firmwares
Click to expand...
Click to collapse
The code other driver is there in the kernel, it's just not used. No idea. It's not like we need Samsung for it: I already talked to a developer at Linaro about some incomplete switcher code that's being currently getting the green-light to be made public. But who knows how long that will take.
Whatever the case, I gather that they can't just let it be in the current state.
AndreiLux said:
The code other driver is there in the kernel, it's just not used. No idea. It's not like we need Samsung for it: I already talked to a developer at Linaro about some incomplete switcher code that's being currently getting the green-light to be made public. But who knows how long that will take.
Whatever the case, I gather that they can't just let it be in the current state.
Click to expand...
Click to collapse
Any way to just disable cortex a15 altogether yet just to see how well cortex a7 will perform in simple texting, browsing, calling and to see what the battery life will be like on that?? (cause cortex a7 only uses like 200 something mw as opposed to 1000mw for snapdragon). I know u wont be able to game. How often does Cortex A15 hits in? cause I would suspect a much worse battery life with incomplete drivers doing the switching if its very often on. But PocketNow reports very similar battery results to snapdragon variant which I find odd
Xdenwarrior said:
Any way to just disable cortex a15 altogether yet just to see how well cortex a7 will perform in simple texting, browsing, calling and to see what the battery life will be like on that?? (cause cortex a7 only uses like 200 something mw as opposed to 1000mw for snapdragon). I know u wont be able to game. How often does Cortex A15 hits in? cause I would suspect a much worse battery life with incomplete drivers doing the switching if its very often on. But PocketNow reports very similar battery results to snapdragon variant which I find odd
Click to expand...
Click to collapse
Use any app to limit the CPU frequency to 600MHz. That'll limit it to the A7 cores running to 1200MHz. Basically you can just use CPU-Spy. Everything <= 600 are A7's mapped at half frequency, everything above it are A15's at 1:1 frequency.
As for PocketNow: irrelevant. The difference is what could be instead of what is, the Snapdragon doesn't play a role in the discussion here.
WOW , thats sucks
Samsung was too rushed and ruined it :/
AndreiLux said:
Use any app to limit the CPU frequency to 600MHz. That'll limit it to the A7 cores running to 1200MHz. Basically you can just use CPU-Spy. Everything <= 600 are A7's mapped at half frequency, everything above it are A15's at 1:1 frequency.
As for PocketNow: irrelevant. The difference is what could be instead of what is, the Snapdragon doesn't play a role in the discussion here.
Click to expand...
Click to collapse
Hey thanks, but I don't have the S4 to test it with since i'm still debating on which to get. I live in Canada and so the only version here which I can get a lot cheaper on a contract is LTE snapdragon, but I wont mind getting the Exynos since it got potential. Besides 16GB internal isn't enough for me. So that's why asking if u seen any improvements in battery when only cortex a7 ran? If a7 doesn't do much in power consumption, then no point spending 800 bucks and loosing LTE altogether...
@bala_gamer please see my PM its important...
Sent from my GT-I9500 using xda premium
Oh wow. Just got word (without further in-depth explanation) that this might actually be a hardware limitation. Coming from a reliable source.
No words...
AndreiLux said:
Oh wow. Just got word (without further in-depth explanation) that this might actually be a hardware limitation. Coming from a reliable source.
No words...
Click to expand...
Click to collapse
Can you elaborate a bit more pls?
Sent from my GT-I9500 using Tapatalk 2
that's not what samsung exynos advertised..
http://www.youtube.com/watch?v=U6UNODPHAHo
Is it possible that we're having a simpler Exynos 5 system technically closer to Exynis 5 Quad (plus 4 A7 cores) than a real seamless Octa-core system? It was strange reading that "Octa-core manufacturing starts in Q2" (April-June) then see Octa-core versions hitting reviewers early April, that's way too low time frame. Maybe this is a 1st-gen 5410. In any case, performance and current-state battery life beats the Snapdragon version, even if only just.
AndreiLux said:
Basically the CPU is running on the cluster migration driver,
Click to expand...
Click to collapse
wtf? Well done Samsung... This is ridiculous...
AndreiLux said:
Oh wow. Just got word (without further in-depth explanation) that this might actually be a hardware limitation. Coming from a reliable source.
No words...
Click to expand...
Click to collapse
WHAT THE [email protected]??!!
Actually WTF is a massive understatement here....!!!
Please can you give more info about this matter whenever is possible? This is very serious...
Is it a specific hardware limitation? Something that Samsung specificly did in GS4 (I9500) ?
Because this can't be a generic exynos octa limitation. It makes no sense... Unless everything we've read from Samsung and ARM about exynos octa, are completely misleading...
A hardware limitation..? They advertised the functionality and to then release a device without it, is just plain stupid. Hopefully it is a just a kernel issue and can be resolved quickly.
Sent from my Galaxy Nexus using xda premium
Probably Samsung will implement it in their Note 3 device? It's a conspiracy so that people buy their next Note phone but this news is sad.
Sent from my GT-I9100 using Tapatalk 2
Now what is this all about? Is this a very serious issue?
So its either all A15s or all A7s?
so would the 'octa' really be a better choice than the S600? That should be powerful enough.. and the S600 is pretty power efficient too
rkial said:
So its either all A15s or all A7s?
so would the 'octa' really be a better choice than the S600? That should be powerful enough.. and the S600 is pretty power efficient too
Click to expand...
Click to collapse
What I understood is its either the full cluster of a7 or a15 is used/ functional based on the load, dynamically turning on one or two cores of a15 to work along with a7 may not be possible it seems.
I may be wrong, waiting for an elaborate exp from andrei
Sent from my GT-I9500 using Tapatalk 2
bala_gamer said:
What I understood is its either the full cluster of a7 or a15 is used functional based on the load, dynamically turning on one or two cores of a15 to work along with a7 may not be possible it seems.
I may be wrong, waiting for an elaborate exp from andrei
Sent from my GT-I9500 using Tapatalk 2
Click to expand...
Click to collapse
I was always under the impression this was the intention of Samsung's particular implementation of it. I thought it was common knowledge that Samsung's version worked on a 4 or 4 (A15) or (A7) basis.
Maybe he was talking about the ability to change that.